The International Linear Collider

Global Design Effort

Baseline Configuration Document

Change History

March 12, 2007 - Updated BDS (single-IR, two-detector push-pull)March 4, 2007 - Updated RTML (to accommodated centraled DRs).Dec. 1, 2006 - Updated Electron SourceNov. 5, 2006 - Updated Parameter and Layout (Centralized DR installation around IR)Nov. 2, 2006 - Updated: Beam Delivery (Surface assembly of detectors)Sep. 23, 2006 - Updated: Beam Delivery (Muon spoiler implementation)Sep. 21, 2006 - Updated: Damping Rings (single ring for positrons)Sep. 21, 2006 - Updated: Beam Delivery (14mrad/14mrad crossing angle)July 13, 2006 - Updated: Main Linac (layout of Q-steering-BPM packages)May 8, 2006 - Updated: Parameter (with the new Layout subsection)Apr. 23, 2006 - Updated: Beam Delivery (Fix broken links)Apr. 23, 2006 - Updated: Ring-to-Main-Linac section (Maintenance update: consistency

with the latest lattice) Mar 28, 2006 - Updated: Ring-to-Main-Linac section (Beam diagnostic station in the

upstream end)Mar 23, 2006 - Resotored missing Fig.1 in “Number of ILC Tunnels” section in the

BCD White Papers (Fig.1 was lost during the archival process in Jan., 2006).

Mar 16, 2006 - Updated: Conventional Facility and Siting section (Asian sample site)Mar. 3, 2006 - Updated: Parameter Section (Linac Angle)Mar. 3, 2006 - Updated: Ring-to-Main-Linac section (Layout)Feb. 27, 2006 - Updated: Damping Ring section (RF frequency)Feb. 4, 2006 - Updated: Operations and Availability section.Dec. 20, 2005 - Updated: Conventional Facility and Siting section.Dec. 12, 2005 - Initial Release, accepted by GDE at Frascati, Italy

Table of Content

1. Baseline Parameter2. Electron Source Description3. Positron Source Description4. Damping Ring5. Ring to Main Linac6. Main Linac7. Beam Delivery System8. Cost and Engineering9. Conventional Facility and Site Development10. Operations and Availability11. Instrumentation and Controls

Baseline Parameters 1

1. Baseline Parameters and Layout

This section outlines the essential aspects of the ILC baseline parameters and the layout of the ILC accelerator complex.

1.1 Baseline Parameters

The tentative parameter set for the ILC [1] was distributed in February 2005. It consisted of two parts:

Linac parameter set Beam parameter range

The working groups suggested possible changes, in particular during the Snowmass workshop. Not all proposed changes were supported by compelling arguments, and were rejected (see “Justification” section, below). The following changes are needed, based on recommendations mainly from WG2 and WG5:

The nominal operating accelerating gradient should be 31.5 MV/m for the 500 GeV stage, and 36 MV/m for the upgrade stage, with Q0=1010 for both cases. The RF system should be designed so as to be able to provide 35 MV/m and 40 MV/m, respectively.

One RF unit for the baseline design for 500GeV stage should consist of one 10MW klystron and 3 cryomodules each containing 8 cavities.

Taking into account these changes, slightly revised parameter sets are given in Table 1.1 The spirit of the parameter sets however, do not change and are described below.

The proposed beam parameters are grouped within 5 sets rather than one. In the past a collider project in most cases used to provide a unique set of parameters which were tuned to give the highest luminosity. In actual machine operations, however, unexpected or underestimated difficulties require the adoption of operating conditions different from those assumed in the initial design. It is desirable to provide for such changes in the initial design. Since these changes are not very predictable, an operating plane is defined, rather than an operating point. Requiring to accept a wide range of parameters may introduce challenges in the design, but the resulting machine operational flexibility is deemed to be more valuable. If the machine, as built, works for a wide range of parameters within the operating plane, then it should be easier to reach the design luminosity.

The five sets are:

Rev. October 28, 2006

Baseline Parameters 2 Nominal set used as a reference to scale to the other sets

Low bunch charge (low Q) reduction in bunch charge by a factor of two

large *y(large Y) increased vertical beam size due to factor two larger emittance growth in LET

Low power (low P) reduced number of bunches (nb) by factor of two

High luminosity (high L) smaller IP beam size and shorter bunch. (This is not a part of the `baseline' but is added just for reference.)

Among these, the first four sets give the same luminosity -- , to be compared with the TESLA TDR peak luminosity , while the high L parameter set gives . As described above the parameter sets are not intended to be considered as fixed sets, but as an indication of the degree of flexibility intended to be built into the machine. Hence each sub-system should accommodate – where possible – the most demanding parameters. The final ‘operating point’ is almost certainly going to be within the space defined by these parameter sets, but will not necessarily correspond to any one of them. The sets of important parameters, which bound the operating plane, are (they are of course related) given in Table 1.1 below:

min nominal maxBunch charge N 1 2 2 x1010

Number of bunches nb 1330 2820 5640Linac bunch interval tb 154 308 461 nsBunch length z 150 300 500 mVertical emitance y 0.03 0.04 0.08 mm.mradIP beta (500GeV) x 10 21 21 mm y 0.2 0.4 0.4 mmIP beta (1TeV) x 10 30 30 mm y 0.2 0.3 0.6 mm

Table 1.1 Baseline Parameter

The alternative of designing to a single ‘default’ parameter set – while appealing from the point of view of the sub-system designers – would effectively remove the margins and flexibility outlined above. Since limiting factors in machine performance are not yet known, it is extremely desirable to maintain all such options in the parameter plane. In addition, allowing for the overhead also keeps open the option of achieving a luminosity greater than the nominal if not as high as High L.Rev. October 28, 2006

Baseline Parameters 3Justification of the choice

A few problems associated with the baseline parameter sets and some possibilities of different parameter sets were pointed out from working groups during the Snowmass workshop.

1. Short bunch

Low Q and High L parameter sets demand a bunch length as short as 150 m, to be compared with 300 m in the nominal set. WG1 concluded that, though a single-stage bunch compressor is just enough for compressing a 6mm bunch from the damping ring to 300 m, a two-stage compressor is mandatory for a shorter bunch length or for a DR bunch longer than 6mm.

Obviously a two-stage compressor is more expensive and requires a longer site. However, exact cost and length differences are not yet known. A shorter two-stage compressor is still under consideration. A two-stage compressor may be desirable even for the nominal parameter set. Thus, it is recommended that the possibility of a shorter bunch be retained.

2. Long Pulse Length in the Main Linac

H. Padamsee and B. Foster suggested a longer beam pulse length (accordingly a longer klystron pulse length) in the main linac with the same pulse charge. If, for example, the beam pulse length is doubled (2ms), the beam current and, therefore, the number of modulators/klystrons would be halved, which reduces the cost of the RF system.

The cost of the cryogenics system, on the other hand, increases due to the higher duty factor. Taking into accout all the above, the proposers of this idea expect the total cost to decrease. It was also pointed out a longer pulse would ease the demand on the MPS and is also better for the detector performance. However, according to C. Adolphsen, the total cost increases slightly. A more detailed study of the cost is needed. In addition to the cost uncertainty, a longer pulse requires a longer modulator/klystron pulse which is not currently available. Thus, this possibility is ranked as an `alternative’ configuration.

3. Larger Number of Bunches

The Low Q parameter set demands a number of bunches as large as about 6000. This is particularly demanding for the design of the Damping Ring, including the need for development of very fast kickers. At this moment we retain the operating condition with up to 5640 bunches per pulse as part of our baseline configuration.

Rev. October 28, 2006

Baseline Parameters 4

4. Very Low Q Parameter Set

J. Gao proposed a `Very Low Q' parameter set in which the bunch charge is 0.6×1010

(40% lower than the nominal). This demands more bunches (6000), a shorter bunch (120 m), a higher repetition rate (8 Hz) together with a tight focusing at the IP. The main motivation is to reduce the space-charge effect in the Damping Ring (it eliminates the need for a coupling bump) and a smaller disruption (Dy=9). With this parameter set the smaller IP spot size (horizontally and vertically) and the shorter bunch length is very demanding. It is necessary to evaluate if these demands outweight the benefits of relaxing the space-charge and the disruption. Accordingly, this option has been rejected.

5. Smaller Number of Bunches with Higher Rep Rate

S. Guiducci suggested reducing the number of bunches by 2 (as in the low P set), but maintaining the 1ms pulse. Luminosity is recovered by increasing the repetition rate to 10 Hz, requiring a factor of two reduction in the damping time (possible with a factor two shorter ring, assuming the same kicker rise time). The resulting factor two drop in linac beam current halves the peak beam power and could potentially halve the number of the 10MW klystrons; but the fill time also increases, thus reducing the efficiency and increases the dynamic cryoload by a factor of two. Increasing the vertical emittance by a factor of two, while reducing the horizontal emittance by the same factor was also discussed (Guiducci); this option increases the beamstrahlung by a factor of four, however, this assumes that no modifications to the demagnification in the final focus are made. Accordingly, this option has been rejected.

6. High L

A. Seryi pointed out a difficulty in designing the IR region for the High L set especially at 1TeV. The disrupted beam has a long low-energy tail due to the strong beamstrahlung. The lowest energy that the dumpline can accept is about 0.3E0. The integrated power below this energy which the detector system can tolerate is less than 10 W. With the present 1TeV High L set, this limit is greatly exceeded. To solve this problem, the beamstrahlung (especially the Upsilon parameter) has to be reduced. A longer bunch (300 m), larger x

and y, a lower vertical emittance 0.023 mm.mrad coupled with an increased bunch charge ( ) has been proposed. This set tightens vertical tolerances (smaller DR emittance and a smaller emittance growth budget) and a larger disruption (kink instability). The requirements can be a little relaxed by choosing the nominal bunch charge but with a slightly decreased luminosity from the High L set.

The required change is within a reasonable range from the standard set and may be feasible by the time when the 1 TeV upgrade is implemented. This option is not listed in the standard parameter sets since the High L set itself is not the baseline.Rev. October 28, 2006

Baseline Parameters 5

Required R&D for the Baseline

The performance of the proposed two-stage bunch compressor needs to be confirmed, including errors. It is also desirable to see if a shorter two-stage compressor is feasible.

A Damping Ring design capable of storing 6000 bunches need to be established.

Required R&D for the Alternative Parameters Long Pulse with 5 Hz or 10 Hz

The technical feasibility and cost of a long-pulse RF system needs to be investigated.

1.2 Layout of the ILC Accelerator Complex

1.2.1 Overview

Figure 1(a) shows a schematic view of the baseline layout for ILC in its phase-1 configuration which supports physics experiments at center-of-mass energies up to 500 GeV. The site footprint is dominated by two main linacs, for accelerating electrons and positrons, respectively, each of which extends over a length of approximately 10 km. Together with the electron sources, damping rings and the beam delivery systems, the total site length is expected to be approximately 31 km.

Figure 1(b) shows a scaled, closed-up view in a 3-dimensional perspective which focused on the positron injection and electron extraction from the damping ring complex.

Figure 1(c) gives a possible layout of the conventional facilities which correspond to the beam lines shown in Figure 1(b), for illustration purposes. Details of the exact topologies are subject to changes through further engineering design efforts.

1.2.2 Electron SourceThe electron source is the point of origin of all ILC beams in normal operation. The positron beam in normal operation will be generated by photons produced by the passage of electron beams through the undulators. The electron source will be part of the Central Injector/Damping Ring complex (see Fig.1(a)). It is located on the e+ linac side of the the Damping Ring which has an electron and a positron ring in the same tunnel. The exact location of the electron source will be determined through optimization of the injection beam line into the damping ring. The polarized electron source system includes one 5GeV linac to inject the beam into the electron damping ring.Rev. October 28, 2006

Baseline Parameters 6(a)

(b)

(c)

Figure 1: (a) Schematic view of the ILC baseline layout for phase-1; (b) Scaled 3-D schematic view on one side of the damping ring injection/extraction; (c) A possible layout topology of conventional facilities for the area shown in (b).

Rev. October 28, 2006

400m: Collimators, 55-deg spin rotation, Bending & Matching Dump

600m: e- Injector, Keep-Alive source, 5GeV Booster

587m: 2% Ramp, 5GeV e- Beam, 0.4GeV e+ Beam

Baseline Parameters 7

1.2.3 Damping RingsThe damping rings (DRs), operated at the beam energy of 5 GeV, will reduce the beam emittance to the level required by the specifications discussed in Section 1.1. The rings are approximately 6.7 km in circumference. The baseline layout has the damping rings located centrally between the linacs and above the level of the beam delivery system. The elevation difference will give adequate shielding to allow operation of the injector/damping-ring (INJ/DR) complex with other ILC systems in open access. One electron and one positron ring in the same tunnel are assumed in this baseline configuration. A study on timing issues at ILC [2], where positrons are produced with undulator photons from electrons of previous bunches, gives the following constraint: The difference in path length, from the production point to collision point, between the positrons and the electrons, including the distance between injection and extraction points in the damping rings, should be an integer multiple of the damping ring circumference. It is noted that simple longitudinal (parallel to the main linacs) relocation of the damping rings does not affect this timing constraint. Consequently, the exact locations of the damping rings may be site-dependent to a certain extent. However, other changes to the layout of ILC systems, such as relocation of the damping rings transverse to the main linacs, the lengths of the linacs or the beam delivery systems will change timing requirements and solutions.

1.2.4 Ring-to-Main-LinacThe RTML beam line begins at the damping ring and transports the damped beam through the linac tunnel to the low energy end where it turns through 180 degrees. The RTML includes a suitable set of beam diagnostics, bunch compression and spin manipulation sections. The 180˚ turn-around allows the application of a feed-forward beam stabilization system. After the turn-around, in the bunch compressor, the beam is accelerated up to 13-15 GeV before injection into the main linacs. These 10 km, 5 GeV beam transport lines from the central complex to the beginning of the RTML bunch compressors are required to maintain the low emittance of the damped yet longitudinally uncompressed beams. They are therefore more complex in instrumentation and correction than the e+ transport lines that they replace in the previous ILC layouts.

1.2.5 Main LinacsThe main linacs receive the beams from the RTML at 13-15 GeV and accelerate them up to 250 GeV in normal phase-1 operation. As discussed below, the main linacs on the electron side are interlaced by undulators to incorporate the positron production system, or segments associated with beam diagnostics, and tune-up dumps for ensuring good operability.

The angle between the two main linacs is irrelevant to the design of most collider systems. An exception is that the Beam Delivery System requires this value for detailed optics design. The baseline design has two beam delivery transport lines leading to two interaction points where the beams interact with a 14 mrad crossing angle. Rev. October 28, 2006

Baseline Parameters 8

1.2.6 Positron SourcePositrons in normal operation are produced with photons originating from the electron beam from the main linac at an energy of 150 GeV. For that purpose a 200 m-long undulator system is incorporated in the electron main linac. The positron production target that converts these undulator photons to positrons is located on the “electron side” of the IR. After acceleration up to 400 MeV, the positron beam is brought to the central INJ/DR complex and accelerated to 5 GeV by a pre-accelerator before injection into the positron damping ring. The trajectory of the positrons upon extraction from the positron damping ring is essentially a mirror image of that of the electrons.

To allow commissioning or debugging of the positron damping rings, positron RTML and the positron main linac, an auxiliary (or keep alive) positron source will be provided at the entrance to the 5 GeV positron pre-accelerator.

1.2.7 Beam DeliveryThe beam delivery system receives the beams of energies up to 250 GeV in phase-1 operation, collimates them, focuses them, collides them at the interaction points, and safely disposes of the exhaust beam at the beam dumps. In the baseline layout, the ILC supports two interaction regions which provide collisions at beam crossing angles of 14 mrad..

1.2.8 Elevation LayoutIn the baseline configuration, the bulk of the main linacs will be built approximately following the “local horizontal” lines that are defined by the local gravity in its neighborhood. This is the preferred elevation layout from the standpoint of straightforward engineering implementation of the liquid He leveling within the main linac cryostats.

However, for sake of ensuring emittance controls and beam tuning, the entire beam delivery system beam line is to be built to stay in a single mathematical plane normal to the gravity vector at a point near the interaction regions. The INJ/DR complex will also be on a plane parallel to that of the BDS but displaced vertically from it by 10 to 20 meters.

1.2.9 Path Length ConstraintsThe fact that positrons are produced by undulator photons which are produced by electrons within previous machine pulses lead to special issues of timing controls. They lead to constraints or preferences on various aspects of the ILC design, including the choice of the circumference of the damping rings and all the beam line path lengths.

A detailed analysis of this issue has been reported by A.Wolski et al in [2]. It is understood that an integral relationship is highly desired between the circumference of e+ damping ring and the path length a new positron bunch travels in a round trip through the INJ/DR

Rev. October 28, 2006

Baseline Parameters 9complex, the RTML, the main linac and the beam delivery.1

1.2.10 Hardware Layout within the TunnelsFollowing the study presented in the Section 12.3 “Number of Tunnels” of GDE White Papers (http://www.linearcollider.org/wiki/doku.php?id=bcd:bcd_home ), the hardware for the main linacs (and many of other subsystems) will be implemented in two parallel-running tunnels. More discussions are found in the “Conventional Facilities and Siting” section of BCD.

1.2.11 Energy Upgrade and the Accelerator LayoutThe first phase (500 GeV centre-of-mass) will be constructed using a tunnel long enough to achieve 250 GeV final energy in each linac with an average gradient of 31.5 MV/m (~2×10 km assuming a 0.75 fill factor). A second phase (phase 2) upgrade to 1 TeV centre-of-mass will then require extending the tunnel (away from the IR) an additional ~2×9.3km assuming cavities capable of 36 MV/m operational gradient.

The Beam Delivery System will be configured so that it can support operation at 1TeV with only minor upgrades. The main dump systems will be configured for the 1 TeV option from the start.

Injectors will be configured for phase 1 so that there is a minimum impact on them when upgrading to phase 2. Prospective sites must be chosen with the TeV phase 2 machine in mind; specifically the availability of both the total required land and power.

References

[1] http://www-project.slac.stanford.edu/ilc/acceldev/beamparameters.html, http://www-project.slac.stanford.edu/ilc/acceldev/beampar/Suggested ILC Prameter Space.pdf

[2] “Recommendations for ILC Configuration Satisfying Timing Constraints,” by H.Ehrlichmann, S.Guiducci, K.Kubo, M.Kuriki and A.Wolski, submitted to GDE Executive Committee on April 7, 2006. Available from: http://www.linearcollider.org/wiki/lib/exe/fetch.php?cache&media=bcd: timingrecommendations-revapril17 .pdf .

1 The previous baseline solution for these issues was the insertion of 1.2 km into the positron linac comparable to the positron production region in the electron linac. This insert would contain a few hundred meter path length adjuster which is no longer required as the interaction regions and detectors are now at the same longitudinal location. The whole 1.2 km insert in the e+ linac has been removed and the “Keep alive source” is now part of the central complex. See 1.2.6. The positrons now pass through an additional half turn in the damping ring between injection and extraction and this partially corrects the e+/- timing. Further correction will be required but the amount is dependent on the lengths of so many systems which are presently under review, that a decision on the optimum correction methodology will be delayed until some time in the future. There are many possible timing correction strategies which will be considered at this time.and compared with the cost of a 1 km insert in e+ linac.

Rev. October 28, 2006

Electron Source 1

2. Electron Source Description

The conceptual design is shown in Figure 2.1. The injector consists of two DC guns incorporating photocathodes illuminated by two Ti:Sapphire drive laser. At one time only one gun and laser system will be in operation, the second system provides redundancy to achieve high source availability. The thereby produced long electron microbunches (~ 2 ns) are bunched in a bunching section consisting of two sub-harmonic bunchers operating at 108 MHz and 433 MHz, respectively. A β matched traveling wave L-band buncher/pre-accelerator reduces the bunch length further to ~ 20 picosseconds (FWHM) and accelerates the beam to ~ 80 MeV. A chicane provides energy collimation before the bunch train is injected into the superconducting booster linac for acceleration to 5 GeV. A transfer line connects the electron source with the damping ring. The bending angle is chosen to provide spin precession by 90 degrees. Superconducting solenoids are used to rotate the spin vector by 90 degrees into a vertical orientation prior injection into the damping ring. Following the spin rotating solenoidal system, a single cryomodule will provide energy compression to achieve the energy spread specification required for damping ring injection. The source beam parameters for the ILC are summarized in table 1.

Figure 2.1 : Overview of the electron source injector concept.

The primary source requirements are given in Table 1.

Parameter Symbol Value UnitsElectrons per bunch1 4x1010(2x1010)† Number

Bunches per pulse 2820 (5640) † Number

Electrons Rev. Nov.28, 2006

Electron Source 2Microbunch repetition rate fmicro 3 (6) † MHzPulse Repetition Rate 5 Hz

DR energy Acceptance E/E 1 % (FW)DR Transverse Acceptance A=2J 0.09 m-radElectron Energy E0 5 GeVElectron Polarization Pe >80 %

1 twice IP requirement† Low charge parameters

Table 2.1: Electron source system parameters requirement prior to injection in damping ring:

The requirement for the laser parameters are shown in Table 2.1. The photocathode material choice will be GaAs/GaAsP strained superlattice cathode, which can provide the required polarization.

A detailed design of a possible solution for the injector is described in Reference [Curtoni and Jablonka TESLA-NOTE 2001-22]. A bunching section consisting of two sub-harmonic bunchers (SHB’s) operating a 108 and 433 MHz was proposed. The sub-harmonic bunchers will be followed by a beta matched traveling wave bunching/pre-accelerating section, which provides improved performance compared to the β=1 L-band bunchers described in the reference mentioned above. The sub-harmonic bunchers are scaled versions of SHB cavities in use at the TTF facility and do not present technical challenges. The bunching section is immersed in solenoids to control the transverse beam envelope.

Laser parameter Value UnitMicro-pulse energy at photocathode

~3 (1.5) J

Micro-pulse length ~2 ns# micro-pulse per train 2820 (5640) NumberIntensity jitter 2 % (rms)Micro-Pulse spacing 337 (169) nsRepetition rate 5 HzWavelength 750-850 nm

Table 2.2: requirements on the photocathode drive-laser in laser room (assumes a photocathode QE of 0.5 %):

Parameter Tables

Electrons Rev. Nov.28, 2006

Electron Source 3http://www.astec.ac.uk/id_mag/BCD/Electron_Source_Parameters.pdf (table)http://www.astec.ac.uk/id_mag/BCD/Electron_Source_Laser_Parameters.pdf (table)

Supporting Documentation

TESLA Report 2001-22

Required R&D

R & D Plan proposed by SLAC

Electrons Rev. Nov.28, 2006

Positron Source 1

3. Positron Source Description

Table 3.1 lists the performance requirements of the ILC undulator-based positron system and outlines key aspects of the system. The undulator based system uses a 100m long helical undulator placed at the 150GeV point of the ILC electron linac. The ILC electron beam passing through this undulator generates circularly polarized photons. The photons pair-produce in a relatively thin, high strength Ti-alloy target. The positrons from this process are collected and accelerated up to the damping ring energy of 5 GeV. A yield into the damping ring of 1.5 positrons per electron through the undulator has been chosen for the design as an operational safety factor. This overhead is manifested in extra photon beam power incident on target and in the power and peak energy handling capabilities of the pair-production target system as well as the power load considerations of the downstream capture systems.

Figure 3.1 shows the undulator located at the 150 GeV point of the electron main linac. There are several layout options for the undulator in the linac (chicane, dog-leg, etc). The undulator will consist of a number of identical segments, each segment being several meters long. The active length of the full undulator is 100 m. The example helical undulator parameters used are K=1, and a period of 10 mm. This undulator system is designed to operate at a fixed electron drive beam energy of 150 GeV with an overhead factor of 1.5 in the estimated captured positron yield. In order to describe undulator operation, the ILC linac energy can be divided into ranges of 50 to 150 GeV and 150 GeV and above. In the energy range of 150 GeV and above an electron beam of 150 GeV is passed through the undulator and the remainder of the linac is used to increase the energy to the value required. For energies below 150 GeV electrons of 150 GeV are still passed through the undulator and the beam is then decelerated in the remainder of the linac to the required energy.

Figure 3.1: Schematic overview of Undulator Based Positron Source in which the v.Dec.12, 2005

Primary e-

source

e-

DR

Target e- Dump

Photon Beam Dump

e+

DR

Auxiliary e- Source

Photon Collimators

Adiabatic Matching

Device

e+ pre-accelerator

~5GeV

150 GeV 100 GeVHelical

UndulatorIn By-Pass

Line

PhotonTarget

250 GeVPositron Linac

IP

Beam Delivery System

Positron Source 2undulator is located at the 150GeV point in the electron main linac. The option of the undulator in a chicane is shown here.

A key feature of the undulator based source is the capability of producing positrons with a longitudinal polarization of about 60%. The upgrade to a polarized positron source is easily achieved through the addition of ~100 m more undulator, photon collimation, polarimeters, and spin rotators as appropriate.

Photons are converted in a rotating target wheel of high strength, radiation hard, Ti-alloy. Positrons are captured downstream in an L-band RF linac operating at a nominal gradient of 15 MeV/m. After acceleration to 250 MeV, the captured positrons are separated from captured electrons in a magnetic chicane and injected into the 4.75 GeV booster linac for acceleration to the full damping ring energy of 5 GeV. The booster linac is a standard ILC L-band superconducting system operating at a gradient of about 25 MeV/m.

A “hot spare” positron conversion target and capture system is included in the design to enhance system availability. A keep alive source has been included in the design to add flexibility during commissioning, reduce the loss of machine time due to downtime (scheduled or otherwise) of the primary electron beam, and provides a source of polarized electrons for e-e- and ILC options. The proposed intensity for the keep-alive source has been specified at 10% of nominal charge.

Parameter Value UnitsPositrons per bunch ( )† numberBunches per pulse 2820 (5600) † numberPulse Repetition Rate 5 HzPositron Energy 5 GeVElectron Drive Beam Energy 150 GeVElectron Drive Beam Energy Loss

3.23 GeV

Undulator Period 10 mmUndulator Strength 1 -Undulator Type Helical -Undulator Length (unpolarized source)

100 m

Photon Energy (1st harmonic cutoff)

10.7 MeV

Max Photon Beam Power (unpolarized source)

147 kW

Target Material Ti-6%Al-4%V -Target Thickness 0.4 r.l.Max Target Absorption 11 kWIncident Spot Size on Target 0.75 mm, rms

v.Dec.12, 2005

Positron Source 3Positron Polarization (upgrade) 60 %

† Low Q Parameters

Table 3.1: Undulator-based positron system parameters:

Alternative Positron Source Description

The Compton scheme can also generate the high intensity positron beam which meets the ILC requirements with some margin. In contrast to the undulator scheme, the Compton scheme has advantages as follows;

It is a completely independent system from the electron arm. It avoids complex interference between two arms that is important feature especially for the commissioning. This is also important for good availability of the collider.

Because of the independency, there is much flexibility to change the beam structure, intensity, etc.

The performance can be much improved by introducing some new technology, e.g. more powerful laser system, more precise high-gain optical cavity,etc. These improvement can be made after enough confirmation of the new technology in the laboratory without any risk.

Because the operation is completely independent from the electron linac, there is no limitation at the lower energy operation. The collision at the energy from 5 GeV up to 250 GeV (500GeV eventually) can be easiliy made without any cost. On the other hand, in the undulator scheme, the energy spread is larger and/or the luminosity is lower at some energy region.

The injection scheme is totally different from that of the undulator scheme. Any fast kicker, which is one of the most difficult device in ILC injector, is not necessary.

It can generate the polarized positron. The polarization depends on the laser helicity which can be switched pulse by pulse. This polarization switching is very important for the physics experiment. It is hard to switch pulse by pulse the polarization in the undulator scheme requiring an extra section to implement the switching.

Addition to these advantage, the critical elements of the Compton scheme can be demonstrated prior to the real construction and partly demonstrated already. Then the Compton system can be developed step by step. This feature is remarkable in contrast to the undulator because the undulator scheme is hard to demonstrate with a reasonable scale

v.Dec.12, 2005

Positron Source 4prior to the construcution. Even these advantages compare to the undulator scheme, the Compton scheme is still in an initial stage to develop the real ILC positron source because a high current electron beam needs to be produced and a high power multi-bunch laser scheme is needed. However, this concept is so attractive that the people keep their aggressive R&D efforts towards the well matured design based on the Laser Compton scheme.

Compton Scheme Design

The Compton based polarized positron source [1] consists of (1) an electron linac which is the injector of the Compton ring, (2) a electron storage ring named Compton ring, (3) a laser which send laser beam to pulse stacking cavities, (4) laser pulse stacking cavities which are installed in a straight section of the Compton ring, (5) conversion target and capture section, (6) positron injector linac, and (7) a damping ring.

Now two versions of design are under consideration. One uses CO2 laser and the other uses YAG laser. The former is called CO2-version, and the later is called YAG-version. In following description, the number corresponds to CO2-version is firstly shown in each statement, then that of the YAG-version is shown in parentheses.

Compton ring is a high current electron storage ring. The energy of the ring is 4.1 GeV (1.3 GeV). The ring has long straight section in which 30 laser stacking cavities are installed. The circumference of the ring is 649.2 m (276.7m). Two trains are (one train is) circulating in the ring. In both CO2- and YAG-versions a train consists of 280 bunches with bunch-to-bunch spacing of 3.077 nsec, and the bunch population is 6.2E10. An electron linac which energy is 4.1 GeV (1.3 GeV) is employed to inject electron to the Compton ring. The linac is not necessary to be high current, because electron population loss due to Compton scattering is negligible. The collisions of electrons circulating in the Compton ring and photons stored in laser stacking cavities create polarized gamma rays. In one turn of the Compton ring, 280 (280x2) gamma ray bunches are created. Polarized positrons are created from those gamma rays on thin (~0.5 X0) target, then positrons in the high energy side of the spectrum are corrected in the capture section.

Each laser stacking cavity stored a photon bunch which energy is 210 mJ (600 mJ). The laser stacking cavity is designed to have enhancement factor of 100. Thus, each bunch of laser beam delivering from a laser should have energy of 2.1 mJ (6.0 mJ). The laser operates at 100 Hz and each pulse of laser contains 3.6E4 (2.9E4) bunches with bunch-to-bunch spacing of 3.077 n sec. The duration of the laser pulse is 110s (90s). Single laser provides a laser beam to 30 stacking cavities in daisy chain. The laser operates at 100Hz, thus laser cavities are filled by photons every 100ms.

Then positron injector linac accelerate positrons up to 5 GeV. A cold linac is employed. The linac is almost identical to the main linac, except that the injector operates at 100 Hz. Since the linac operates at 100 Hz, necessary cooling power per module is 4 times larger

v.Dec.12, 2005

Positron Source 5than that of main linac. However, the excess of total cooling power of the collider is not significant. After acceleration, positrons are sent to the damping ring. In the Compton scheme the damping ring has two functions: stacking and damping. Actually, main damping ring itself is the ideal choice of stacking ring, because it can store full number of positron bunches, it can be designed to have short damping time of ~10 m sec, and it has large longitudinal bucket area. The circumference of the damping ring is chosen to be 3247 m (2767 m), which is 5 times (10 times) larger than that of the Compton ring. The damping ring stores 10 trains with inter-train gap of 133 n sec (61 n sec).

The procedure of collision in the Compton ring and positron stacking in the damping ring is as follows. The collisions with laser photons make negligible loss of electron population in the Compton ring, however collisions make bunch lengthening. Due to bunch lengthening, number of gamma rays created by collisions decreases as a function of turn number. Therefore number of gamma rays becomes practically zero, if laser photons always exist in cavities. Pulse mode operation is applied to cure this problem. Laser cavities are filled by photons only in 110s (90s). This corresponds 50 (100) turns of the Compton ring. Then laser is turned off in ~9.9ms of cooling time. In the cooling time, electron bunch length become shorter and go back to primary length. Average number of gamma rays per turn is 1.8E10/bunch (1.4E10/bunch) in the energy range of 23-29 MeV. Then the average number of positrons is 2.4E8/bunch (1.9E8/bunch). 50 (100) turns of the Compton ring corresponds 10 turns (10 turns) of the damping ring. During this 110s (90s), injection to the damping ring continues and 10 times of stacking in each bucket are made in the damping ring. This stacking is performed using large longitudinal phase space of the damping ring. After the first injection, there is a ~9:9ms of damping time which corresponds the cooling time of the Compton ring. In a ~9:9ms, the area in the longitudinal phase space occupied by injected positrons is damped to the size which is small enough to accept the next injection. Since the laser operates at 100 Hz, next injection starts at this timing. This alternate cycle of injection and damp is repeated 10 times. Therefore, 100 times of stacking in each bucket are made in total. At this moment, the damping ring still has 100ms which is enough to make the emittance of positron beam fully damped. The simulation shows that the average stacking loss is 18%. Taking this into account, total number of stacked positrons in the damping ring is 2.0E10/bunch (1.6E10/bunch).

Compton Scheme R & D

There are several important components and technologies for the Compton scheme. Among of them, a couple of essential components were already demonstrated in KEK-ATF as follows,

High gain laser cavity[2],

The Compton scattering to yield enough gamma rays[3] with polarizaion[4], and to yield polarized positron[5],

v.Dec.12, 2005

Positron Source 6

that means the proof of principle of Compton scheme is reasonably confirmed. There are several issues to be studied for the Compton scheme as follows.

High current Compton Ring.

Positron beam stacking into the main DR.

High power CO2 or YAG laser system.

Control the chain of the laser cavities.

R&D Plans for these items are under scheduling as international collaboration. Regarding to High current Compton Ring.[6][7] and Positron beam stacking into the main DR., since we have developed of full beam tracking simulation code, research of the realistic values on beam parameters will be finished within one year. Then, parameters on laser systems and laser cavity will be redetermined.

On the other hand, we already started the design of the double chain of the YAG laser cavities and the installation will be planned at ATF. Test of laser focus to 5m will be established at the center point in the cavity in this JFY and we will install the double chain of the laser cavities into ATF damping ring next JFY and generate gamma rays for the demonstration.

Another challenge of our scheme is the necessity of high power and high quality laser. For the YAG-version, we have to develop long laser pulse amplification system using commercial available components. In the R&D program, both flash lamp excitation and solid state excitation are under consideration for this long pulse operation. Furthermore, the study to make the pulse stacking cavity which has higher enhancement factor[9], which reduce requirements to the laser. For the CO2, slicing of a long laser pulse into multiple bunches is the point. Essentially, the technology of slicing is already established[10]. However, in the Compton scheme, we need many bunches, thus heat load of the Ge-plate is not negligible. The engineering study to cure this problem is necessary.

References

[1] S. Araki et al., Conceputual Design of a Polarized Positron Source Based on Laser Compton Scattering. A Proposal Submitted to Snowmass 2005 CLIC Note 639, KEK Preprint 2005-60, LAL 05-94[2] M. Nomura et al., Proceedings of EPAC 2004, Lucerne, Switzerland., pp.2637-2639[3] I. Sakai et al., Physical Review Special Topics-Accelerators and Beams, 6, 091001(2003).[4] M. Fukuda et al., Phys. Rev. Lett. 91, 164801 (2003)[5] T. Omori et al., arXiv:hep-ex/0508026, KEK Preprint 2005-56

v.Dec.12, 2005

Positron Source 7[6] Eugene Bulyak and Vyacheslav Skomorokhov. Parameters of x.ray radiation emitted by Compton sources. In Proc. EPAC.2004 (Luzern, Switzerland), 2004. http://accelconf.web.cern.ch/accelconf/ e04/papers/weplt138.pdf.[7] Eugene Bulyak. Laser cooling of electron bunches in Compton storage rings. In Proc. EPAC.2004 (Luzern, Switzerland), 2004. http://accelconf.web.cern.ch/accelconf/ e04/ papers/thpkf063.pdf.[8] F. Zomer, Habilitation, Universit´e Paris 11-Orsay (2003) LAL/0312.[9] F. Zomer, Habilitation, Universit´e Paris 11-Orsay (2003) LAL/0312.[10] I. V. Pogorelsky et al., IEEE J. Quantum Electron., 31 (1995) 556. Alternative Configuration Document for ILC Positron Source Based on Laser Compton Scattering, August 2005

v.Dec.12, 2005

Damping Rings 1

4. Damping Rings

Updated: Aug. 04, 2006, CCB Approval Recommended: Sep. 19, 2006,

EC Approval: Sep. 21, 2006Initial Release: December 6, 2005

Contents

Introduction...........................................................................................................................2Circumference.......................................................................................................................2Beam Energy.........................................................................................................................2Injected Emittance and Energy Spread..................................................................................2Bunch Train Length and Bunch Charge................................................................................2Extracted Bunch Length........................................................................................................2Injection/Extraction Kicker Technology...............................................................................2Damping Wiggler Technology..............................................................................................2Main (Non-Wiggler) Magnets Technology...........................................................................2RF System Technology..........................................................................................................2RF Frequency........................................................................................................................2Vacuum Chamber Aperture...................................................................................................2Vacuum System Technologies...............................................................................................2Nominal Parameter and Performance Specifications............................................................2

Rev. April 23, 2006

Damping Rings 2

IntroductionThe initial recommendations for the configuration of the ILC damping rings were the result of discussions held during a meeting at CERN on November 9-11, 2005. The first part of the meeting was devoted to hearing the results of detailed studies of a range of configuration options. These studies were carried out over the previous six months by nearly 50 researchers, and the results of the studies form the basis on which recommendations for the damping rings configuration were made. A detailed report of the results of the configuration studies has been produced2. This report includes a description of the systematic process by which the various configuration options were evaluated, and choices between them made. Here, we simply present a summary of the present baseline and alternative configurations, together with a summary of R&D required in support of the configuration options. The studies of the various configuration options were based on nominal parameter and performance specifications for the damping rings: these specifications are given on page 10.

2 A. Wolski, J. Gao and S. Guiducci, “Configuration Studies and Recommendations for the ILC Damping Rings”, LBNL-59449, February 2006.

Rev. April 23, 2006

Damping Rings 3

CircumferenceThe positron damping ring should consist of one (roughly circular) ring of approximately 6 km circumference. Electron-cloud effects make a ring of circumference much less than 6 km unattractive, unless significant progress can be made with mitigation techniques. Design of the baseline positron ring (including the injection and extraction systems) should not preclude later installation of a second positron damping ring (located in the same tunnel) should electron cloud effects limit performance. Space-charge effects will be less problematic in a 6 km than in a 17 km ring, and achieving the required acceptance will be easier in a circular ring than in a dogbone ring.

The electron ring should also consist of a single 6 km ring, assuming that the fill pattern allows a sufficient gap for clearing ions. The injection and extraction kickers, as well as ion effects, are more difficult in a 3 km ring than in a 6 km ring. A 17 km ring could ease ion effects (by allowing larger gaps between minitrains), but would likely be higher cost. We have no recommendation on whether the electron ring needs a separate tunnel from the positron ring, but we note that having separate tunnels will ease installation and commissioning issues, make injection and extraction more straightforward, better “decouple” the operation of the positron and electron areas of the facility, and facilitate the installation of a second positron damping ring should it be needed later..

Although R&D is still required for the injection/extraction kickers for a damping ring with 6 km circumference, it is expected that existing programs will demonstrate a solution.

The exact circumference of the damping rings should be chosen, if possible, to allow flexibility in the fill patterns and number of bunches in a bunch train.

The feasibility of the baseline depends on: further progress with developing and demonstrating techniques for suppressing

electron cloud effects (positron ring); development of a satisfactory lattice design, e.g., with properties that mitigate ion

effects (electron ring) demonstration of kickers meeting the specifications for rise/fall times, kick

amplitude stability and repetition rate.

Alternatives1. If techniques are not found that are sufficiently effective at suppressing the electron

cloud, then a pair of (roughly 6 km) damping rings will be needed for the positrons; these rings could be located in the same tunnel. Design of the baseline configuration should not preclude later installation of a second positron ring as an upgrade.

Rev. April 23, 2006

Damping Rings 42. If electron cloud mitigation techniques sufficient for the baseline positron ring are not found, then a 17 km ring is a possible alternative; this would require addressing space-charge and acceptance issues.

R&D Requirements

Baseline

Techniques for mitigating the electron cloud to acceptable levels are needed. This R&D needs to be given very high priority.

A lattice design is needed that simultaneously satisfies requirements for acceptance and beam stability, and can be tuned easily for low emittance.

Alternative 1 (a pair of 6 km positron rings in the same tunnel)

Support and alignment issues for stacked rings in the same tunnel need to be addressed.

Designs are needed for the injection and extraction systems that will separate the injected beam between, and combine the damped beam from, the two positron rings.

Alternative 2 (17 km positron ring)

Techniques for suppressing space-charge tune shifts without driving betatron and synchrobetatron resonances are needed.

A lattice design is needed that simultaneously satisfies requirements for acceptance and beam stability, and can be tuned easily for low emittance.

General R&D requirements

Kickers that simultaneously meet specifications on rise/fall time, pulse rate and stability need to be demonstrated.

Assessment of mitigation techniques for ion instabilities are needed for the electron ring.

Ion-induced pressure instabilities in the positron ring need to be addressed. A range of classical collective instabilities must be properly understood, with

analysis based on a detailed impedance model. The effectiveness of low-emittance tuning techniques needs to be assessed.

Beam EnergyThe damping ring energy should be approximately 5 GeV. A lower energy increases the risks from collective effects; a higher energy makes it more difficult to tune for low emittance, and potentially has an adverse impact on the acceptance.

Rev. April 23, 2006

Damping Rings 5

Injected Emittance and Energy SpreadAn injected beam with maximum betatron amplitude up to 0.09 m-rad and energy spread up to 1% (full width) is preferred for the damping rings, over a distribution with larger energy spread but smaller betatron amplitude. Achieving good off-energy dynamics in the damping ring lattices is likely to be more problematic than achieving a large on-energy dynamic aperture. A smaller energy spread is likely to improve the margin for the acceptance of the injected beam.

AlternativeIf the acceptance issue (with realistic errors) can be addressed successfully, a larger energy spread on the injected beam (up to 2% full width) could be accommodated.

R&D Requirements

Baseline

Studies of the positron production indicate that an injected full-width energy spread of 1% should be achievable; however, a thorough investigation including realistic models for collimators, energy compressors etc. is still needed.

Alternative

A lattice design is needed that shows an energy acceptance with some margin beyond 2% full-width, while satisfying other requirements.

Bunch Train Length and Bunch ChargeA train length of around 2800 bunches is preferred because the kickers, ion effects and electron cloud are easier with a smaller number of bunches. If the electron ring is completely filled with no gaps (as may be the case with around 5600 bunches) the ion effects could be extremely difficult. However, there may well be other acceptable options with numbers of bunches between 2800 and 5600: further studies are needed to specify the gaps in the fill needed to keep ion effects under control.

If the positron ring is uniformly filled with 2800 bunches, the bunch separation is around 7 ns. Studies suggest that because of electron-cloud effects, the bunch separation should not be reduced much below this; this would prevent operation with larger numbers of bunches per train.

It is possible that the fill patterns in the electron and positron rings may need to be different, so as to allow a large bunch spacing between positron bunches (because of electron cloud), and gaps between minitrains of electron bunches (because of ions). This would require electron and positron rings with different circumferences, and would limit

Rev. April 23, 2006

Damping Rings 6flexibility on timing solutions.

AlternativesIncreasing the number of bunches beyond 2800 could be possible if electron-cloud and ion effects are found to be manageable, and sufficiently fast kickers can be demonstrated.

R&D RequirementsStudies are needed to determine:

the minimum bunch (or minitrain) spacing needed to keep electron-cloud effects under control;

the minimum gap between minitrains needed to keep ion effects under control. A demonstration is needed of kickers meeting the specifications (appropriate to each option for the number of bunches in a bunch train) for:

pulse rise and fall times; kick repetition rate; kick amplitude stability.

Extracted Bunch LengthA 9 mm bunch would be helpful for mitigating single-bunch collective effects in the damping rings (except, possibly, in the case of electron cloud), but a 6 mm bunch also appears to be a viable option.

R&D RequirementsStudies of bunch compressors suggest that a 9 mm bunch from the damping ring is acceptable, for a final bunch length of 300 m. Thorough studies, including tuning simulations for emittance preservation are in progress. Studies of beam dynamics effects in the damping rings with bunch lengths between 6 mm and 9 mm are needed to quantify the benefits (and drawbacks) of longer bunches.

Injection/Extraction Kicker TechnologyThe damping ring kickers should be based on “conventional” strip-line kickers driven by fast pulsers, without use of RF separators. The basic technology is available, and is close to a demonstration of most of the performance specifications. Using RF separators has potential cost implications, and could adversely affect the beam dynamics; for these reasons, it is preferred to avoid the need for RF separators if possible.

AlternativesRF separators may prove useful if it is decided to fill the rings with large numbers of bunches, pushing the bunch spacing to small values. Studies should be continued, to understand fully the beam dynamics and engineering issues.

Rev. April 23, 2006

Damping Rings 7Because Fourier pulse-compression kickers provide a very different approach, it is worthwhile continuing studies to develop a more complete understanding of the benefits and limitations of these systems.

R&D Requirements

Baseline

Kickers need to be demonstrated meeting all specifications for: pulse rise and fall times; pulse repetition rate; kick amplitude stability.

Alternative 1 (RF separators):

The beam dynamics and engineering issues associated with the RF separator scheme need to be fully understood, and limitations overcome.

Alternative 2 (Fourier pulse-compression kickers):

A more complete understanding is needed of the technical issues involved in Fourier pulse-compression kickers.

Off-Axis Injection

The usual operation mode of the damping rings requires on-axis injection, which prevents accumulation of current by stacking charge within RF buckets over many turns. Most conventional storage rings - e.g. in synchrotron light sources - use off-axis injection, in which radiation damping is used to merge injected (off-axis) charge with stored (on-axis) charge. The availability of off-axis injection would be of benefit in the damping rings for commissioning and tuning; a high beam current could be stored in the damping rings even with an injector system operating at less than full capacity, or with a separate, low-intensity source.

The possibility of designing the injection system of the damping rings to operate in either on-axis or off-axis mode should be investigated.

Damping Wiggler TechnologyThe damping wigglers should be based on superconducting technology. The requirements for field quality and aperture have been demonstrated in existing designs, and the power consumption is low.

AlternativesNormal-conducting electromagnetic and hybrid technologies are both viable alternatives. Issues with field quality and aperture can be addressed (at increased cost) in wigglers

Rev. April 23, 2006

Damping Rings 8based on either technology. The power consumption in a normal-conducting wiggler is a concern, though this technology could provide a device with potentially better resistance to radiation damage than the superconducting or hybrid options.

R&D Requirements

Baseline

The CESR-c wigglers have demonstrated the basic requirements for the ILC damping ring wigglers. Designs for a superconducting wiggler for the damping rings need to be optimized.

Alternatives

Designs with acceptable costs for normal-conducting electromagnetic and hybrid wigglers need to be developed that meet specifications for aperture and field quality. In the case of a normal-conducting electromagnetic wiggler, the design also needs to show acceptable power consumption.

Main (Non-Wiggler) Magnets TechnologyWe recommend that the main magnets in the damping rings be electromagnets. Using electromagnets simplifies tuning issues, and allows polarity reversal, e.g. for storing electrons in the positron ring.

AlternativePermanent magnets may still be considered as a possibility for the main magnets in the damping rings, if it is decided that polarity reversal is not required.

R&D Requirements

Baseline

Designs for electromagnetic dipoles, quadrupoles etc. should be straightforward, but still need to be developed and optimized.

Alternative

The problem of polarity reversal needs to be addressed. A demonstration is needed of a permanent magnet system with good tunability and resistance to radiation damage.

RF System TechnologyEach damping ring should use a superconducting RF system. Compared to a normal-conducting RF system, a superconducting RF system requires fewer cavities, (with

Rev. April 23, 2006

Damping Rings 9advantages for cost and keeping HOMs low); the power dissipation is lower; and smaller phase transients are expected.

AlternativeA normal-conducting RF system could still satisfy the requirements for the damping rings.

R&D RequirementsDesign and optimization of a 650 MHz superconducting RF system (including HOM-damped cavities) is needed.

RF FrequencyThe damping ring RF systems should use an RF frequency of 650 MHz. This is a simple subharmonic of the main linac RF frequency (1.3 GHz), which will facilitate phase synchronization between the damping rings and other RF systems. However, 650 MHz is a non-standard RF frequency that will require modest R&D.

R&D RequirementsDesign and optimization of a 650 MHz superconducting RF system (including HOM-damped cavities) is needed.

Vacuum Chamber ApertureA chamber diameter of (not significantly less than) 50 mm in the arcs, 46 mm in the wiggler and 100 mm in the straights is required. The wiggler chamber needs a large aperture to achieve the necessary acceptance, and to suppress electron cloud build-up. The large aperture also reduces resistive-wall growth rates, and eases the requirements on the feedback systems.

R&D RequirementsEven with a large aperture chamber in the damping rings, a bunch-by-bunch feedback system will be needed in the transverse and longitudinal planes to suppress coupled-bunch instabilities driven by the resistive-wall impedance. Although the required performance of the feedback systems should be within the range of existing technology, studies are needed of the level of residual beam jitter, and possible emittance growth.

Vacuum System TechnologiesRecommendations on the various options for the vacuum system technologies are yet to be made.

R&D RequirementsA number of issues regarding the vacuum system remain to be addressed, including:

What are the required levels of residual gas pressure needed to avoid ion effects?

Rev. April 23, 2006

Damping Rings 10 What kind of chamber preparation (NEG coating, TiN coating, grooves etc.) is

needed for suppressing the electron cloud, and what are the implications, e.g., for impedance?

Can (or should) clearing electrodes be used to suppress electron cloud or ion effects?

What length of time is allowed by the commissioning schedule for conditioning the vacuum system in the damping rings?

Further studies are needed to resolve these issues.

Nominal Parameter and Performance Specifications

Baseline Alternative (I)

Alternative (II)

Bunch train length 2820 5640Train repetition rate 5 HzInjected bunch separation 330 ns 165 nsMaximum injected normalized betatron amplitude (e+)1

0.09 m-rad

Injected full-width energy spread (e+) 1%Normalized injected transverse emittance, rms (e–)

45 m

Injected energy spread, rms (e–) 0.1%Injected bunch charge 21010 11010

Extracted bunch separation 330 ns 165 nsExtracted bunch charge 21010 11010

Extracted normalized horizontal emittance 8 mExtracted normalized vertical emittance 0.02 mExtracted rms energy spread 1.410–3

Extracted rms bunch length 6 mm 9 mmMaximum extracted vertical jitter 0.1

1 The normalized betatron amplitude is defined as Ax+Ay where:

and similarly for Ay. is the relativistic factor, and x, x, x are the Twiss parameters.

Rev. April 23, 2006

Ring to Main Linac 1

5. Ring to Main Linac

Overview

The ILC Ring to Main Linac (RTML) is the collection of beamlines which transfer the beam from the damping ring to the main linac on each side of the collider. In addition to transporting the beam between these two regions, the RTML beamlines perform all the manipulations in the beam conditions which are required to match the parameters of the beam extracted from the damping ring to the parameters required of the beam injected into the main linac. The parameter matching includes:

Collimation of beam halo generated in the damping ring Transformation of the beam polarization direction from the direction required in the

damping ring (nominally vertical) to the direction required by experimenters at the IP compensation of the beam jitter introduced in the damping ring, during extraction from

the damping ring, or during transport to the ends of the ILC site Compression of the bunch length from the equilibrium length of the damping ring to

the shorter length required in the linac and at the IP.

In addition, the RTML must provide instrumentation and diagnostics sufficient to measure and control emittance growth, beam jitter amplification, spin dilution, and other beam quality reductions which would otherwise be introduced by the beamlines of the RTML; and the RTML must provide a set of dumps and stoppers which can stop the beam from entering downstream systems during beam tuning at upstream locations or during maintenance accesses at downstream locations. Last but not least, the RTML must be made as cost effective as possible within the constraints of achieving its technical goals.

Baseline

A baseline configuration for the RTML has been selected.

Description

Below is a description of the RTML, divided according to sub-beamlines in longitudinal (”S-position”) order. Figure 1 shows an approximate topology of the layout of RTML and its beamline composision.

Rev. March 4, 2007

Ring to Main Linac 2

Figure 1: Approximate topology of the RTML layout and its beamline composision.

Skew Correction and Emittance Measurement

This beamline uses a set of four orthonormal skew quadrupole magnets to couple all four xy coupling terms in the beam matrix; this allows coupling introduced by the process of beam extraction from the damping ring to be corrected globally. The coupling correction section is followed by an emittance measurement station, which checks to make sure that the emittance correction is properly performed. The current baseline calls for a system which can measure only the projected horizontal and vertical emittances, and not the full beam matrix and the normal mode emittances and the coupling terms. The system for emittance measurement uses four laser wires driven by a common laser, with 45 degree phase advance between the wires in each plane. There is also a small chicane which separates the full-energy beam from the low-energy particles and photons produced by Compton scattering in the laser wire. At the end of this beamline there is a pulsed extraction system and a full-power (220 kW) dump which can be used to stop the beam from passing into the collimation section and other downstream areas. This stopper can be used when the beam extracted from the damping ring is being tuned up, when downstream systems are unprepared to receive beam, or duing Machine Protection System interrupts which occur within a single bunch train.

Transverse Collimation

This beamline uses a spoiler/absorber scheme to collimate halo particles which are generated in the damping ring. Collimation is 2 phases x 2 planes x 1 iteration, and both planes are nominally collimated at the same depth. The spoiler positions and apertures are adjustable; the absorbers may also need to be adjustable. The beam is sufficiently enlarged at the spoiler locations to prevent damage to the spoilers in the event of a direct hit from the beam core by a small number of bunches (requiring either that damping ring extraction

Rev. March 4, 2007

Ring to Main Linac 3is halted after that number of bunches); the spoilers, in turn, enlarge the beam core to prevent damage to the absorbers from a direct hit from several bunches of the beam.

Damping Ring Stretch

This is a simple coasting beamline, the purpose of which is to make up the distance between the damping ring extraction and the Escalator Beamline (see below); the damping rings are fixed to be at or near the center of the ILC site, while the Escalator’s longitudinal position is determined by the specific point at which the RTML beamline’s entry into the linac tunnel is desired. The total length and composition of the DR Stretch is to be determined.

Escalator

This beamline consists of a vertical arc, a coasting FODO lattice at a shallow angle, and an additional vertical arc. The purpose of the Escalator is to transport the damped beam from the beamlines listed above, which are at the elevation of the DR housing, to the beamlines listed below, which share the main linac tunnel. Depending on the exact layout adopted for the ILC, there may also be some horizontal bending sections in the Escalator; however, all bend magnets will be either pure-vertical or pure-horizontal bends. The Escalator section also contains 3 stoppers with burn-through monitors which are part of the Personnel Protection System (PPS), and are used to ensure the safety of workers in the main linac when beam is present in the Skew Correction section.

Return Line

This beamline is a coasting FODO lattice which runs anti-parallel to the main linac in the same tunnel. Its purpose is to transport the beam from the escalator to the extreme upstream end of the ILC site, where it is turned around and transported forward again to the IP.

Transverse Collimation

This is an additional collimation system, which is needed to collimate beam halo particles generated in the Damping Ring Stretch, Escalator, and Return beamlines. Unlike the Transverse Collimation section near the DR extraction point, the Transverse collimation system which follows the Return line cannot be protected by the simple expedient of stopping damping ring extraction since it is too far away from the damping rings and therefore too many bunches are already in the RTML before extraction can be switched off. The second Tranverse Collimation system relies on the general ILC MPS, which uses pilot bunches and polling of correctors and other devices to ensure the safe passage of the beam through the ILC. The second Collimation section also contains the beam position monitors which are used for trajectory correction via feed-forward across the Turnaround.

Turnaround

Rev. March 4, 2007

Ring to Main Linac 4This beamline reverses the direction of travel of the beam so that it is headed in the direction of the main linac. The purpose of the turnaround is to allow the bunch-by-bunch trajectory measurements to be fed forward over a shorter path length to the Trajectory Correction section. The turnaround will introduce emittance growth from synchrotron radiation:the baseline “cat-bone” style turnaround (with a 90 degree bend followed by a 262 degree reverse bend) limits emittance growth to about about 9% of the emittance at extraction from the damping ring. The turnaround also contains spoilers and absorbers for collimation of off-energy particles generated in the Return line or via multiple coulomb scattering from the collimators in the second collimation section.

Spin Rotator

This beamline uses 4 strong solenoid magnets to allow the beam polarization vector to be set to any orientation desired by the experimenters. The first half of the system contains two solenoids which are powered in series and separated by an Emma rotator (a beamline which performs a +I transformation in the horizontal plane and a -I in the vertical), to allow the polarization to be adjusted without introducing coupling from the solenoids. This is followed by an achromatic arc of approximately 8 degrees, which completes the turnaround of the beam trajectory; the arc is followed by another pair of solenoids separated by an Emma reflector. The combination of the two solenoid pairs and the bending system allows the polarization to be pointed in any direction required by the experimenters.

Trajectory Correction

This beamline is a simple FODO array with 2 horizontal intra-train dipole correctors separated by 90 degrees in betatron phase, and 2 vertical intra-train correctors with the same phase separation. Bunch-by-bunch trajectory information is measured in the upstream collimation h section, and fed forward to this location to correct the beam jitter generated in the damping ring and during extraction (i.e., by jitter in the extraction kicker amplitude or driven by quad vibrations in the Return line).

Emittance Measurement and Coupling Correction

This beamline uses a set of 4 laser wire profile monitors to make sure that the emittance correction is properly performed. The current baseline calls for a system which can measure only the projected vertical and horizontal emittances, but cannot measure the full beam matrix, the normal mode emittances, or the coupling terms. The emittance measurement station must be capable of measuring emittances during multibunch operation, using many bunches to measure the emittance within 1 train, and must also be capable of measuring emittances during single-bunch operation, using many pulses (at 5 Hz) to complete one measurement. This section also contains a system with 4 skew quads, similar to the Skew Correction section described above, immediately upstream of the

Rev. March 4, 2007

Ring to Main Linac 5emittance measurement station. This is necessary due to the large number of betatron wavelengths between the initial Skew Correction Section and the Emittance station.

First Stage Bunch Compressor

The first stage bunch compressor is divided into the following subregions:

An RF section which generates the necessary correlation between longitudinal position and energy. This section contains 24 9-cell RF cavities arranged in 3 cryomodules of 8 cavities each, based on the assumption that this is the cryomodule configuration which will be used for the ILC main linac. Because the bunch is long in this section, relatively strong focusing is used to limit the emittance growth from transverse wakefields: quad spacing is 1 quad per cryomodule, with 90 degree phase advance per cell in x and y. The cavities are phased near the zero-crossing (-100 degrees is typical), and require gradients of up to 18.4 MV/m. There are no spare modules in this section, but there is a spare klystron and modulator which can be connected to the cryomodules via an RF switch in the event that the BC1 klystron or modulator should fail.

A wiggler based on 6 90 degree FODO cells with chicanes placed in the space between each pair of quads (12 chicanes total). Each chicane contains 8 bend magnets, with bends 1 and 8 in each chicane powered in series by one power supply and bends 2, 3, 4, 5, 6, and 7 each comprising an additional string of magnets (ie. The wiggler proper has a total of 7 bend strings). There are also 3 bends at the upstream end of the wiggler and 3 at the downstream end for dispersion matching; these 6 bends can all be powered in series from a single power supply. The wiggler also contains normal and skew quads for tuning the horizontal and vertical dispersion, instrumentation for measuring beam energy, and adjustable energy collimators which can tolerate being struck by several bunches without being damaged.

A longitudinal diagnostics section, which permits measurement of the central energy, energy spread, arrival time, and bunch length. This section includes an approximately 35 cm-long dipole-mode RF structure (“crab cavity”) which is room-temperature and operates at S-band. The structure is used as part of the bunch length diagnostic system.

A short region which extracts the beam from the straight-ahead channel to a tune-up dump. This region is equipped with pulsed bends (which can driven by a DC power supply or else pulsed to take bunch trains out to the tune-up dump) and also with a set of kicker magnets (which can rise from zero to full strength in 100 nsec or less, which permits a subset of the train to be extracted).

Second Stage Bunch Compressor

Rev. March 4, 2007

Ring to Main Linac 6The second stage bunch compressor is divided into the following subregions:

An RF section that generates the necessary correlation between longitudinal position and energy. This section contains 384 9-cell RF cavities arranged in 48 cryomodules of 8 cavities each. There is 1 quad per 2 cryomodules and a phase advance of 60 degrees is used in each plane. Of the 48 cryomodules, 3 are spare and 45 must be accelerating the beam. The phase of the RF is between -22 degrees and -58 degrees depending on the exact configuration, and the maximum gradient required in the accelerating sections is 31.1 MV/m.

A wiggler with optics identical to the wiggler in the first-stage bunch compressor, but with weaker bends. Also, the bends used for dispersion matching are longer than in the BC1 wiggler.

A longitudinal diagnostics section, which permits measurement of the central energy, energy spread, arrival time, and bunch length. This section has one dipole mode structure which is identical to the one in BC1.

Launch into Main Linac

The launch into the main linac performs the final conditioning of the beam necessary for main linac injection. This region includes a station for the measurement of projected emittances. There is also an extraction line which permits DC extraction, train-by-train extraction (via pulsed bend magnets), or sub-train extraction (via kicker magnets with 100 nsec or less rise time), similar to the system at the end of BC1.

.

Rev. March 4, 2007

Ring to Main Linac 7

Parameter Tables

Beam ParametersParameter Nominal Value HighLumi Value LongDRBunch Value

init momentum 5 GeV/c init espread 0.15% init emit 8 m 20 nm init x jitter 1.0 ? init bunch length 6 mm 6 mm 9 mm final bunch length 0.3 mm 0.15 mm 0.3 mm final momentum 15.0 GeV/c 13.0 GeV/c 15.0 GeV/c final espread 1.1% 2.5% 1.6% final x jitter 0.1 ? ISR emit growth 0.90 m 0.74 m 0.90 m emit growth budget 1.0 m 4.0 nm?

All system lengths and element counts are approximate pending development of lattice files describing the RTML baseline configuration.

Magnet CountsRegion Bends Quads Sextupoles Dipoles Kickers Solenoids Septa

Skew Correction 1 0 16 0 18 0 0 0Emittance 1 5 8 0 6 8 0 4Collimation 1 0 12 0 18 0 0 0DR Stretch 0 36 0 54 0 0 0Escalator 16 34 0 51 0 0 0Return 0 336 0 504 0 0 0Collimation 2 0 12 0 18 0 0 0Turnaround 116 66 0 99 0 0 0Spin Rotator 6 45 0 67 0 4 0Skew Correction 2 0 16 0 18 0 0 0Emittance 2 0 15 0 22 4 0 0BC1 103 29 0 27 8 0 0BC1 Extraction 8 7 10 2 4BC2 102 55 0 28 0 0 0BC2 Extraction 10 9 13 2 6Linac Launch 5 15 0 22 10 0 0Total 371 711 0 975 34 4 14

Rev. March 4, 2007

Ring to Main Linac 8

Rev. March 4, 2007

Ring to Main Linac 9

1.3GHz RF Components in Bunch CompressorsRegion Cavities Modules Klystrons BC1 24 3 1+1 BC2 384 48 16 Total 408 51 17+1

NC S-Band Dipole-Mode RF components (for diagnostics)

Region Structures KlystronsBC1 1 1Linac Launch 1 1Total 2 TBD

Instrument Counts

Region BPMs Wires BLMs OTR

ScreensPhase

Monitors Skew Correction 1 12 0 0 0 0Emittance 1 8 4 0 0 0Collimation 1 12 0 0 0 1DR Stretch 36 0 0 0 0Escalator 34 0 0 0 0Return 336 0 0 0 0Collimation 2 12 0 0 0 0Turnaround 60 0 0 0 0Spin Rotator 45 0 0 0 0Emittance 2 15 4 0 0 0BC1 27 0 1 1 1BC1 Extract 6 0 0 1 0BC2 52 0 0 0 0BC2 Extract 9 0 0 1 0Linac Launch 15 4 1 2 1Total 679 12 2 5 3

Rev. March 4, 2007

Ring to Main Linac 10System Lengths

Region Length Skew Correction 1 27 mEmittance 1 27 mCollimation 1 400 m DR Stretch 600 mEscalator 600 mReturn 13200 mCollimation 2 400 mTurnaround 218 m Spin Rotator 82 m Emittance 2 27 m BC1 238 m BC1 Ext 60 mBC2 758 m BC2 Ext 63 mLinac Launch 89 m Total 16789 m Total excluding Extraction lines 16666 mLength of Footprint 1263 m

Rev. March 4, 2007

Ring to Main Linac 11

Justification

The emittance measurement station immediately downstream of extraction is required because of the well-known sensitivity of the extracted beam emittance to beam position in the damping ring septum magnet. Only some form of emittance measurement system immediately following the extraction can be used to determine the optimum extraction orbit for emittance preservation. Since the beam is stopped just after the emittance section during access to the linac or other parts of the ILC, it is necessary to have an emittance diagnostic at this location which can operate during full power extraction from the ring and which is non-invasive, thus the complement of laser wires. The skew quads upstream of the emittance station are used to correct coupling from the DR extraction septum and are tuned by minimizing the projected emittance measured on the subsequent laser wire scanners as a function of skew quad settings.

During SLC operation, calculations of likely beam halo populations due to linac scattering processes did not explain the large observed beam halo, which was removed by collimators at the high-energy end of the SLAC linac. Because of the end-linac ILC beam parameters (energy, power, and emittance), it will be quite difficult to collimate the beam halo at the high-energy end of the linac. The risk of intense halo formation in the damping ring is mitigated with a relatively simple collimation section in the RTML, where the energy and beam power are relatively low and the geometric emittance relatively large. Because of the large scattering potential in the DR stretch, escalator, and return lines, an additional simple collimation system is required immediately upstream of the turnaround.

The intra-train jitter requirements for the extraction kicker are extremely tight (0.07% RMS), and represent a luminosity risk for the ILC. In addition, the tight vertical beam jitter requirement imposed at the IP by the strong disruption (0.05 sigmay) also represents a luminosity risk. The jitter measurement and feedforward permit that risk to be mitigated. The turnaround is required to delay the arrival of the beam sufficiently for the beam jitter data to be processed and the jitter correction to be applied to the magnets. It also permits the beam to be turned around from the “outbound” direction to the “inbound” direction, which is required to accommodate the central Damping Ring complex.

The polarization of the beam is rotated into the vertical to preserve it during storage in the damping ring. The polarization at the IP has to be completely adjustable and tunable, and the adjustment/tuning is not permitted to dilute the emittance. The most straightforward method identified is to use solenoidal spin rotators with the lattice properties described above to cancel out the emittance growth from xy coupling that the solenoids would otherwise generate. Since the solenoids rotate the polarization from vertical to horizontal, the 8 degree arc between the two solenoid pairs is required to rotate the polarization from horizontal to longitudinal; thus the first solenoid pair plus the arc puts the polarization in the longitudinal; if the first solenoid pair is turned off and the second pair is turned on, horizontal polarization is generated; if both solenoid pairs are turned off, the polarization remains oriented in the vertical.

Rev. March 4, 2007

Ring to Main Linac 12

Since the coupling correction of the paired solenoids is not likely to be perfect, the coupling correction and diagnostic lattices are required to globally correct any residual coupling. In principle a single set of skew quads might be acceptable to correct the coupling errors from the DR septum, the spin rotators, and all the coupling errors in between; in practice this is a very large number of betatron wavelengths, and recent studies have shown that positioning the skew quads immediately upstream of the wire scanners used to tune them (as opposed to 20-40 betatron wavelengths away) dramatically improves the convergence of coupling correction. For this reason we have elected to include two skew correction sections.

Bunch compression in the ILC is a necessity, given the opposing requirements of the damping ring (where long bunches are needed to limit collective effects) and the IP (where short bunches are needed to match the small values of betay which are mandated by the high luminosity goals). This compression is complicated by the large longitudinal emittance generated by the damping rings, which means that bunch compression leads to large energy spread after compression. Because of the energy spread, a single stage for compression from 6 mm to 0.3 mm RMS length was already marginal, and shorter bunches such as 0.15 mm RMS, which are required in the parameter tables, are not achievable in a single stage. The two-stage system works around this by accelerating the beam between stages of compression to limit the maximum fractional energy spread at any point in the ILC. The large and complex configuration of the wigglers is driven by the requirements of flexibility in the initial and final bunch lengths and by the requirements of dispersion tuning quadrupoles which do not introduce betatron mismatches or x-y coupling.

The final emittance measurement station is required to tune the emittance of the large energy spread beam generated by the compressor, prior to injection into the main linac. The collimation in the main linac launch is primarily machine protection segmentation: it ensures that a mistuned or mis-steered beam in the RTML will not result in a machine protection incident in the main linac. This segmentation is vital to simplify design of the active MPS.

Required R&D

The following R & D steps are required in order to produce a complete design for which the cost can be estimated:

Tuning and tolerance studies (analytic and simulation). Since tuning studies are often fairly sensitive to the details of the simulation, it is necessary that this step be taken by at least two different people/groups working semi-independently. The exact effort required for this study is hard to estimate as nobody has yet made a serious attempt at studying the combined transverse and longitudinal tuning of any RTML lattice for any linear collider design. Since the RTML tuning is in a much less mature state than the main linac, unpleasant surprises are a possibility that can’t be ruled out.

Development of component and system tolerance specifications. Review of component and system tolerance specifications by qualified engineers.

Depending on the outcome of this step, further iterations of the design, tuning, and tolerancing studies listed above may be required.

Rev. March 4, 2007

Ring to Main Linac 13

It is worthwhile to note that, compared to the other regions of the ILC, the RTML pushes the limits of technology in very few places. The electromagnets can easily be designed to fall within the limits of existing accelerator magnet technology; the RF components are duplicates of the main linac versions, and in fact are generally down-rated in their power and gradient requirements; the pulsed kickers for extraction of a runaway beam are based on similar technology to the system at the end of the main linac, and benefit from the much lower beam energy compared to the latter system; the bunch length monitors can be based upon very successful systems in use elsewhere for the measurement of much shorter bunches. The main technological issue for the RTML is likely to be the required RF system phase stability, which is on the order of 0.1 to 0.2 degrees of L-band. This phase stability must be maintained for a period which is long enough for a beam-based feedback to determine that an unacceptable phase change has occurred, as indicated by variation in the beam arrival times at the IP; thus, a stability period of a few seconds is probably sufficient.

Lattice Files

The lattice files for the baseline configuration will be made available at the ILC RTML wiki page as soon as they are prepared: http://www.linearcollider.org/wiki/doku.php?id=rdr:rdr_as:rtml_lattice .

Alternatives

Description

Single Stage Bunch Compressor

Another possible cost savings would be to return to a single-stage bunch compressor similar to what was included in the TESLA design. Note that although the site length table implies a site savings of almost 1 km per side, this is partially compensated by the fact that the beam energy at the end of the RTML is reduced from around 15 GeV to 4.4 GeV, and thus the linac must be lengthened in this option; as a result the actual net savings is about 0.5 km per side, plus all 100 of the BC2 bends and about 30% of the BC2 quads and RF elements. Such an option can only be pursued if the emittance tuning strategies for the RTML and main linac can be shown to function reliably in the presence of an RMS energy spread in excess of 4%, and if the parameter sets assuming a longer bunch in the damping ring or a shorter bunch at the IP are eliminated. Because this change to the design would reduce the parameter reach of the ILC, the Low Emittance Transport working group at Snowmass 2005 does not favor pursuing this alternative.

Shorter Two Stage Bunch Compressor

At the Snowmass workshop of August 2005, a design for a two stage bunch compressor with short single chicanes in place of the wigglers used in the baseline. This design has compression capabilities comparable to the baseline design (ie, both the nominal 300 micrometer RMS bunch length and the shorter 150 micrometer RMS length can be achieved),

Rev. March 4, 2007

Ring to Main Linac 14

but uses a single chicane for each stage of compression, rather than the 12 chicanes used in the baseline design. This design eliminates about 190 bends per side from the total BC1/BC2 system, along with about 25 quads per side. The site-length savings is about 400 m per side relative to the baseline. Since this design eliminates the emittance-tuning features of the wiggler in the baseline, and eliminates the symmetries of design which make those features possible, there is a risk of unacceptable emittance dilution which must be studied. In particular, the tuning strategy and installation tolerances of the shorter system must be carefully reviewed. Because the shorter two-stage bunch compressor design has not been studied to the same degree as the longer design, it is the longer design which has been selected as the baseline. However, it is highly recommended that this alternative be studied, since a design does exist and since, if this design proves tractable from the point of view of emittance tuning, it would permit a significant cost savings without sacrificing performance or parameter reach.

Lattice files for the shorter two-stage bunch compressor can be found here.

Required R&D

If the RTML is to be built with a single-stage bunch compressor, more intense studies of emittance preservation in the presence of a larger RMS energy spread will be required. In addition, it will be necessary to verify that the damping ring can achieve its required stability stability with short (6 mm RMS) bunches, and to verify that IP conditions will be tolerable with long (0.3 mm RMS) bunches.

If the RTML is to be built with the shorter two-stage compressor, more intense studies of emittance preservation will be required, concentrating on the absence of dedicated dispersion tuning quadrupoles.

.

Rev. March 4, 2007

Main Linac 1

6. Main Linac

Baseline: Modulator

To power the 10 MW MBK’s, the baseline choice is to use the bouncer-compensated pulse transformer modulator that was developed initially by FNAL and industrialized by DESY (with seven units built). Although the units perform well, they are expensive, require multi-ton, oil-filled transformers, and are susceptible to single-point failures.

FNAL recently designed a lower-cost, more-reliable pulse-transformer modulator that has 5 ms pulse capability (desired by the Proton Driver project). Two units are being built and will be commissioned in 2006. They use the latest generation of high power transistors for the switches, which SLAC has designed and will provide for these units. In addition, DESY continues to work with industry to improve the reliability of their modulator design in preparation for the XFEL, which requires 35 units (these units will likely be run at about 80% of the peak power level required for ILC).

Alternatives

The main alternative is a Marx-style generator, which is being developed at SLAC and through the DOE-funded Small Business Innovative Research (SBIR) Program. It has a modular design with built-in redundancy and should be easier to mass produce and repair compared with the baseline modulator (in part, because it has no transformer and is forced-air cooled). Preliminary cost estimates suggest that it may be up to 50% less expensive than the baseline choice. The first full-scale prototype is expected to be tested near the end of 2006. Other approaches being considered include a direct-switch (DTI will build a prototype through SBIR funding) and a DC-to-DC converter (LANL is simulating operation of a higher power version of their SNS design).

For these alternatives, if a TDR-like tunnel layout were adopted with the modulator separated from the klystron by up to 2.5 km, the transport and impedance matching (cable to klystron) of the 120 kV pulses would require further development. The Marx approach would be the easiest to operate with an impedance mismatch as its turn-on can be ramped step-wise.

Baseline: Klystron

The 10 MW Multi-Beam Klystrons (MBK’s) being developed by Thales, CPI and Toshiba are the baseline choice. The basic tube design appears to be robust while alternative approaches have not been fully designed nor are currently funded to be developed. At worst, if the MBK’s do not meet availability requirements, the commercial, single-beam, 5 MW tube from Thales could be used (it has been the ‘work-horse’ for L-band testing at DESY and FNAL). Although it is less efficient (42% vs 60-65%), this tube has been in service for over 30 years with good availability.

Rev. July 13, 2006

Main Linac 2

Alternatives

The three alternatives discussed were a 10 MW Sheet-Beam Klystron (proposed by SLAC to reduce cost), a 5 MW Inductive Output Tube (proposed by CPI to improve efficiency) and a 10 MW, 12 beam MBK (proposed by KEK to reduce the modulator voltage, and the modulator plus klystron cost).

Cavity Shapes

Overview

The cavity shape determines the fundamental mode as well as the higher order modes properties. The aperture of cavity cells determines the cell-to-cell coupling in the fundamental mode, the loss factor of wakefields, and the higher order mode propagation. There are several options under consideration for ILC BCD and ACD. These options differ in terms of the following cavity parameters,

The ratio of the peak magnetic field to the accelerating gradient (Hpk/Eacc). The ratio of the peak electric field to the accelerating gradient (Epk/Eacc). The product of the geometry factor G and R/Q (G x R/Q). The cell-to-cell coupling factor (kc). The loss factors of longitudinal (k_l) and transverse (k_t) wakefields. The Lorentz detuning factor (K_L).

The choice of a specific shape has profound impact on the cavity performance, beam quality, beam stability and manufacturability. The mature TESLA shape has a favorable low Epk/Eacc, acceptable cell-to-cell coupling and wakefield loss factors. It has lower risk of field emission and dark current.

Two major new shapes, the Cornell re-entrant shape and the DESY/KEK low-loss shape, are under initial developments. Both new shapes have a lower Hpk/Eacc and a higher G x R/Q. They have a higher ultimate gradient reach since Hpk is the fundamental limit, and lower cryogenic losses. Both shapes carry higher risk of field emission and dark current since Epk/Eacc are 20% higher than the TESLA shape. The iris aperture is a major geometrical difference between the two new shapes. The DESY/KEK low-loss shape has a smaller iris aperture by about 15%, whereas the Cornell re-entrant shape has the same aperture as that of the TESLA shape.

The ‘superstructure’ concept improves the cavity packing fraction in the cryomodule by a few per cent and reduces the number of input couplers by a factor of two. The idea can be applied to any of the cavity shapes besides TESLA.

Baseline: TESLA Shape

Description

Rev. July 13, 2006

Main Linac 3

The TESLA cavity is the benchmark that all other designs must be compared to.

Pros: It has low Epk/Eacc, large cell-to-cell coupling and small wakefields. It has been studied and tested extensively:

o Single-cell cavities achieved up to 43 MV/mo It has achieved ~ 35MV/m gradient at Q = 10^10 for a number of multi-cell

cavities, best achieved 40 MV/m at 10^10.o High power tests were done as well as a test of one cavity in a module with

beam. o Over 100 cavities have been constructed. Several companies have fabricated

successful cavities showing that the procedures are well understood. Wake fields and HOM damping have been thoroughly investigated. Cavities and modules have operated in TTF for considerable time. Cavity Data base, processing and test history is extensive. The cost basis for limited quantity is well established and a number of vendors have

produced these cavities. Industrialization studies of fabrication and processing have been carried out. These cavities are planned for the XFEL. These cavities are closest to meeting ILC requirements at this time.

Pro & Con:

35MV/m is close to the ultimate gradient (42MV/m ?) for this cavity shape. This points to the maturity of the R&D program.

Cons: This shape has a higher Hpk/Eacc (than the alternatives under development) and a

higher risk of premature quench induced by a higher surface magnetic field for gradients > 40 MV/m.

These cavities do not have the ultimate gradient potential that some of the other ACD designs have. However other designs reduce Hpk/Eacc at the expense of Epk and/or iris diameter. These new shape cavities have not yet been proven as a module operating at 35MV/m with a beam.

Required R&D

Most critical R&D is the establishment of proof of principle of 35MV/m modules with acceptable dark currents, and long-term operation at 35 MV/m with acceptable trip rates.

Topics relevant to cavities listed under R1 to R4 in the TRC should be addressed with existing or planned test facilities.

Further work is necessary to establish an adequate gradient safety margin and to get reproducible high gradient results from cavity to cavity, and from processing to processing.

Rev. July 13, 2006

Main Linac 4

There needs to be a measurement cryogenic power deposited by HOMs to be sure that this is less than 20% as required to keep the overall cryo load under control.

Potential Mods to BCD

A number of minor modifications and improvements could be implemented without impact to the basic cavity design. These include:

Slight modifications to the HOM coupler, and pickup design for ease of fabrication, fundamental power rejection, and thermal stability.

Design modification to the helium vessel end walls for more strength. Shortening the beam tube lengths to their acceptable minimum to improve the packing

fraction. Review of the overall mechanical design, including flanges, sealing gaskets, number

of nuts and bolts, and end group fabrication, with an eye toward industrial production.

Technical advantages, increased tech potential:

Savings in cavity length (and interconnect) will shorten the tunnel required. HOMs would have better power margin.

Potential cost impacts:

If a 5% cavity slot length reduction could be realized, this would impact the tunnel length and cost (but probably less than or ~ 1% of total cost.) Greatest cost impact is probably in the design for industrial production if good ideas emerge.

Risk and Reliability impacts:

Better design with more margins should decrease risk and improve reliability. This is especially true if a reliable and simple flange design (or weld) can be developed.

Cost Estimation

Lattice Files

Parameter Tables

Supporting Documentation

Original Snowmass document by R. Geng

TESLA TDR Cavity Chapter

P. Schmueser et al.; The Superconducting TESLA Cavities; PRST-AB 3 (9) 092001

Rev. July 13, 2006

Main Linac 5

Alternatives1. A number of different cavity shapes are being proposed. These shapes tend to

decrease Hpk/Eacc (Pro) and increase G x R/Q (Pro), but increase Epk/Eacc and may have smaller iris diameters (Cons). The most work to date has been done on the DESY/KEK low-loss.

2. The ’superstructure’ concept improves the packing fraction and reduces the number of couplers. The idea can be applied to any of the cavity shapes.

Alternative cavity Shapes

Technical advantages, increased tech potential

Reduced risk of premature quench due to lower Hpk/Eacc for gradient > 35 MV/m. Higher gradient possibly up to 45-50 MV/m. Has higher G x R/Q so lower cryogenic power loss. Needs shorter tunnels. Gradient improvement could be used for operating margin.

Risk and Reliability impacts

Has higher Epk/Eacc. Dark current (exponential with Epk) may be a greater problem. Operating at higher gradient implies greater reliability issues, and greater risk,

especially during commissioning and early operation.

DESY/KEK low-loss shape

Pros: Most work done to date, several single cells tested, some multicells fabricated. Successful tests of 45-47 MV/m with 1.3 GHz single cell cavities

o (45 MV/m with 2.2 GHz single cell cavity). Computational analysis of wakefields underway. Test of 9 cell cavity underway. Lorentz force detuning analysis underway.

Cons: Has smaller iris than TESLA. May have less mechanical strength. Manufacturabillity (danger of making it re-entrant) Needs much development and testing to reach maturity of TESLA.

Cornell Re-entrant shape

Pros: Has expectation of higher gradient with TESLA like iris diameter.

Rev. July 13, 2006

Main Linac 6

Single cells at 1.3 GHz tested up to 52 MV/m. Successful HPR with single cell done to give record Epk (> 100 MV/m). First order HOM analysis of multicells complete. Lorentz force detuning analysis underway. 9 cell cavity fabrication underway.

Cons: Weaker mechanical strength. HPR is problematic because of re-entrant design. This is a critical complication, as the

Epeak/Eacc ratio is also enhanced. Needs much development and testing to reach maturity of TESLA.

Modifications/ variants of re-entrant: smaller aperture re-entrant half-re-entrant.

Required R&D

Considerable R&D will be required and different check points:

Wake fields:o The allowed iris diameter must be specified from theoretical analysis. This is a

trade off between allowable emittance growth (luminosity) and cost.o Complete wake field analysis must be carried out computationally and

checked with measurements.o Cold tests of wake fields must be carried out on two or more adjacent cavities.o Wake fields must be checked in modules with beam.

Gradient and Q:o Gradient and Q expectations up to at least 35MV/m must be achieved first in

9-cell cavity tests then in modules with beam. Time scales for R&D

o Full program to bring one of these cavity ideas to the state of understanding of the TESLA cavity may be of order several years with substantial funding.

o The rules for when the ACD would be considered to replace the present BCD should be proposed. Such a point might be when ~6 cavities have achieved gradients in excess of 35MV/m with Q >10^10, and when HOM damping has been checked in at least a two-cavity (9-cells each) string without beam.

Superstructure

Pros:

Superstructure has possibilities for significant cost savings through the use of only one input coupler per two cavities.

Significant design work has been carried out.

Rev. July 13, 2006

Main Linac 7

A two superstructure module (with two pairs of 7-cell structures) has been tested with beam at DESY.

Wake fields have been investigated and the mode analysis understood.

Cons:

A main drawback of the superstructure is how to process and test such a long assembly, either with BCP or EP processing. This would take significant infrastructure development beyond that needed for single cavity structures.

Alternatively a superconducting joint might be developed to join the superstructure pair after processing. This has been attempted recently at DESY without success. Jlab has a program to continue superconducting joint work.

Technical advantages, increased tech potential

The main technical advantage would be the reduction in the number of input couplers by a factor of two (elimination of 8000 couplers). These couplers would need to carry double power.

Wake fields are less

Potential cost impacts

The cost saving might be the cost of 1/2 the couplers. However if coupler fabrication cost is reduced significantly then the impact would be less.

Supporting Documentation

Original Snowmass document by R. Geng

Superconducting superstructure for the TESLA collider: A concept, J. Sekutowicz et al., Phys. Rev. ST Accel. Beams, 2, 062001 (1999)

Test of two Nb superstructure prototypes, J. Sekutowicz et al., Phys. Rev. ST Accel. Beams 7, 012002 (2004)

Cavity Materials

Overview

Specifications for high purity niobium (RRR) sheet, used for fabrication of cavity cells and auxiliary components such as HOM- and FP-couplers and beam pipes have been developed over the years. Material produced to these specifications by various companies in Japan (Tokyo Denkai), Germany (W.C.Heraeus) and the US (Wah Chang) and used for cavity fabrication has resulted in high performance prototype and production cavities, when combined with appropriate QA measures during sheet production such as e.g. clean rollers or eddy current or squid scanning for defects prior to deep drawing of cavity cells.

Rev. July 13, 2006

Main Linac 8

Recently, single cell and multi-cell cavities have been produced at JLab from large grain ingot material or from single crystal, cut directly from the billet by either wire EDM or saw cutting. This is an exciting new development, and has the potential of simplifying the production sequence and consequently the cost. Initial experience indicates that very smooth surfaces can be obtained with the single crystal material or even the large grain material using the BCP (chemical) etch process only, thus avoiding the necessity for using the more complex electro-polishing (EP) processing. This might be related to less defects, a reduced intrinsic strain in the single crystal material and a significantly reduced number of grain boundaries.

Nb/Cu laminated material has been successfully used to produce high gradient single cell cavities at DESY and KEK from Nb/Cu tubes by hydro-forming. Explosion bonding, HIP bonding, back-extrusion and hot rolling techniques were successfully used to produce the composite tubes. The laminated Nb/Cu approach takes advantages of the bulk Nb performance (Nb layer ~ 0.5 mm thick) combined with the increased thermal conductivity and stiffness of the copper backing resulting in possible significant cost savings. Welding presents a difficulty in that the Cu must be removed at the weld joints before e-beam welding and that there are risks of contamination/leaky joints. This material is probably best suited when used with hydro-formed multi cell assemblies.

Low frequency (< 500 MHz), lower gradient cavities historically use a thin layer of Nb deposited on Cu. Cavities made from deposited Nb suffer from very strong Q slope and do not appear suitable for high gradient (>15 MV/m), high Q ILC application. Research continues with different deposition techniques (e.g. vacuum- arc and ecr plasma-deposition) to try to understand and improve the SRF properties.

Other types of superconductors, such as Nb3Sn, NbN and MgB2 are experimental and material property evaluation stage, and far from being useful for project application. There are also fundamental questions related to the limiting RF field and its dependence on kappa in these high kappa materials.

Options under consideration Nb RRR fine grain sheet Nb single crystal or large grain Nb/Cu lamination Nb deposited on Cu Other superconductor

Please note that the second part of the original Snowmass document by P. Kneisel contains some general R&D topics which are not solely related to one of the technologies described here. Please see also section Fundamental Material R&D.

Rev. July 13, 2006

Main Linac 9

Baseline

Nb RRR fine grain sheet

Description

Pro: Specifications exist and are - with some exceptions - met by industry. This material is best known. Measurements of the thermal conductivity, Kapitza

resistance (under different surface conditions), mechanical properties and mechanical anisotropy, texture and formability have been published. Post purification with Ti at different temperatures and durations has been studied.

Many examples of high performance cavities (single+multi-cell) made from this material exist.

Studies of possible cost-savings in a mass-production scheme were performed.

Con: For accelerating gradients > 28 MV/m EP for final surface treatment is necessary to

give a smooth surface finish, even though there exist examples of cavity performances beyond this level after BCP treatment (rougher surface) only. In any case, ‘in situ’ baking is necessary to remove the ‘Q-drop’, typically starting at gradients > 22MV/m, corresponding to peak magnetic surface fields > 80 mT (G.Ciovati,SRF2005). In-situ baking seems to be more effective in electropolished cavities to remove the Q-slope at high fields.

At this time it is statistically unclear, if titanization at 1200°C-1400°C is required for best performance of this material. The titanization with subsequent etching increases the RRR from ~300 to ~600, providing better thermal stability of the material. However, the mechanical properties degrade significantly (yielding!) and the process adds cost.

The process of producing sheet from ingot material is inherently expensive because of rolling, cleaning and annealing steps and loss of material (edges, etc.) Also, in comparison to large grain ingot material the sheet manufacturing process appears more prone to introduction of defects. Reproducibility of mechanical properties from sheet to sheet is still an issue although the process is well advanced. The issue of skin rolls, affecting texture and micro-structure still needs to be addressed. Other related issues are microyielding (Myneni, SRF2005), spring back (half cells formed from different sheets/heats of material end up with different frequencies), grain size distribution and texture. Interesting work is ongoing on Equal Channel Angular Extrusion (ECAE), consisting of extruding the niobium through an angled, narrow channel. ECAE promises to produce even smaller grain size, better uniformity and better formability. It is unknown, whether this process introduces unwanted impurities.

Potential Mods to BCD with impact (tech, cost, difficulty/time scale)

Rev. July 13, 2006

Main Linac 10

The effect of impurities in the Nb is being investigated (e.g. Tantalum). It may be possible to relax specifications on impurity content without compromising the cavity performance. Such a possibility is indicated by recent prototype cavities made from higher Ta content material at JLab (800 and 1500 wt ppm), which reached high fields of 34 – 36 MV/m.

Understanding and optimizing the industrial production process, e.g. number of melts to reach the specified RRR/impurity content, should lead to high quality material at a cost savings. ECAE and other procedural steps, for instance, could yield material with better formability. Benefits from this research are also expected for the alternate cavity fabrication technologies, such as for hydro-forming.

Required R&D

The impact of the impurity content on the cavity performance (e.g. Tantalum) should be studied further; the starting point of the existing study (JLab, DESY, Reference Metals) was the claim of cost reduction benefits by the participating material supplier. It is not clear how many melts are needed to achieve the specified RRR value; however, the impurity level of interstitial impurities such as H,C,N,O affect RRR significantly and possibly improved vacuum conditions during EBM would reduce the number of melts and therefore cost. In this context new measurements of the content of light as well as heavy impurities in bulk and surface (and their effect on RRR) need to be conducted. This work would also produce improved specifications for the fine grain Nb sheet.

The mechanical properties of fine grain rolled sheet material need further exploration to study: best crystal orientation for forming, Reproducibility, Grain size control, Yield strength, Spring back, Texture, Equal Angle Extrusion, best grain size. This is especially important for alternate production technologies such as hydroforming.

Current sheet quality control measures (eddy current scanning, optical inspection) allow detection of ~100 um size defects. Thermal model calculations indicate that detection of much smaller normal inclusions is needed to guarantee ultimate cavity performance for the specified RRR-value of > 300. DESY, the University of Giessen, Heraeus, Amac Int. and Fermilab are developing a Squid Scanner aiming at 10 um resolution

Better understanding of different surface treatments, i.e. chemical polishing and electro-polishing on the cavity performance, surface roughness, oxide thickness and composition in sheet material is also a R&D priority (this has at various levels been done in the past with little success and is again at various levels going on now, e.g. CARE for EP, L.Philipps - SRF2005, Cr.Boffo SRF2005, Geng SRF2005)

Need to establish best value of RRR needed; high performance cavities have been made with Nb from RRR = 200 – 400. Nominal specification is 300.

Cost Estimation

Potential cost savings in Nb are currently expected from the single-crystal (or large grain) or the Nb/Cu laminate approaches. Another possible avenue of cost reduction is

Rev. July 13, 2006

Main Linac 11

via a relaxation of the impurity content specifications, in particular for Tantalum, provided Eacc > 35 MV/m can be reliably reached.

Cost savings can also be generated as a result of mass-production for an ILC size project (back-flow of ‘scrap-material’, economy of scale). The cost impact of large production quantities needs to be better understood, however.

Parameter Tables

Supporting Documentation

Original Snowmass document by P. Kneisel

TESLA TDR Cavity Chapter

P. Schmueser et al.; The Superconducting TESLA Cavities; PRST-AB 3 (9) 092001

Alternatives1. The most exciting new idea is the large grain/single crystal material, (started at Jlab)

mainly because it opens the possibility of ’streamlining’ the procedures at comparable performance, which could result in significant cost savings.

2. Nb/Cu laminate has the potential for significant cost savings. Single cell cavities have been made with good performance.

Single crystal and large grain

Description

Jlab has made single cell 1.3 – 1.5 GHz cavities with large grain Nb from various suppliers reaching Eacc = 34 - 36 MV/m. JLab has purchased and received a 500 kg ingot from CBMM with Ta =800 ppm and will fabricate 2 TESLA cavities from this material after initial qualification with single cells (the material is sufficient for ~ 16 nine-cell cavities). DESY and Cornell have started production of single cell cavities from various large grain material suppliers. Fermilab is currently ordering such material. Wah Chang advertised to offer large grain/single ingots in the near future.

Pro:

Single crystal or large grain material promises the following potential advantages:

Reduced costs Technical advantages may lead to simplification of fabrication and processing. Consistent gradient/Q results Possibly lower residual resistances would also lead to cost savings. Very smooth surfaces are achieved with BCP, a potential elimination of EP would

simplify production

Rev. July 13, 2006

Main Linac 12

Final cleaning of smoother surfaces might be more effective which may lead to less dark current

Good or better mechanical performance than fine grain material (e.g. predictable spring back)

Less material QA (eddy current/squid scanning) if proven by scanning a large number of sheets.

Con As a relatively new but exciting development, little experience exists at present. 9-cell

cavities with stiffeners and couplers need to be made and tested. Technology to provide large single crystals needs to be developed (it is the preferred

option to use single grain material) Large grain material with a few crystals might be acceptable.

Required R&D

Initial R&D is underway at Jlab, DESY, Cornell and Fermilab using material from various suppliers

Fast, inexpensive cutting techniques need to be identified and tested. The presently preferred wire EDM method is too slow.

The dependence of mechanical, etching and oxidation properties with crystal orientation need to be better understood.

Important topics are: acceptable yield strength of material cut directly from ingot, uniformity during forming of half cells, slippage of grains during forming, vacuum leaks through grain boundaries, grain boundary problems during EBW.

Appropriate acid agitation during BCP needs to be developed to achieve smooth surfaces and uniform material removal.

Specifications need to be developed and the first material received from the different vendors needs to be qualified.

The theoretical and experimental investigation of the effects of grain boundaries need to be pursued further [University of Wisconsin].

NIST has applied for a grant to the NRC to investigate the forming process on large grain and single crystal material.

Supporting Documentation

Original Snowmass document by P. Kneisel

Rev. July 13, 2006

Main Linac 13

Laminated Nb/Cu

Description

The laminated Nb/Cu approach takes advantages of the bulk Nb performance (Nb layer ~ 0.5 mm thick) combined with the increased thermal conductivity and stiffness of the copper backing resulting in possible significant cost savings. DESY has demonstrated the technology for multi-cell (3-cell) cavities at the laboratory level.

Pro The Nb/Cu laminate approach promises cost reduction because of reduced amount of

Nb per cavity. Gradients comparable to the best bulk niobium cavities have been achieved with

prototypes.

Better thermal stabilization as a result of the copper backing.

Stiffening against Lorentz-forces can be obtained without significant performance and cost penalty by increasing the thickness of Cu layer. In particular, the stiffening can be varied cell-to-cell.

Seamless fabrication technique (hydro-forming) allows the elimination of equator welds in the high magnetic field region of the cavities.

Most processes used for treatment of bulk niobium cavities (except for post-purification with titanium) are applicable.

Cons A first pass production technique was developed at DESY and subsequently at KEK.

Further refinement of the technique might be needed, especially in the bonding process between Nb and Copper (explosive vs hot rolling vs HIP bonding).

The e-beam welding requires cutting away the copper to make a pure Nb weld. There is a risk of contaminated welds, which might leak because of cracks in the weld. Also, local RRR reduction typically ensues following the welding.

Cu/Nb laminate cavities still quench, despite additional thermal stabilization. Thermo-currents introduced during the quenches (or during processing in a barrier)

lead to frozen-in flux, which lowers the Q-value. Cool-down has to be very uniform over the cavity volume, because thermocurrents/

frozen-in flux will destroy the Q-value. The presently used methods of cooling down cryo-modules will most likely not work

with the Nb/Cu composite material. This technology is probably not applicable with single crystal unless one would apply

standard fabrication techniques (half cell forming from composite sheets and welding).

Cracks appear in iris area during fabrication, when heat treating below the recrystallization temperature of the Nb. Heat treatment at higher temperatures causes softening of the copper. Doping of Cu with Zr was tested. Intermediate temperature

Rev. July 13, 2006

Main Linac 14

heat treatments were not sufficiently explored. More effort is required to get the tube-material to state where it can be transferred to industry.

Industrialization process has not been started yet.

Required R&D

A number of choices need to be made:

Choice of bonding method: explosion bonding (DESY), hot rolling (KEK-Nippon Steel Co.), HIP, back extrusion (DESY);

Fabrication techniques: e beam, hydro-form, Lorentz force stiffeners; How the end groups are handled: composite material (lots of EBW)?, solid niobium?

Sputtered Nb on copper with end groups flanged onto cell structure? Superconducting joint?

Can such composite cavities be ‘pipe-cooled’, which would change the whole cryo-concept?

Complete cavities with end groups need to be fabricated and tested; Cavities must be placed in modules and tested; How the cavity is made rigid against Lorentz force must be developed in detail

(varying the Cu thickness, for example); A demonstration of this technology on complete 9-cell cavities with end groups is

necessary!

Supporting Documentation

Original Snowmass document by P. Kneisel

Other Superconductors

Films and other SC considered not feasible at this time for LC project.

Fundamental Materials R&D

Overview

Over the last decade a set of procedures has been developed for the fabrication, surface treatment and assembly of superconducting niobium cavities, which lead to high performance cavities, if applied properly. These procedures include improved material QA by scanning for defects, extensive QA during cavity fabrication (e.g. cleanliness of weld joints), appropriate amount of material removal prior to heat treatments at 600 -800°C for hydrogen removal (to avoid ‘Q –disease’), BCP and EP, high pressure ultra-pure water rinsing (HPR) for extended periods of time, clean room assembly, ‘in situ’ baking and finally acceptance testing. The application of these procedures has in many cases now led to performance levels which approach the ultimate limits of the material.

Rev. July 13, 2006

Main Linac 15

Nevertheless, the underlying physics is in some areas not well understood and a fundamental material R&D program should be aimed at clarifying the physical phenomena and aid to optimize the processes. In the context of a project of the scale of the ILC improved understanding of the processes is likely to benefit performance and cost.

Fundamental materials R&D distinguished from R&D in direct support of the project- as discussed here - has to do primarily with the basics of rf superconductivity, such as the theoretical rf critical magnetic field, the loss mechanisms at high field, the nature of the RF surface ,including the Nb-oxide interface and surface contamination.

The two key performance criteria are Gradient and quality factor Q. The main performance limitations in today’s bulk Nb cavities are:

Limitation I: Ultimate cavity gradients will be limited by the theoretical RF critical magnetic field. The prevailing theory claims this is the superheating field, but its value for Nb is still under debate. Experimental data suggests the magnetic field limit to be around 185 mT.

Limitation II: Field emission at high surface electric fields. For the new shapes (discussed in the Cavity Shapes section) the surface electric fields exceed 100 MV/m for Eacc > 45 MV/m. Therefore, major efforts have to go into control of contamination, the development of clean processing and assembly procedures- especially in complex assemblies such as cavity strings and cryomodules- and the prevention of re-contamination.

Limitation III: defects in the material, which limit the achievable fields to values HRF < Hcrit. Scanning methods need to be enhanced to eliminate defects as small as 10 um as suggested by thermal model calculations.

Limitation IV: At gradients > 20 - 25 MV/m a strong degradation of the Q-value (‘Q-drop’), which significantly increases the cryogenic losses and limits the achievable gradients due to heating. This Q-drop can be eliminated/reduced by ‘in-situ’ baking at a temperature of ~ 120°C for a duration >12 hrs. Smoother surfaces (e.g. electro-polished ) give more significant improvements.

Limitation V: Residual resistances of a few nOhm have been achieved, but not on a regular basis. Low residual resistances (<3 nOhm) would allow to take advantage of the decrease of the surface resistance with decreasing temperature and an accelerator such as ILC could be operated at e.g. 1.8 K, reducing the cryogenic load and the operation costs.

R&D investigations (ongoing or under consideration)

The overarching goals of fundamental R&D should be to understand and thereby remove performance limits, ultimately leading to cost reduction.

Theoretical studies on the RF critical magnetic field Measurements of the RF critical field Preparation of field emission free surfaces

Rev. July 13, 2006

Main Linac 16

Studies to reduce field emission including emitter processing. Improvements in scanning methods to pre-screen defects Theoretical and experimental studies on the high field Q-slope and its reduction by

baking. Theoretical and experimental studies on the medium field Q-slope. Surface analytical studies of niobium using state-of-the-art instrumentation such as

XPS, Auger, SIMS, 3DAP and others. Studies to delineate the role of impurities (e.g. O, N, C, H, Ta…) on the

superconducting RF performance of Nb Studies to delineate the role of grain boundaries and other mechanical imperfections

on the superconducting properties of niobium. Studies to improve understanding of chemical treatment processes such as

electropolishing. Basic studies (similar to those describe above) on large grain and single grain

niobium.

Cavity Fabrication

Overview

The accelerating structures for ILC will need to be fabricated from well-controlled materials according to established and well-controlled methods. The scope of fabrication starts from the receipt of starting materials and ends with the completion of tuned structures meeting specified mechanical configuration criteria. It is critical that the applied methods yield a consistent ‘defect-free’ interior rf surface. The fabrication steps must not add defects to the rf surface. The options available vary somewhat with the principal starting material.

The materials options appear to be bulk Nb fine-grained, bulk Nb large-grain, bulk Nb/Cu, or, perhaps longer term, even thin film Nb on Cu. It may be possible to have combinations in the future.

Since structure design is treated separately, one may note that fabrication R&D bears only weakly on couplings to other design elements. Therefore, competitive pressure to develop lower cost methods of providing the chosen structure may proceed until construction start.

A serious industrialization study has only been made for the current ’standard’ fabrication method (1.a.). More similar studies may be necessary.

Principal opportunities for cost savings:

Reduce material costs by reducing the amount of required Nb, substituting cheaper material where possible.

Reduce the number of components and steps. Reduce the ‘hands-on’ time for each cavity. Develop methods with low-effort based QA

Rev. July 13, 2006

Main Linac 17

Options under consideration1. Bulk fine-grained niobium from sheet, bar, and plate stock

1. Machining, deep draw forming, mechanical polishing, pre-cleaning chemistry, EBW, inelastic deformation tuning with appropriate intermediate cleaning steps

2. Similar to 1.I., but substituting spinning for the deep draw forming and the EBW of cells.

3. Similar to 1.I., but substituting hydroforming for the deep draw forming and the EBW of cells.

4. Similar to 1.I. for the cavity cells, but treating the fabricating the endgroups differently and using Nb film on copper for these low-field parts.

2. Bulk large or single-grain niobium direct from ingot1. Similar to 1.I., but using wire EDM to form sheets for cell blanks Evidence to

date suggests that the balance of fabrication is not significantly changed with respect to ‘standard’ fine-grained Nb material.

3. Bulk Nb/Cu clad material for cells (and beamtubes?)1. Similar to 1.III. - hydroforming

BCD choice

1. Present ‘standard’ fabrication methods, applied with serious attention to QA. i.e.

1. Bulk fine-grained niobium from sheet, bar, and plate stock2. Machining, deep draw forming, mechanical polishing, pre-cleaning, EBW, inelastic

deformation tuning with appropriate intermediate cleaning steps

(2. is not significantly different in fabrication methods but several complete cavities have to be made with all features and tested. Yield strength of ingot material may be significantly lower than 1.I, affecting final rigidity of finished structures)

Description

Pros There exists a strong experience basis for describing and implementing appropriate

fabrication methods for bulk sheet Nb cavities. ~1000 such cavities exist. Costs are well understood from low-volume production runs. Serious production analyses exist for high-volume scale-up of the ’standard’

fabrication methods.

Cons ’Touch’ labor is relatively high, with piece part handling, cleaning, and inspection

required for EBW steps, many nuts and bolts in flanged connections to keep clean. The endgroups require as much fabrication attention as the cells. This seems less

justified for the cost.

Rev. July 13, 2006

Main Linac 18

Mass production analysis for TESLA-type cavities showed 77% of residual fabrication cost in machining operations.

Required R&D

Reduce material costs by reducing the amount of required Nb, substituting cheaper material where possible.

Reduce the number of components and steps. Reduce the ’hands-on’ time for each cavity.

Supporting Documentation

Original Snowmass document by C. Reece

TESLA TDR Cavity Chapter

P. Schmueser et al.; The Superconducting TESLA Cavities; PRST-AB 3 (9) 092001

Alternatives

Overview

Motivations for considering alternate fabrication choices are almost exclusively related to cost. No presently considered alternative methods claim to directly improve ultimate performance.

Cost reductions may result e.g. from methods which reduce the cost of required material or aid the automation of fabrication. Thus the hydroforming, spinning, and film coating might be considered. Hydroforming and spinning offer the prospect of seamlessly forming the cavity cells, while eliminating several machining, chemical cleaning, and EBW steps.

Hydroforming of cell structure

Pros Technique quite suitable for factory production with automation and reduced total

fabrication costs. The technique can produce equally performing Nb cavities. (One-cell cavities: 42

MV/m and Q - value ~ 1010 test cavity without EP) (KEK & DESY) 3-cell structures have been built. If applied to Nb/Cu clad tubing, the quantity of required high-purity Nb for cells

could be reduced by 75%. Highly consistent interior cell geometry expected, thus less required tuning. Avoidance of machining and welding in the high field regions of the cavity eliminates

some potential sources of defects which could degrade ultimate cavity performance.

Cons

Rev. July 13, 2006

Main Linac 19

Less experience translates into less awareness of subtle difficulties. With Nb/Cu, must manage the end transitions, e.g. Cu removal to avoid Nb weld

contamination. Tube production needs development, this is underway at INFN Must assure tubing QA for consistent forming properties.

Potential cost impacts Elimination of multiple machining steps and expensive EBW time. Expected time required to form 9-cell structure from tube: ~6-8 hours If applied to Nb/Cu clad tubing, would reduce quantity of required Nb. Needs serious study to determine cost benefit

Risk and Reliability impacts Impact on production yield is unknown. Delamination of Nb-Cu from stresses due to multiple thermal cycles.

R&D Work is proceeding under JRA1 – CARE program. Fabrication of seamless bulk Nb tubes of the length sufficient for 9 cell cavity from

one piece (ca. 1.8 m long) Development of ‘industrial’ production routine and qualification of cavities. How to avoid or suppress the trapping of magnetic flux caused by thermo - coupling

effect in Nb/Cu cavities? New methods of bimetallic tube fabrication More appropriate material for Nb clad cavities instead Cu (Cu alloys etc.)? End group cost reduction for Nb/Cu clad cavities Seamless cavity of new shapes (low losses, re-entrant etc.)

Time scales for R&D Bulk Nb “seamless” 9 cell TESLA shape cavity (e beam weld three 3-cell units)

suitable for installation in the cryomodule (2006). New machines (capital investment) needed for 9-cell hydroforming Multi cell NbCu clad cavities from special copper without Cu layer inside (2006)??

Spinning

Pros Technique quite suitable for factory production with automation and reduced total

fabrication costs. The technique can produce equally performing Nb cavities up to 40 MV/m in single-

cell test cavities (INFN Legnaro) 9-cell structures have been built, awaiting test.

Cons

Rev. July 13, 2006

Main Linac 20

Nb tube production needs development Must assure tubing QA for consistent forming properties. Multi-cell performance not yet demonstrated Substantial inside surface grinding required to remove fissures. This may be reduced

by 800 C annealing before spinning tube. Less experience translates into less awareness of subtle difficulties.

Technical advantages, increased tech potential Avoidance of machining and welding in the high field regions of the cavity eliminates

some potential sources of defects which could degrade ultimate cavity performance. Limited experience makes evaluation difficult.

Potential cost impacts Elimination of multiple machining steps and expensive EBW time. Expected time required to form 9-cell structure from Nb tube: 4 hours in mass

production (INFN) Estimated potential fabrication cost reduction: (Needs Industrial study.) Possibility of spinning thicker segments for stiffening against Lorentz-Force detuning

Risk and Reliability impacts Impact on production yield is unknown. Substantially more material removal is necessary if fissures are present Wall thickness uniformity

R&D Work is proceeding under JRA1 – CARE program. Production of Nb tubing for multi-cells needs development Development of industrial production routine with multiple cavities and RF

qualification of those cavities.

Nb films on Cu for endgroups

Pros Potential for reduced cost by using Nb film coating in low-field regions. Improved thermal conduction in portions of accelerating structure outside of the

helium vessel. If successfully made demountable, (as proposed by KEK) could maximize QA of high

field region of cells by providing opportunities for restructuring inspection, cleaning, and assembly sequence.

Successful SC flanging would create a new type of modularity that could be exploited.

Cons Complexity of endgroup shape makes confident coating difficult.

Rev. July 13, 2006

Main Linac 21

Added complexity due to material interfaces at joints etc. Still at very early conceptual stage Lack of experience base leaves potential problems unrecognized.

Potential cost impacts Material cost could be reduced Greatly reduced need for EBW steps if Cu endgroups can be formed as a brazed

assembly, followed by Nb film coating and endgroups are flanged to seamless cell structure.

Studies needed to estimate cost impact

Risk and Reliability impacts Impact on production yield is unknown. Film adhesion problems in complex geometry

R&D Develop Nb coating of HOM coupler and end group assembly. Develop low-profile reliable superconducting flange joint for use just outside of

helium vessel. Develop less complex HOM damping scheme for easier fabrication and coating. Industrial cost study: Is the undeveloped Nb coating cheaper in production than

endgroups of BCD? Is the yield sufficiently high?

Cavity Preparation

Overview

The preparation of the cavities should finally result in fully assembled cavities (incl. power coupler) which are ready for string assembly. After delivery from the welder, there are several important steps:

1. Leak check, mechanical checks, inspection2. Frequency tuning, field flatness3. Cleaning4. Damage layer removal5. Furnace treatments6. Final frequency tuning, field flatness7. Final surface preparation 8. Final cleaning9. Bake-out at 120-130°C10. Low-power acceptance test11. Tank-welding12. Assembly for high power operation13. High-power test

Rev. July 13, 2006

Main Linac 22

All these steps need improvements in QA/QC for mass production. The most challenging one is to define final (electro-)chemical surface preparation to deliver reliable and reproducible performance. Today, highest cavity performance has been achieved with electropolishing as main damage layer removal and final surface preparation process followed by baking.

Baseline

The BCD recommendation for this process is:

Electropolishing for damage layer removal (~120-150 um) 800°C furnace treatment Electropolishing for final preparation (~20-50 um) High pressure rinsing for final cleaning

There is no data on multi-cell cavities which suggests that this procedure can be avoided. Initial results on single-cell cavities with other material (see section Cavity Material) need to be confirmed by multi-cell experiments.

Required R&D

The major risk in the cavity preparation is the contamination of the inner cavity surface which then leads to enhanced field emission and performance degradation. In most series productions of cavities field emission is the reason for performance limitation. Therefore, it is highly desirable to further improve quality control of the processes applied to the cavity. Moreover, any development on cavity shapes with an increased ratio of electric surface peak field to accelerating gradient (Epeak/Eacc) requires an even larger, continuing effort to reduce field emission.

Substantial experience has been accumulated on cavity preparation systems. Still, most of the facilities in operation are small scale systems compared to ILC needs, and there are differences in capabilities of existing systems. The R&D on the cavity preparation should aim for improved preparation facilities of the next generation designed for improved quality control.

There are three main areas in the cavity preparation process where a strong R&D plan is needed:

Electropolishing (EP) system: A generic parameter set for niobium cavity EP was developed at KEK. So far the experience on multi-cell cavities is not as reproducible as desired. Several improvements are suggested.

o Improved parameter control: Development of efficient heat exchangers to improve temperature

stability

Rev. July 13, 2006

Main Linac 23

Compensation for losses of hydrofluouric acid (HF) from evaporation and chemical reaction with the niobium by actively adding HF during the process.

Optimization of the current distribution for homogenous material removal

o Contamination control: Avoidance of sulphur by either improved post-EP rinsing methods (e.g.

alcohol rinse) or changed EP parameters (e.g. bath composition) Definition of measurement techniques for contamination and quality

control (e.g. Nb and HF content of the electrolyte)o Quality control

Development of a roughness measurement of the inner cavity surface that goes beyond the information supplied by witness samples

High pressure rinsing (HPR): Currently, the rinse of cavities with high pressure water is the only effective means of removing particles from the cavity surface. Further R&D is required to reduce performance spread and increase the onset level of field emission.

o Improved parameter control Optimization of the cleaning force by proper adjustment of

Nozzle geometry, material and size. Impact angle and flow rates

o Contamination control Online particulate monitoring on both the high pressure input side and

on the drain water from the cavity Reliable online TOC (total organic carbon) monitoring of input and

output water. Understanding of specifications needed on other contaminants (e.g.

dissolved solids).

Assembly procedures: Currently, the assembly and cleaning of components is not streamlined to fit into a mass production environment. To facilitate this some development is needed to simplify procedures and reduce the amount of parts used during assembly. An investigation on improved tooling and (semi-) automation is necessary.

o Assembly of components For the quality control of assembly procedures it is desirable to

develop a technology to assess particle contaminations of the inner cavity surface.

Methods for mass production need to be developed (e.g. cleaning of screws, gaskets etc.)

The assembly of the cavity string needs a further improved quality control procedure as it is currently not possible to clean the full string after assembly.

Rev. July 13, 2006

Main Linac 24

The overall workflow during inspection, preparation, assembly and testing needs optimization to avoid contamination and increase cost effectiveness.

o Drying An evaluation on drying procedures after HPR is needed to avoid re-

contamination of the surface An integration of the drying with the ‘in-situ’ bake-out seems desirable

and should be explored.

These are critical R&D issues for a cost efficient production with a high yield. Mass production issues need to be addressed for tuning and cleaning, as well as a serious effort to reduce the number of parts to be assembled. Any new R&D infrastructure for the ILC must take into account the enhanced need for quality control. Some of the topics are:

Integration of mechanical measurements (eccentricity, etc.), optical inspection (EB welds etc.) into the cleanroom area

Frequency tuning: Integration into clean room seems desirable (see above).

Vertical cavity tests for all cavities are mandatory (preferably with many cavities inserted into one vertical test dewar) and will potentially allow sorting cavities during production, with some benefit to overall linac performance.

High-power tests on individual cavities are necessary in an initial production phase where the critical assembly steps (e.g. high-power coupler mounting) are set up and where sub-components are qualified (e.g. tuners). With increasing confidence, the number of these tests may be reduced.

Full tests on all modules are mandatory in the first phase of the production. When the production process is established, a statistical approach on testing might be feasible especially if a cool-down of linac sections in the ILC is envisaged, so that major problems (e.g. vacuum leaks) are detected before machine operation starts. A risk analysis of different scenarios is desirable, and should include cavity preparation experts.

Supporting Documentation

Original Snowmass document by L. Lilje

TESLA Report 2004-04 Improved Surface Treatment of the Superconducting TESLA Cavities - L. Lilje, A. Matheisen, D. Proch - DESY; C. Antoine, J.-P. Charrier, H. Safa, B. Visentin - CEA Saclay, DAPNIA; C. Benvenuti, D. Bloess, E. Chiaveri, L. Ferreira, R. Losito, H. Preis, H. Wenninger - CERN; P. Schmueser - Universitaet Hamburg

TESLA Report 2004-05 Achievement of 35 MV/m in the Superconducting Nine-Cell Cavities for TESLA - L. Lilje, D. Kostin, A. Matheisen, W.-D. Moeller, D. Proch, D.

Rev. July 13, 2006

Main Linac 25

Reschke, S. Simrock, K. Twarowski - DESY; E. Kako, K. Saito - KEK; P. Schmueser - Universitaet Hamburg; T. Suzuku - Nomura Plating Co., Ltd.

Alternatives

A brief discussion summary on the alternatives to the BCD is given below. For details refer to the original Snowmass document by L. Lilje.

Postpurification of material at 1400°C with titanium getter Eliminate outside etching Hot water rinsing Tumbling/barrel polishing to reduce amount of EP necessary Dry-ice cleaning instead of high pressure water rinsing Air bake out instead of in-situ vacuum bake at 120 – 130 C.

Required R&D

For all the alternatives

Single-cell R&D is necessary. Multi-cell issues need work.

Supporting Documentation

Original Snowmass document by L. Lilje

High Power Input Coupler

BCD Choice

Our BCD choice is based on the twin cylindrical window architecture of the TTF-III coupler.

Description

Pro:

The principal “pros” of this choice are the following:

Lengthy experience (~ 100,000 coupler-hours) of this device on the TESLA Test Facility.

Some degree of manufacturing experience – 52 couplers built, mostly in industry. Demonstration of use with a cavity at 35 MV/m on CHECIA. Tested at a power of 1 MW, 1.3 ms pulse in TW mode. This coupler has already been accepted for use with the European X-FEL (~ 1,000

units needed) and this implies that there will be considerable experience gained with this coupler before the ILC is launched.

Rev. July 13, 2006

Main Linac 26

In principle, then this power coupler would be sufficient even if the cavities were to be run at 35 MV/m and would meet, at least in TW mode, the needs of a 2 x 9-cell superstructure at 35 MV/m.

Cons:

The “cons” of this choice, at the time of writing are:

The present unit cost is prohibitive. However, the couplers have only been built in small numbers to date.

o Note that the cost issue will be dealt with through an “Industrialisation” study to be carried out by LAL-Orsay (co-financed by DESY and the IN2P3) in the context of the European X-FEL project. We aim for a major reduction in the unit cost of the couplers through this study.

The experience with conditioning indicates that the conditioning time is rather long. o However, note that the scatter in conditioning time is rather large and the fact

that some examples are conditioned rather quickly (< 50 hours) is encouraging.

o The issue of conditioning is currently under study at Orsay in the context of a DESY-LAL collaboration and partly financed through the European Union initiative – CARE (Co-ordinated Accelerator Research in Europe). It is hoped that this activity will lead to procedures offering reduced conditioning times.

A period of ~ two years will be necessary to complete the conditioning and indutrialisation studies (i.e. completion around summer 2007).

Potential Modification

A modification of potential interest is an increase in the diameter of the cold assembly (from 40 mm to 62 mm). This would have the technical benefit of pushing multipactor levels to higher powers and therefore may be of interest in case of a choice of higher gradient ( ~ 45 MV/m).

Four proto-types of such a coupler has been ordered by Orsay and should be delivered in the spring of 2006 and tested soon afterwards. The R&D necessary for such a modification could be complete by the end of 2006.

Required R&D

See above.

Cost Estimation

Supporting Documentation

Original Snowmass document by T. Garvey.

Rev. July 13, 2006

Main Linac 27

Alternatives

There was a general consensus that different coupler designs incorporating two « disk » type windows could be potential alternatives to the TTF cylindrical window. Four such couplers are currently under study and we list them here with no order of priority:

The “capacitive” disk window coupler. The ‘TRISTAN’ like window coupler. The TW60 coupler. The AMAC window coupler. (this last coupler was added by T. Garvey after the

meeting. It is currently being developed by AMAC, in collaboration with DESY and CPI with funding from a DoE SBIR grant).

Each of these couplers were presented in more or less detail at Snowmass and a description can be found in the presentation on the Workshop web site.

Description

Pro:

Some ‘pros’ are as follows:

Disk windows are (or can be expected to be) relatively free from multipactor. Disk windows are mechanically easy to fabricate and therefore may be cheaper (but

the ceramic is not a cost driver for the coupler). Thin disk windows can be positioned at low values of the SW electric field. Disk windows should be relatively easy to braze into the coupler. One should note that windows based on the TRISTAN design have a history of

success.

Con:

Some ‘cons’ are as follows:

The current version of the capacitive coupler cannot be DC biased. The present capacitive and TRISTAN like couplers have no possibility to have their

external Q variable. Disk ceramics are in the “line of sight” of the cavity beam pipe. This was seen to be a

problem on the early CEBAF linac design, which was a waveguide coupler. However, the co-axial version may be less problematic as the window surface is smaller and is further from the beam pipe, thus reducing the solid angle presented to x-rays or electrons which might impinge on the ceramic.

For all of these alternatives it is too early to estimate the cost impact.

Required R&D

Rev. July 13, 2006

Main Linac 28

Time for R&D

As for time scales for R&D, the AMAC coupler is currently under test at DESY. A proto-type of the TW60 coupler will be tested at Orsay in the summer of 2006 at the earliest. The capacitive coupler will be tested at high power at KEK early in 2006.

Supporting Documentation

Original Snowmass document by T. Garvey.

HOM Couplers and Beam Line Absorbers

Overview

The accelerated ILC beam, if similar to the proposed in the TDR beam of the TESLA 500 GeV collider, will consist of 3.2 nC bunches which rms length will be 300µm. The 14100 bunches accelerated per second will be grouped within five 950 µs long RF pulses. The time separation between sequential bunches in a RF-pulse will be 337 ns. This beam will generate spectrum up to 0.4 THz. The beam deposited power in a cryo-module housing 12 TTF-like 9-cell structures will be ~24 W if no synchronous excitation of parasitic modes will take place. Big fraction of this power (17.4 W) is carried out by propagating modes above 5 GHz.

The beam deposited power must be removed from cryomodules to avoid an additional heat load in 2K environment and to maintain the high quality of the accelerated beam (preserving the low emittance). This will be achieved by means of two kinds of devices:

HOM couplers Beam line absorbers

Each accelerating structure will be equipped with two HOM couplers suppressing mainly the non propagating part of the spectrum (below 5 GHz). Beam dynamics simulations showed that preservation of the low emittance demands suppression of high impedance dipoles to Qext of the order of 10^5. This suppression will also ensure stable operation if resonant excitation of some high impedance modes takes place.

One beam line absorber will be installed in each interconnection between cryomodules. It is at present proposed and being under development version should absorb 80% of the energy propagating out of neighboring cryomodules. The first prototype will be ready for beam test by the end of October 2005.

Baseline

The suppression scheme as proposed in the TESLA TDR

Description

Rev. July 13, 2006

Main Linac 29

Required R&D

The careful studies of HOM suppression at TTF linac showed that almost all dipole modes, but two, are well damped and satisfy the specification. To improve damping of these modes which Qext was above the spec one of two HOM couplers for all recently produced cavities has been positioned in a different way. The expected improvement can be first verified when a cryomodule with new cavities will be installed in the TTF linac. This should happen in 2006.

Further R&D is needed on getting reproducible properties of beam line absorber material. We also expect that the prototype of the beam line absorber will be installed and tested at TTF linac also in 2006.

The fraction of HOM power deposited at 2 K should be investigated.

Supporting Documentation

Original Snowmass document by J. Sekutowicz

TESLA TDR Cavity Chapter

P. Schmueser et al.; The Superconducting TESLA Cavities; PRST-AB 3 (9) 092001

TESLA Note 2002-05 Higher Order Mode Absorption in TTF Modules in the Frequency Range of the 3rd Dipole Band M. Dohlus, S.G. Wipf - DESY, V. Kaljuzhny - MEPI

Alternatives1. TDR scheme with minor improvements2. Coaxial coupling fitting in the beam line longitudinally and radial.3. Other solutions

TDR scheme with minor improvements

The proposed TDR HOM suppression scheme with minor improvements should work properly. Some additional improvements like higher heat conductivity feedthroughs for the output lines can be implemented at any time of cavity production with minimal risk.

Options under consideration

Further improvements reducing costs of HOM couplers and beam line absorbers are also considered for XFEL cavities and can be later implemented for the ILC collider. Three of them are listed below:

Radial positioning of the HOM coupler output in the plane of so called F-part. Version of HOM coupler with hidden output capacitor. Version of HOM coupler without output capacitor.

Rev. July 13, 2006

Main Linac 30

Coaxial coupling fitting in the beam line longitudinally and radial.

A more advanced and revolutionary change should lead to a device that replaces the two HOM couplers rigidly attached to a cavity and that is mountable between cavities. This could simplify cavity production if remaining body will be cylindrically symmetric. For this also FP coupler should be removed. The Cu model of beam line FP coupler (very similar to the Darmstadt linac solution) was built and tested on Cu model of 9-cell cavity. At the moment there is no conceptual design for the HOM coupler.

Supporting Documentation

Original Snowmass document by J. Sekutowicz

TESLA Note 2002-05 Higher Order Mode Absorption in TTF Modules in the Frequency Range of the 3rd Dipole Band M. Dohlus, S.G. Wipf - DESY, V. Kaljuzhny - MEPI

Frequency tuner

Overview

Main objectives for the frequency tuner are to provide means to tune the cavity on resonance, detune a cavity to by-pass operation if needed, and to compensate Lorentz-force detuning. The tuner further needs to allow for a high linac fill factor (compact design), should be hysteresis free, and should not cause cross-tuning of neighboring cavities. Long life time of the tuner is essential; see discussion below. All this needs to be achieved with lowest cryogenic losses, and at low cost.

ILC requirements:

Coarse tuning range: 500 kHz (1.6 mm at 315 Hz/ïm) Coarse resolution: <5 Hz

Fast tuning range (static at 2K): o delta_f=2*K*E^2 (factor 2 for dynamic operation overhead) o 2.5 kHz (for K=1 Hz/(MV/m)^2 cavity at 35 MV/m)o 3.2 kHz (for K=1 Hz/(MV/m)^2 cavity at 40 MV/m)o 4.0 kHz (for K=1 Hz/(MV/m)^2 cavity at 45 MV/m)

The requirement on fast running range is not well known at this point. Significant spread in the dynamic Lorentz-force detuning constant has been seen (factor 2 at TTF). Unless K can be controlled well in the ILC cavity/LHe vessel production, more fast tuning range is required. Also, significant difference can exist between the static range and the dynamic range (maximal frequency shift within the RF pulse length). A factor of 2 is included in the above numbers for the ratio static / dynamic range. Dedicated experiments are needed to define the actually required fast tuning range.

Rev. July 13, 2006

Main Linac 31

Options under consideration:

The following existing tuners could provide a basis for the ILC frequency tuner (see original document for pictures):

Original Saclay / TTF tuner: o This type of tuner is in use at TTF/VUVFEL since several years and thus is

well tested. It does not have a fast tuning element, though a piezo actuator has been added for proof-of-principle tests of Lorentz-force detuning. The range however of the fast tuner is below 500 Hz, and the tuner was not initially developed to implement a piezo actuator.

Modified Saclay tuner: o Similar the original Saclay tuner, this tuner is located at one end of the cavity.

The redesign however is more compact, and incorporates piezo actuators (about 1 kHz tuning range). First tests of this tuner are expected by end of 2005.

INFN / DESY blade tuner: o The first version of this tuner did not include fast actuators, and was tested at

TTF with the superstructure. This tuner is located around the LHe vessel, thus does not require any clearance at the cavity ends. The recent version of this tuner includes piezo actuators (about 1 kHz tuning range), and will be tested late 2005.

TJNAF Renascence tuner:o This tuner was developed for the TJNAF upgrade cryomodule. It does

incorporate piezo actuators (about 1 kHz tuning range), and eight tuners of this type will be operated and tested in a cryomodule test late summer 2005.

Saclay original

Saclay modified

INFN Blade tuner

TJLAF upgrade

KEK Slide Jack Tuner

KEK coaxial ball screw

Coarse Range [kHz] 440 500 500 500 1100 >4000

Coarse Res. [Hz] <1 <1 <1 <1 <100 <120

Fast actuator (Piezo) Piezo Piezo Piezo/ Magnetostr.

Piezo Piezo

Number of fast actuators (1 - 2) 2 2 2 1 1

Fast range [Hz] <500 1000 1200 1000/ 30000 1900 2500

Position of fast 5 K, 5 K, 5 K, 5 K, vacuum 5 K, vacuum 80 K,

Rev. July 13, 2006

Main Linac 32

actuator vacuum vacuum vacuum vacuumPosition of

motor 5 K, vacuum

5 K, vacuum

5 K, vacuum

5 K, vacuum Warm, outside

80 K?, vacuum

KEK slide jack tuner: o KEK design for cavity operation at the baseline gradient of 35 MV/m. The

uniqueness of this design is that the motor is placed at room temperature outside of the vacuum vessel.

KEK coaxial ball screw tuner: o This tuner was designed for the ICHIRO 9-cell cavity at 45MV/m. Both, the

motor and the piezo are placed at intermediate temperature inside of the vacuum vessel. A first prototype test is planed for late summer 2005.

Risk and Reliability impacts:

All but one of the above mentioned designs have a cold drive motor inside the vacuum vessel. In none of the designs can the fast actuator be replaced without cryostat warm-up. Highest reliability / lifetime are therefore essential. The main risk is a failure of the motor, the fast actuators or the gearing. All designs with cold drive can use the same type of motor, gearing and fast actuator, so that there is no principle difference in risk and reliability between these tuner designs. A well tested version of the motor and gearing exists. Tests on piezo performance and reliability are underway.

However, the reliability of a single cold drive might not be sufficient, as was pointed out by the U.S. Linear Collider Technology Options Study: ‘The cavity tuners and cavity piezo tuners designs both require opening the cryostat to effect repairs and had over 50 failures per year. This is an unreasonable amount of work even for the 3 month shutdown. The tuners will either have to be made very reliable (probably via redundancy) or their failure prone components made replaceable without warm-up.’

To improve reliability, the following options exist:

Redundant motor and piezo, if inside of vacuum vessel Improved design with highest reliability for motor and/or piezo, if inside of vacuum

vessel Warm motor

It should be pointed out, that the operation of the fast actuator is essential at high fields. A failure of this element will result in lower operating fields. The motor on the other side is only operated during cool-down and warm-up, and to correct for slow frequency drifts. A failure of the motor will not immediately impact the cavity operation. It can be expected that the total required step-count of the motor is moderate. The failure mechanism and the MTBF in this operating mode need to be studied in detail, to verify if a cold motor is acceptable.

Rev. July 13, 2006

Main Linac 33

Fast actuator:

Two options are under consideration:

1. Piezo actuator (used in all designs as baseline). o Detailed studies have been done to verify pulsed cryogenic operation in

radiation environment.2. Magnetrostrictive actuator (similar size to piezo, so can be used instead of piezo in all

discussed tuner designs): o Has a significant larger stroke than a piezo at 5 K, produces less heat and

might have a higher lifetime and higher tolerance for preload change than piezo. First tests at cryogenic temperatures have been done. A detailed characterization is need. This actuator needs a high drive current, and its residual magnetic needs to be studied. Also, the cost of this actuator type might be higher.

Baseline

Not available. No existing tuner design fulfills the specification on fast tuning range above 30 MV/m. The above mentioned designs give a good starting point for an ILC tuner and for a cost estimate. The tuner needs to provide 500 kHz slow tuning range and more than 3 kHz fast tuning range.

The fast actuator should be located inside the vacuum vessel for best performance during Lorentz-force compensation. A redundant design for the fast actuator is important for reliability.

Required R&D

Tuner design for 40+MV/m operation and prototype tests including demonstration of Lorentz-force detuning at highest fields with BCD cavity.

Reliability (MTBF) studies of motor / gearing / piezo / magnetostrictive actuator, including failure mechanisms and improved estimate of requirements.

Performance of magnetostrictive actuator. Cavity design with smaller Lorentz-Force detuning. Cost estimation for external motor.

Parameter Tables

Supporting Documentation

Original Snowmass document by M. Liepe and S. Noguchi

Cost Estimation

Rev. July 13, 2006

Main Linac 34

Close to BCD:1. Modified Saclay tuner:

o Pros: Relative simple and compact design Redundant design for piezo element Original Saclay tuner was tested in detail

o Cons: Maybe difficult to increase fast tuning range Redesign needed with increased fast tuning range Poor maintainability of stepping motor and fast actuator Requires some length between cavities (located at cavity end)

o R&D necessary: Design with increased fast tuning range Fast actuator R&D Prototype tests with Lorentz-force compensation at 35 MV/m Verification of sufficient MTBF for cold motor

2. INFN blade tuner:o Pros:

Compact design (not at cavity end) High stiffness Tested (without fast actuator) Relative easy to increase fast tuning range

o Cons: Redesign needed with increased fast tuning range Poor maintainability of stepping motor and fast actuator

o R&D necessary: Design with increased fast tuning range Fast actuator R&D Prototype tests with Lorentz-force compensation at 35 MV/m Verification of sufficient MTBF for cold motor

Supporting Documentation

Original Snowmass document by M. Liepe and S. Noguchi

Alternatives1. TJNAF Renascence tuner:

o Pros: Compact and simple design (not at cavity end) Low cost?

o Cons: Not originally designed for ILC cryomodule. May need some redesign

to fit. Redesign needed with increased fast tuning range

Rev. July 13, 2006

Main Linac 35

Poor maintainability of stepping motor and fast actuatoro R&D necessary:

Design for ILC cryostat Design with increased fast tuning range Fast actuator R&D Prototype tests with Lorentz-force compensation at 35 MV/m Verification of sufficient MTBF for cold motor

2. KEK slide jack tuner:o Pros:

Motor outside of vacuum vessel (inexpensive motor) Piezo can be replaced (cryostat warm-up required) High stiffness

o Cons: Feed-through to outside needed (penetration of shields and vacuum

vessel) Some static losses (0.05 W?) Redesign needed with increased fast tuning range Poor maintainability of fast actuator; no redundancy

o R&D necessary: Design with increased fast tuning range Fast actuator R&D Prototype tests with Lorentz-force compensation at 35 MV/m

3. KEK coaxial ball screw tuner:o Pros:

Wide tuning range Compact design with common technology Cost effective High stiffness Maybe access to piezo (warm-up required; need to pass through all

thermal shields)o Cons:

Poor maintainability of stepping motor Poor maintainability of fast actuator; no redundancy Heavy weight Some static losses Redesign needed with increased fast tuning range

o R&D necessary: Choice of coating material for balls Weight reduction Design with increased fast tuning range Fast actuator R&D Prototype tests with Lorentz-force compensation at 35 MV/m Verification of sufficient MTBF for cold motor

Supporting Documentation

Rev. July 13, 2006

Main Linac 36

Original Snowmass document by M. Liepe and S. Noguchi

Rev. July 13, 2006

Main Linac 37

Gradient

Baseline

The WG5 recommendations call for TESLA-like cavities to be used. They would be qualified to operate at a gradient of at least 35 MV/m with a Q > 0.8×1010 in CW tests (cavities not meeting these requirements would be rejected or reprocessed). Only a small fraction of the cavities and cryomodules would be pulsed-power tested. With such screening, they expect that a 31.5 MV/m gradient and Q of 1×1010 would be achieved on average in a linac made with eight-cavity cryomodules. This assumes that (1) the rf system would be capable of supporting 35 MV/m operation throughout the linac (2) some of the poorer performing cavities would be de-Q’ed so the associated cryomodule can run at a higher gradient and (3) the cryomodule power feeds would include attenuators so the average gradient in each unit can be maximized. For a future upgrade, they recommend that cavities of the low-loss or reentrant type be used and that they be qualified to at least 40 MV/m with Q > 0.8×1010 in order to achieve 36 MV/m and Q = 1×1010 on average in the linac.

Alternatives

Since improvements in cavity performance will likely continue, one design strategy would be to choose a gradient significantly higher than that currently achievable. However, the linac cost is a weak function of gradient in the 30-50 MV/m range, and operating close the ultimate 45-50 MV/m gradient limit would prevent extending the machine energy by lowing the beam current (and depending on the cooling overhead, lowering the machine repetition rate). Thus a better strategy would be to design for a gradient around 30 MV/m, and if the cavities that are eventually installed perform better than the initial requirement, use this capability to extend the machine energy reach (e.g., up to 750 GeV if 45 MV/m operation is eventually achieved). The luminosity would decrease with higher energy, but still may allow for discovery-level measurements. The WG5 ACD gradient recommendation for 500 GeV operation is the same as given for the BCD upgrade.

Rev. July 13, 2006

Main Linac 38

Cryomodule and Lattice

Baseline

Adopting the WG5 gradient and cavity recommendations and assuming TDR-like rf distribution losses and overheads, a reasonable baseline rf unit is a 10 MW klystron driving 24 TESLA cavities. This configuration allows 35 MV/m operation with 7% rf distribution losses and an 11% power overhead (below klystron saturation). The overhead is an estimate of that required to stabilize of the cavity voltages and to operate in a reasonably linear regime: an overhead based on operation experience with ILC-like cryomodules should eventually be used.

The cavities would be divided into three cryomodules instead of two since (1) this is the configuration that has been used and will continue to be used for several years (2) there is no significant cost savings with longer cryomodules, which would be more difficult to build, transport and make vibrationally stable and (3) the cavity gradient variation can be more efficiently dealt with if there are less cavities per cryomodule, assuming the power to each cryomodule can be controlled using attenuators.

Every third cryomodule in the linac would include a superferric or cos(2*phi)-type quadrupole (this corresponds to a constant beta lattice with one quadrupole every 24 cavities). The quad He vessel would be supported from above by the 300 mm diameter gas return pipe, which itself would be supported by three posts extending down from the top of the cryomodule. The quad would be located below the center (fixed) post, and attached to its upstream face would be a BPM with 10 micron or better bunch-to-bunch position resolution for the nominal bunch charge of 2×1010. On the downstream face of the quad, a combined horizontal and vertical dipole corrector magnet would be attached.

A laser wire scanner would be located in each of the three warm sections between the 2.2 km long cryogenic units within each linac. The 1.2 km undulator region in the electron linac, and the corresponding section in the positron linac, would each contain emittance diagnostics and energy and energy spread measurement capability.

The TDR cavity spacing of 283 mm was optimized based on the flange connection and bellow scheme used at TTF. Shorter spacing is possible (see ACD below), but the TDR spacing is assumed until the engineering implications are better understood. With this choice, the linac packing fraction (ratio of active to actual length) would be about 70%.

For 500 GeV operation, 316 (312) such rf units would be installed in 15-250 GeV electron (positron) linac, where the additional units in the electron linac would restore the energy lost in the undulator. These unit counts provide a 3% energy overhead with 5 degree off-crest beam operation. Overall, 71% of the peak rf capability would be transformed into beam power.

Rev. July 13, 2006

Main Linac 39

Alternatives

Several variations on the TDR-like layout were discussed including:

Having up to 12 cavities per cryomodule instead of 8 to reduce the number of inter-connections.

Shortening the distance between cavities to 250 mm based on the connection scheme used at JLab, or as close as possible to the 180 mm limit from cavity cross talk and heat losses at the transitions.

Instrumenting the HOM readouts to provide a measure of the average beam position in the cryomodules. To better center the beam, the cryomodules would either be moved manually during down periods or equipped with remote-controllable movers to allow corrections during machine operation.

Putting the quad and BPM in a separate cryo-section to better stabilize them vibrationally and allow them to be moved independently from the cavities. This would eliminate the need for corrector magnets.

Putting movers on the middle support post of the gas return pipe to allow adjustment of the quad and BPM positions (the nearby cavities would move as well). To allow independent control of the quad and cavity (average of eight) positions, the cryomodules would also have to be supported on movers.

Reducing the quad aperture by half (to about 35 mm) to allow the use of superferric quads, which will likely have more stable magnetic centers with respect to quad shunting. Likewise, reducing the BPM aperture by half to yield higher resolution and smaller common-mode errors. These changes would increase the short-range wakefield by about 10%.

Improving the BPM resolution to about one micron to allow measurements of beam jitter at a level smaller than the vertical beam size.

RF Distribution

Baseline

The baseline choice is a TDR-like distribution system that includes a circulator in each cavity feed followed by a tuner (three-stub or E-H type) to allow control of the cavity phase and Qext. Currently DESY uses off-the-shelf components for the distribution system: a customized, integrated design would likely yield significant cost savings.

Alternatives

The circulator is the largest cost component at about 25% of the rf distribution cost. There are several alternative distribution schemes that eliminate the circulators but require more precise

Rev. July 13, 2006

Main Linac 40

cavity-to-cavity phasing, and make it harder to deal with the variation in the maximum cavity gradients. Also, in the event of rf breakdown in a coupler or cavity, these schemes would allow some fraction of the reflected power to propagate to the other cavities. How this additional power would effect performance has yet to be determined. KEK plans to test such a distribution scheme at the STF in the next year.

Cryogenic System

Baseline

The basic layout choice is that outlined in the TDR except the refrigerator spacing would depend on the choice of operating gradient, expected cavity Q and the desired cooling overhead (the spacing is about 5 km in the TDR – cryo-engineers at FNAL will recommend a spacing for the BCD). The length of the 2K, two-phase lines depends on the tunnel slope. For slopes up to 0.3 mrad (e.g., segmenting the machine into about 10 ‘flat’ regions), the 167 m long, 8.5 cm diameter lines specified in the TDR could be used. Larger slopes would require shorter lines, and for values in the 1-4 mrad range, a canal-like system would be used (for reference, a laser-straight tunnel would have a 3 mrad maximum slope). The maintenance length (i.e., the length that would need to be warmed up to repair a cryomodule) is half of the refrigerator spacing.

Alternatives

Thermal cycling long strings of cryomodules will be slow and may cause vacuum leaks. To reduce the maintenance length, U-tubes or turnarounds can be included at periodic locations to allow one section be thermally isolated (the refrigerators on either side of this region would cool the other cryomodules). If such sections (each 1.5 m long) were installed every 500 m, the number of cryomodules thermally cycled would be reduced by a factor of five, the warm-up time would be reduced by a factor of two and the cool down time would be reduced by a factor of 10.

Supporting Documentation

bcd:main_linac:ilc_bcd_cryogenic_chapter_v3.doc

Tunnel

Baseline

Relative to a linac with bends, a laser-straight linac would make dispersive emittance control easier and reduce the likelihood of off-energy beams intercepting the accelerators. However, this choice would limit the possible sites, especially those near the surface, and would require that the two-phase He be distributed along the cryomodules in more costly manner.

Until on-going beam dynamics simulations show otherwise, the linac will follow the curvature of the earth, unless a site-specific reason (cost driven) dictates otherwise.

Rev. July 13, 2006

Main Linac 41

Alternatives

Laser-straight linac or one constructed from straight-line segments. Final choice depends on cost considerations, primarily for the cryogenics and any site constraints, and the results of further beam dynamics studies.

Required R&D

Aggressive beam dynamics simulations of emittance preservation (beam-based alignment) for the continuously curved linac, including tolerance studies.

Studies of viable and cost effective cryogenic solutions for a tilted linac.

Baseline

The baseline choice is for the rf sources to be located outside of the beam tunnel so they would not be subject to radiation and could be accessed for repairs while the machine is running. To minimize rf power losses and cable runs, the sources are to be distributed along a second tunnel (or surface gallery) that runs parallel and nearby to the beamline tunnel. The rf power is transported into the beamline tunnel through three WR650 waveguide runs from each rf unit (one waveguide per cryomodule).

Alternatives

For cost savings, the TDR design is a reasonable choice. The beamline is in a near-surface tunnel (< 30 m deep) and the modulators, sans transformers, are clustered in surface buildings located every 5 km (the beamline tunnels contain the modulator transformers and klystrons). With the transformer in the tunnel, only relatively low voltage pulses (~ 10 kV) need to be transported, and the required cable impedance is typical of that used commercially (although four cables would be used per modulator). The disadvantage of separating the modulator from the klystron is that the cables are expensive ($70 million in the TDR), they pose a fire risk and they are not easily repairable once the spares are depleted.

For either the baseline or alternative choice, locating the beamline near the surface would allow easier access and shorten the power and cooling distribution lines that connect to the surface. The main disadvantages in that case would be larger ground motion and limited site availability. The final choice in this regard will likely be depend on the surface terrain and the population density at the proposed sites.

Rev. July 13, 2006

Beam Delivery 1

7. Beam Delivery

Beam parameters

The present BDS design is being evaluated for various ILC parameter sets [Raubenheimer_20050228]. The nominal parameters for 500GeV and 1TeV CM are acceptable. However, the parameter sets which have large beamstrahlung, may turn out to be problematic  from the point of view of extraction line energy acceptance and from background (pairs hitting vertex) point of view. In particular, the high Lumi 1TeV parameter set is not working and an alternative set was suggested [Seryi_20050817]. This alternative set pushes the vertical emittance and thus affects DR & LET working groups. Similarly, an alternative parameter set for high L 500GeV CM could be suggested. Some other parameter (e.g. Low P) may have the same problems with high beamstrahlung and associated power loss in extraction line (is being evaluated). In terms of the effect of different parameter sets on background, the low Q option is most preferable (but may be a concern for DR working group, as it pushes number of bunches), while the large Y size, low P and high L are much less preferable, especially at 1TeV. In particular, these three sets seem to exceed the tolerance on hits of the first layer of vertex detector (assuming CCD technology -- this may affect choice of technology for vertex detector) [Kozanecki_20050824].

    The maximal energy to which the hardware, layout, design of BDS beamlines (including in particular the extraction lines and SC quad) is specified, need to be defined. This is important, in particular, because higher energy can be reached in the linac with reduced current, provided that cavities can withstand increased gradient. Should this max energy be strictly 500GeV, 1TeV, or 35/31.5*1TeV? This affects, and feedback expected from, other working groups, in particular Parameters.  {This is response to GDE#2}

Required R&D    -- continue studies of various parameter sets, including alternative sets, on BDS and detector performance     -- develop alternative high Lumi parameter set for 500GeV CM    -- study how the limit of max operating energy is affecting BDS design optimization 

Overview of baseline

The recommendation for the BDS baseline have been chosen taking into account the recommendation of the particle physics community [ILCSC_scope], which requested that  ILC will have two detectors, which could possibly focus on different physics programs, and allow for different approaches to the search for new physics. The paradigm of two IRs (and, independently, two detectors) is being recently revisited, and is being discussed by the whole community. Possible configuration for the case of a single IR will be discussed below.

Baseline Configuration: Single BDS, 14mrad, 2 detectrors, single IR hall at z=0

Rev. March 12, 2007

Beam Delivery 2

The present baseline configuration consists of one Beam Delivery System with a beam crossing angle of 14mrad and a single IR hall at z=0 which houses two detectors to be placed at the interaction point for experiments in a push-and-pull fashion.

Elimination of one of the Beam Delivery Systems from the previous baseline configuration means elimination of both the civil construction and beamline components for one electron beam delivery and one positron beam delivery systems. This is motivated by the desire to reduce a substantial fraction of construction cost. From a technical standpoint this is supported by the perceived prospects for achieving a rapid switch-over of two detectors at the interaction point which is felt feasible when focused engineering efforts are made in the engineering design phase of ILC.

The beam crossing angle of 14mrad is chosen on the basis of the technical issues and a significant cost increment that is expected if a smaller crossing angle were chosen in the present status of beam optics design. More discussion on this topic appears in subsequent Alternative Configurations.

This baseline configuration assumes on-surface detector assembly (which saves 2-2.5 years). Details are given below in the section “Detector assembly”.

It is expected that the two main linacs are to be built at a relative angle of 14mrad.

Alternative Configuration: two BDSs, 14/14mrad, 2 detectors, single IR hall at z=0    This configuration consists of two Beam Delivery Systems with identical crossing angle of 14mrad, two detectors, and a single IR hall at z=0. This was in the past introduced as a baseline configuration which allowed cost saving and performance improvement [Vancouver1414] in comparison with previous baseline (two IRs with 20/2mrad crossing angles), but is presently maintained as an alternative configuration, since the singleBDS, single IR, two detecrtors, 14mrad crossing angle is chosen as the latest baseline. The 14/14 configuration, as opposed to 20/2mrad, is chosen to reduce the cost and improve the performance. A large fraction of the cost saving comes from the extraction lines due to reduction of the cooling capacity needed for 14mrad extraction line in comparison with 2mrad extraction line, reduction of the quantity and the cost of the magnets and their power supplies, reduction of the length of the vacuum system, reduction of the number of the extraction line collimators and photon beam dumps. A fraction of reduction also comes from removing the stretches in the optics, reduction of the length of the tapered tunnels in the beam switch-yard region, and savings due to common collider hall. As for the performance, the proposed baseline will have improved radiation conditions in the last part of the beamline, due to smaller beam losses in the extraction line; better performance of downstream diagnostics; much easier design and operation of extraction optics and magnets; reduced back-scattering from the extraction line elements and possibly better overall background; etc.

Rev. March 12, 2007

Beam Delivery 3

The configuration assumes on-surface detector assembly (which saves 2-2.5 years). Details are given below in the section “Detector assembly”.

It is expected that the two main linacs are to be built at a relative angle of 14mrad.,

Issues with the 20/2mrad configuration can in part be attributed to its relatively immature design of the beamline to achieve the 2mrad part of its layout. It is important to stress that we need to make sure that the design efforts and R&D studies for the 2mrad to continue since a configuration with a smaller crossing angle has certain merits from the technical and the physics analysis standpoint.

Alternative Configuration: two BDSs, 20/2mrad, 2 detectors, 2 longitudinally separated IR halls    {GDE#5 & GDE#15}    An alternative configuration for the case of two IRs consists of two Beam Delivery Systems with crossing angle 20mrad and 2mrad, two detectors, two independent and longitudinally separated IR halls [WG4_Snowmass_Summary]. This used to be the initial baseline configuration for layout of BDS. However, presently, it is maintained as an alternative configuration, first superceded by the two BDS, 14/14mrad and then by the single BDS, 14/14mrad configuration.

The large crossing angle IR features stable and mature design, separate incoming & extraction beamlines which allow achieving high luminosity, potentially cleaner downstream diagnostics, expect to provide good operational margins and flexibility, minimizing the risk to achieve nominal parameters, is upgradeable for gamma-gamma, but has somewhat larger backgrounds. The small crossing angle IR is a more recent design, it provides better background and better detector hermeticity, relies much less on crab crossing technique, but achieve lower luminosity than other IR, the downstream diagnostics may have higher background, the design is more constrained and operation may be more difficult.     Longitudinal separation of collider halls in this alternative configuration (dZ about 130m) provide possibility to build or upgrade one detector while another is taking data. This decision may affect other groups. In particular, with collider hall separated longitudinally and with undulator e+ source, there may be difficulties providing collisions at both detectors with different ( /2 or *2 ) time separation if the fast (train to train) interleaved operation is considered. However, assuming that fast interleaved operation is excluded, and that DR need to have turn-around for feed-forward, it should be possible to provide collisions in both IRs regardless of train structure (this may need building additional turn-around beamlines). {GDE#17}    The linacs in this alternative layout are pointing to the large crossing angle IR. This is configuration choice which does not a priori preclude multi-TeV upgrade. This may be a necessary but not sufficient condition to provide multi-TeV compatibility. Other requirements will be discussed below in the section "Multi-TeV issues".

Rev. March 12, 2007

Beam Delivery 4

    Design of the optics for this alternative is in advanced stage. Optics files for both IRs, for incoming and extraction beamlines are available [BDS_optics]. Geant or equivalent models of IR regions are being developed [BDS_IR_models].     The R&D which a specific to this alternative configuration choice are briefly the following. The compact SC quads and crab cavity for 20mrad IR, and large aperture SC quads and special extraction septum quads for 2mrad IR. The compact SC quads are being developed at BNL and recently demonstrated and exceeded the design gradient with 38cm short prototype [Parker_20050816], while the work on a longer prototype and preparation for stability study of such quad are under way. The crab-cavity design will build up upon the 3.9GHz deflection cavity being developed at Fermilab. Development of large aperture SC magnets within LARP program and at Saclay are relevant for 2mrad IR design [Napoly_20050816], while r&d on extraction magnets for 2mrad IR yet has to be started. (See more details on these r&d in sections "R&D specific to..." as well as in "IR magnets",  "Crab cavity systems", etc).  

Discussion of other alternative configuration candidates    {GDE#5 & GDE#15}This section presents some discussions on prospective configurations for the BDS layout .

The earlier design of head-on scheme, where a number of problems were identified by the TRC study [ILC_TRC2], has been reconsidered with attempts to improve it. In one case the electrostatic separator was suggested to be replaced by an rf-kicker [Iwashita_20050818], which however introduced severe MPS concerns and other issues. In another case, the extraction optics was modified so that the kick required from electrostatic separator was reduced about twice [Keller_20050818], improving its feasibility even at 1TeV. There are number of issues with this design: full design of extraction line is absent; it is not clear if downstream diagnostics would be possible; requirements on pressure (~1nTorr) in the separator are tight; radiative bhabha’s are hitting the separator plates; there are parasitic bunch crossing. Would these issues be solved, and required r&d successful (see below) the head-on configuration may give somewhat simpler design of the forward region and somewhat smaller background; simpler FD and extraction magnets; absence of the need for crab crossing. At Snowmass, the group did not achieve consensus whether the issues outweigh possible benefits, therefore the head-on configuration remained an alternative choice to be studied in more details. The head-on case was further studied and an updated design was presented at EPAC2006. In this case, the optimization focused on 500GeV CM energy, which allowed shortening the FD and the separator so that the beam separation at the parasitic collision is increased sufficiently to avoid the instability for the nominal bunch separation of 307ns. The extraction is done with help of 0.5mrad kick from electrostatic separator (combined effect of electric and magnetic fields), assisted by the 1.8mrad kick from the defocusing quadrupole QD2A of the incoming beamline, and the following 5mrad kick from a dipole. The beamstrahlung photons are collimated at several places, pass through large aperture bends B1 and B2 (full gap 12-15cm) of incoming beamline and is absorbed at the beamstrahlung dump 320m from

Rev. March 12, 2007

Beam Delivery 5

IP. A complete optics of the extraction line with downstream diagnostics was put together. Further analysis have shown that the off-energy tail of the beam is considerably overfocused and would need to be collimated up to dE=-10% quite early in the extraction line. Feasibility of such collimation and of the 1TeV CM design needs to be studied.     The additional r&d needed for head-on alternatives include feasibility study, prototyping and beam tests of electrostatic separator or rf-kicker. This may require several years. The r&d for alternatives need to be compared with r&d needed for baseline. In particular, the large bore SC quads for head-on may be similar as quads for 2mrad. See more in "R&D specific to...".

 

 

Ranking of BDS configurations – 20mrad, 14mrad, 2mrad, and head-on (with reduced strength electrostatic separator and with rf kicker) – for various criteria

Ranking example: "best A, B and C, then D, worst E". In this case all A,B,C have the same high rank, D would be intermediate rank (separated by "then"), and E would be the worst rank.

Rank 1 – directly affecting energy and luminosity reach, background and precision measurements of beam properties, or a single point failure:

Luminosity reach – best 14 and 20mr, worst 2mr and head on o In 2mr and head-on, to extract the disrupted beam, it is bent by a separator, rf

kicker or field off-center of the final quadrupole. Large energy spread of disrupted beam causes beam losses and limits the luminosity reach by more than a factor of two in comparison with 20 and 14mr

Crab-crossing – best head-on, then 2mr, then 14mr, worst 20mr o No need for crab cavity in head-on o Small to moderate luminosity loss (5%, 10% or 30% for low Q, nominal or large

sigma Y parameters) in 2mr without crab cavity. Full recovery requires crab cavity that would need to be placed a km away from IP which may be an issue for phase stabilization, however the needed kick and phases stability are ten times more relaxed than in 20mr. Partial luminosity recovery may be done with dispersion at the IP, without crab-cavity, for the cost of larger energy spread in the beam and induced energy bias of luminosity spectrum

o Crab cavity is essential for 14 or 20mr. Luminosity loss without crab cavity is 60-75-90% in 14mr and 75-85-95% in 20mr (for low Q, nominal or large sigma Y parameters)

o Warm NC transverse cavities are in use now, SC cavities are not yet. A deflecting SC CKM cavity is being built at FNAL. Crab cavity system can be built and experimentally verified during TDR phase, before start of ILC operation. 

Rev. March 12, 2007

Beam Delivery 6

Fast feedback hardware and its integration into IR  – best 20 and 14mr, then head-on, worst 2mr

o In 20 and 14mr, feedback BPMs and kickers do not see other beam o In head-on with shared aperture, BPM sees other beam and need to be directional,

there may be losses of low energy beam tail on the kicker. o In 2mrad, the feedback BPM has to be placed in front of FD, where disrupted

beam envelope is still small, there is offset of outgoing beam in the BPM, the kicker should be large aperture, there are potential losses on the kicker.

o Performance of IP feedback, with all effects of beam losses included, is difficult to guarantee from simulations (which have advanced significantly) or from simplified beam tests. Eventual verification cannot be done before start of ILC operation.

Hermeticity & min veto angle – best head-on and 2mr, then 14mr, worst 20mr o The configuration with separate apertures for incoming and outgoing beams has a

larger un-instrumented area in the forward region by comparison with a single aperture.

o The minimum accessible veto angle is defined by the radial extent of the aperture required for the exiting beam (which is determined by the collimation aperture, L* and the strength of the solenoid field) and can be made equal regardless of the crossing angle choice.

o The detection efficiency for small angle electrons from copious 2-photon processes is reduced for crossing angle configurations with two holes, which results in higher backgrounds for interesting (SUSY) physics processes whose signature in the detector is missing energy. This loss of efficiency can be compensated by higher luminosity.

Pairs background – best head-on, 2mr and 14mr, worst 20mr o In small angle cases the pairs produced in beam-beam collisions are not dispersed

by crossing the solenoid field and background is smallest o In larger crossing angle cases pairs traverse solenoid field and disperse more. In

optimized 20mr configuration the background can be two times higher than in 2mr or head-on

o Use of DID increases background, while anti-DID reduces background o In 14mr the effective crossing angle can also be zeroed for outgoing beam with

use of anti-DID, and background can be as good as in head-on or 2mr Flexibility of extraction optics and possibility of downstream diagnostics – best 20 and

14mr, then 2mr, worst head-on o Extraction optics is independent of incoming optics and downstream diagnostics

design in the 20 and 14mr options o Extraction optics in 2mr depends on shared FD magnets, in particular relying on

focusing arising from going through final sextupole SD0 with an offset. Design of downstream diagnostics exists. Further studies have shown that downstream diagnostics in both 20/14 and 2mrad extraction lines seem to be possible at 500GeV CM. Polarization precision of 0.25% seems achievable. There are higher beam losses and SR in 2mrad. The SR losses in 2mrad vary with beam conditions and for 200nm horizontal offset at IP the losses vary by 25MeV which is close to

Rev. March 12, 2007

Beam Delivery 7

the required accuracy of energy measurements of 100ppm. At 1TeV CM the much higher SR losses in 2mrad extraction make downstream diagnostics extremely challenging.

o Extraction optics in head-on depends on shared FD magnets and QD2A quad. Early focusing of disrupted beam by FD sextupoles, as in 2mr, cannot be done in head-on. Investigation of feasibility of downstream diagnostics remain to be done. Further studies confirmed this assessment. A first version of extraction optics (designed for up to 500GeV CM only) with downstream diagnostics was created. Due to over-focusing by FD, the chromatic variation of the beam size and orbit is very large, and early collimation of the beam tail (below 90% E0) is needed. Feasibility of such collimation and other design optimization need to be studied.

Losses and background conditions in downstream diagnostics – best 20 and 14mr, then 2mr, worst head-on

o Losses are smallest in 20 and 14mrad extraction line o In 2mrad extraction line, need to collimate several tens of kW in vicinity of

downstream diagnostics o For head-on, there is no design of downstream diagnostics at the moment. First

version of design exists, assessment on the losses is correct.  Losses in extraction affecting IR background - best 20 and 14mr, worst 2mr and head-on

o In 2mrad and head-on extraction line, there are higher losses which may create, e.g., neutron flux to detector. Losses of FD generated SR at a surface of extraction line magnet create a flux of backscattering photons which exceeds by several times the flux created by pairs. Possibility of mitigation of this issue need to be studied.

Rank 2 – may affect energy, luminosity and background indirectly, e.g. via reliability of operation (integrated luminosity):

Parasitic crossings – best 20,14,2mr, worst head-on o For bunch spacing considered for ILC, there are no parasitic crossings for

20,14,2mr o For head-on with electrostatic separator, there are parasitic crossings,

especially an issue with halved bunch spacing Vertical orbit correction in IP – best head on and 2mr, then 14mr, worst 20mr

o Vertical orbit is flat or negligible in head-on and 2mr o Vertical orbit through detector solenoid is nonzero, and IP vertical angle is 70-

100 micro-radians in 14-20mr (relative vertical angle between e+e- is still zero).

o If required for accurate polarization measurements, the vertical angle can be compensated with DID field or less locally with FD correctors, without DID.

o If the vertical IP angle is not corrected, one can also tilt the beam orbit in downstream polarimeter, to match the IP angle.

o In case when angle at IP or in downstream polarimeter will be corrected, such correction is energy dependent (if DID or anti-DID is used, its setting is independent on energy, and proportional to detector field). 

Rev. March 12, 2007

Beam Delivery 8

Tracking, in particular TPC operation and calibration – best head on and 2mr, worst 14 and 20mr

o If DID or anti-DID is used, TPC calibration may be more complex. The calibration procedure can be eased with modified DID, which provides significantly reduced transverse field in the center (|z|<0.5m) of the detector

Radiation in solenoid field – best head on and 2mr, then 14mr, worst 20mr o Vertical trajectory deviation in detector solenoid is proportional to crossing

angle and synchrotron radiation effects grow as fifth power of the angle, independent on energy. The effect on the luminosity is less than one percent when crossing angle is less than 32mrad (SiD), 27mrad (GLD) and 21mrad (LDC).

Extraction line clearance for beamstrahlung photons – best 20 and 14mr, worst head-on and 2mrad

o The aperture opening for beamstrahlung photons is +-1.25mrad in 20 and 14mr IR, while it is +-0.5mrad in 2mrad, leading to higher losses on photon collimators and more stringent steering constraints. In head-on with separator, the photons collimated to 0.38mrad at 80m from the IP.

Photon losses in FD, direct sight to vertex – best 20,14 and head-on, worst 2mr o In 2mr, photons from beam halo hit poles of warm quad QF1 in FD. There is a

direct sight to vertex detector from QF1 poles. For certain assumptions about beam halo, photons and neutrons backscattered from QF1 to vertex contribute about 10% in comparison with beam-beam pairs

Extraction devices affecting MPS – best 20,14, worst 2mr and head on o In 2mr, because of the large size of outgoing beam, there is difficulty to locate

protection collimator on both sides of FD in its vicinity o In 2mr, quench of QD0, which provides extraction angle, may cause the beam

to hit near IR beamlines o In head-on with electrostatic separator, a breakdown would cause beam to hit

collimators on another side. o In head-on with RF kicker, its failure may cause the beam to hit the IR region.

The dark current from parasitic rf-buckets can be swept to IR region if BDS energy collimator is opened to more than 2% energy acceptance. 

Extraction devices affecting collision stability – best 20,14 & 2mr, worst head-on o In head on with separator, its field needs to be stable at the level of 1E-5, to

prevent beam separation at the IP o In head on with rf kicker, its field needs to be zero within 5E-6 at the moments

of passage of incoming bunches

 Rank 3 – affecting only cost, difficulty of r&d and of the design:

Difficulty of final doublet magnets – best 20 and 14mr, then head-on, worst 2mr o The 20 and 14mr SC magnets are based on direct wind technology which was

used for HERA upgrade, for BEPC upgrade, etc. o The head-on SC FD quads may benefit from LARP and Saclay multi-year

programs which may come with a prototype in 2006

Rev. March 12, 2007

Beam Delivery 9

o The 2mrad SC FD magnets must have larger aperture than for head-on, especially for sextupole. The 2mrad QD0 need to have 3mm thin tungsten liner, to reduce the energy density due to radiative Bhabhas.

Length of extraction line – best 20 and 14mrad, worst 2mr and head on o Min length of extraction line defined from min transverse separation from

incoming beam is about 400m in 20 and 14mr and about 800m in 2mr and head-on

Difficulty of final doublet integration in detector – best 20, 14mr and head on, worst 2mr

o The external size of FD cryostats with supports fits to r =20cm envelope in the detector for 20,14mr and head-on

o For 2mr, with larger aperture magnets, the external size will be larger and the envelope radius will be increased to about  r=25cm or more

Special extraction magnets – best 20 and 14, then head on, worst 2mr o The 2mr require special septum quads with large aperture for extracted beam

and zero field region for incoming beam. Design work on magnets has shown that it is difficult to design and fit especially the QEX1-3-5 quadrupoles in the places where the space is limited by the near-going photon beam and B1/B2 magnets of the incoming beam.

Special coils for detector solenoid – best 2mr and headon, worst 14 and 20mr o DID or anti-DID coils need to be included in detector solenoids for 14 or 20mr o An accurate 3D field map with (anti-)DID must be measured, for several

settings

 Special Rank  – compatibility with other physics programs and upgrades (Relative weight of this category should be discussed and determined by the whole community):

Compatibility with gamma-gamma – best 20mr, then 14mrad, worst head-on, 2mr o The +-10mrad extraction cone acceptance for can be provided in 20mr

for L* of about 4.5m (with latest self-shielding quads). If displacement of FF beamline and modifications of tunnels is considered, both 20mrad and 14mrad IR can be reconfigured for 25mrad IR for gamma-gamma. (Changes needed for 14mrad are larger than for 20mrad, however, 14mrad has an advantage of better separation for the gamma-gamma dump.) Smaller angle is not compatibility with

Compatibility with e-e- – best 20 and 14mr, then head-on, worst 2mr o The 20 and 14mr can provide same beta-functions at IP, collision angle can be

corrected o In head-on with electrostatic separator, a defocusing quad is relied upon to

increase the extraction angle. In e-e-, its polarity needs to be changed, leading to modification of FF optics. Impact of this change was not studied.

o Most difficult for e-e- is 2mrad, where, to preserve focusing of outgoing beam by FD sextupoles, the polarity of FD need to be changed, leading to thirty times higher Y beta function at IP 

Rev. March 12, 2007

Beam Delivery 10

Compatibility with multi-TeV – best 20mr and 14mr, worst head on and 2mr o Parasitic crossings in CLIC-like scheme require crossing angle of 20mr and

14mr may be also fine (with new CLIC parameters, TBC) o Small crossing angles not compatible with multi-TeV, because of parasitic

beam-beam crossings and because one cannot separate the outgoing beam

R&D specific to 20 and 2mrad alternative, 14/14mrad baseline and head-on alternatives

Common items are not listed (see subsystems chapters below). Items listed mean detail engineering & prototyping, except when it says “study, evaluate or design” which means paper & engineering study. Text in () gives present status. If parenthesis () are absent, the activity still need to be started. Ranking of the R&D items (e.g. as TRCs R1-R3) is planned to be done.

20mrad alternative and 14mrad baseline:

Prototype compact SC quads (short BNL prototype tested, self shielded quad tested) Crab cavity system and phase stabilization (3.9GHz FNAL cavity developed, redesign

coupler to open aperture) Study integration of DID into solenoid (preliminary consideration for SiD done

[Smith_Wands_2005]) Study vertical orbit correction & polarimetry (preliminary considerations done) Study TPC calibrations procedures with DID Stability of compact SC FD (preparations at BNL) Design crossing angle IR with all acc. (bpms, kickers, supports) and mdi (lumi,

beamcal) components, vacuum chamber w. assembly considerations Study performance of downstream diagnostics in presence of beamstrahlung cone

 14mrad design baseline:

All items are same as for 20mrad, plus Integration of final quads into common cryostat (eng. studies at start)

 2mrad alternative:

Large aperture SC quads (r=35mm), may use NbTi for long L*, likely need Nb3Sn for L*=3.5m and E upgrade (LARP and European programs for LHC ongoing, design must be accommodated to ILC requirements)

Large aperture SC sextupoles (r=90mm) Design tungsten liner into the QD0, to reduce energy density due to radiative Bhabhas Study integration of large aperture quads and sextupoles within the detector Pocket coil warm quads QF1 (first look done) Evaluate neutron and photon backscattering due to synchrotron radiation (produced

upstream and by bending in FD) and charged halo particles impinging on apertures of FD and extraction line, in particular on the face of QF1

Rev. March 12, 2007

Beam Delivery 11

Special extraction septum quads Crab cavity system and phase stabilization (less critical than for larger crossing angle)

(3.9GHz FNAL cavity developed, redesign coupler to open aperture) Phase stability of crab cavities separated by a km Study consequence of 0.5mrad photon aperture (1.25mrad in 20mrad) Study possibilities to place protection collimators in FD and MPS issues Study performance of downstream diagnostics in presence of tens of kW losses on

nearby collimators Feedback kicker inbore of large aperture sextupole or near Study effect of losses in extraction beamline to IR background Stability of larger aperture SC FD Design straight IR with all acc. (bpms, kickers, supports) and mdi (lumi, beamcal)

components, vacuum chamber w. assembly considerations Design region of in, out and photon beam separation with all accelerator and vacuum

components

 Head-on design alternative (with electrostatic separator, reduced strength):

Large aperture SC quads (r>24mm), may use NbTi for long L*, likely need Nb3Sn for L*=3.5m and E upgrade (LARP and European programs for LHC ongoing, design must be accommodated to ILC requirements)

Large aperture SC sextupoles (r>24mm?) Study parasitic collisions at nominal and upgrade cases Study possibility to shorten the FD to increase separation at parasitic collisions Electrostatic separator feasibility and stability (CERN separators in ILC requirements,

beam tests?) Beam, photons & radiative bhabha’s losses on separator plates and protection

collimators (plan beam test at ESA?) Low vacuum (1nTorr) in electrostatic separators Extraction septum quads Design extraction line, study feasibility of downstream diagnostics

 Head-on design alternative (with rf kicker):

List is similar as for previous case of head-on with separator, where the electrostatic separator items replaced by:

RF kicker feasibility and stability (tests at Kyoto so far show too low Q factor, to be continued)

Study MPS issues with RF kicker (to be continued) Study possibility of full elimination of dark current in parasitic rf buckets (by

collimation below 2% at 250GeV beam, 1% at 500GeV beam)

Rev. March 12, 2007

Beam Delivery 12

Overview of BDS and its subsystems

The following sections review properties of beamlines and systems such as feedbacks, beam dumps, and other, highlight baseline choices and existing alternatives, and list the required R&D.

Beamline sequence and design features   

The sequence of beamline sections in the baseline optics is the following: linac, beam emittance diagnostics and coupling correction section, tune-up and emergency extraction beamline, beam switch yard, upstream polarization diagnostics section, betatron collimation, energy collimation, upstream energy spectrometer, final focus proper with secondary clean-up collimation and with tail-folding octupoles, the final doublet, extraction beamline with downstream energy and polarization diagnostics, beam dump. The final focus optics is with local chromaticity correction {GDE#40}. The range of L* (distance from IP to final quadrupole) considered for studies is 3.5 to 4.5m {GDE#34}. The betatron collimation precedes the energy collimation so that off-energy debris from betatron collimation could be cleaned out in the downstream energy collimator {GDE#37}. The tail-folding octupoles are included into baseline and allow to open the collimation gaps by about a factor of three, which provides an additional safety factor, but they are not relied upon in the baseline. {GDE#35}   

The considered BDS baseline assumes that there is no vertical angle between linac and BDS and the nearest region of vertical bends is at least one kilometer away from beginning of BDS. This is a safety factor that would ease extension of BDS if, for example, additional collimation sections found to be necessary {GDE#4}.    

The baseline assumes that there is no undulator source at the end of the linac, before BDS. The recommendation by WG3a to place the undulator source at the very end of the linac contradicts the considered BDS baseline. Would it indeed be placed there, the number of adverse effects and integration issues will arise (see snapshot of discussion here [BDS_mtg_20051101]), requiring detailed studies. {GDE#8}R&D planned at the ATF2 facility will give opportunity to gain experience with local chromaticity correction optics, instrumentation and other BDS aspects.

Tune-up extraction line, MPS, E-error and betatron error diagnostics

Updated November 9, 2005.

The beamline described in this section is located at the end of the main linac at the entry to the BDS and serves several purposes:

provide MPS protection for BDS components with a system of renewable/consumable collimators

measure the beam emittance using a system of laserwires

Rev. March 12, 2007

Beam Delivery 13

provide for correction of transverse coupling using a system of orthogonal skew quadrupoles

monitor incoming betatron orbit errors using BPM measurements; monitor incoming energy errors in a chicane

when necessary, abort the beam to a dump using a system of kickers and a high-bandwidth extraction line

during linac tuning, send the full beam to the extraction line continuously using pulsed bends

Corresponding to these goals, this beamline is subject to the following design choices:

minimum allowable beam size at the laserwire locations and properties of the laser system; required precision of emittance measurements

maximum allowable amplitude of incoming betatron orbit errors which should be handled by the kicker system

maximum number of errant bunches which can be allowed into the BDS which could hit the betatron spoilers

maximum number of errant bunches which can be allowed into the BDS which could hit the energy spoiler

Other design choices, such as how to detect the signals from the laserwires (scattered photons or electrons), and how much emittance dilution from synchrotron radiation can be allowed in the chicane, will also influence the layout and design performance of the system.     One of the most critical R&D for this beamline is the laserwire [Blair_Snowmass_20050817]. The minimum beam size that can be measured with the laserwire (the size of the laser spot at its focus) has a direct impact on the length of the diagnostics section. For example, if a 3micron laserwire is used, the beam size at the laserwires should be not less than 3microns, which drives the length of the diagnostics system to several hundred meters [Woodley_20050927, Nanobeam_Blair_20051018, Nanobeam_AngalKalinin_20051018]. The required accuracy of emittance measurements and the beam emittances for which the system is designed (250GeV or 500GeV/beam) are obviously very important questions here.     To limit the diagnostics section to a reasonable length, the baseline would be based on one micron laserwires. This will require higher laser power, frequency tripled light and f/1 optics. The length of the diagnostics section is chosen to be 100m with 65m maximum beta functions so that an emittance measurement accuracy of <5% would be expected for a 500GeV beam with normalized emittance 2e-8m-rad. The unattractive alternative is to lengthen the diagnostics section considerably. Signals from the laserwires can be observed either by detecting Compton-scattered photons in the chicane or by detecting degraded electrons near the beampipe somewhere downstream.    The maximum amplitude of betatron error defines the power of the kicker, which scales as the fourth power of the betatron amplitude [Nanobeam_Mattison_20051020].  The apertures

Rev. March 12, 2007

Beam Delivery 14

in the main linac and BDS are quite different -- r=35mm in the linac and typically r=10mm in BDS. It is important to note that probability of very large betatron errors is very small but never zero. With this in mind, the BDS baseline incorporates a passive consumable (survivable to be investigated) MPS collimation system that would limit the amplitude of betatron errors incoming to BDS to approximately r=8mm. This system would be integrated into the emittance diagnostic section, perhaps requiring large apertures for the skew correction magnets which are immediately upstream.    The number of errant bunches allowed to get through to the downstream BDS systems before the kickers reach full strength is defined by the survivability of the betatron and energy spoilers, which will be able to survive up to two bunches of 2e10 particles at 250GeV/beam. The design goal for the emergency extraction system is to allow no more than two errant bunches through. Thus the rise time of the kickers should be about 100ns, somewhat shorter than the inter-bunch spacing.

Required R&D     -- Develop laserwire with one micron spot-size, tripled light, f/1 optics    -- Study detection of laserwire signal with scattered photons or electrons    -- Design a system of consumable (renewable?) devices to limit the amplitude of betatron orbit errors    -- Design a system to reliably determine betatron and energy errors

Collimation and Backgrounds

Section updated September 19, 2005. Primary authors responsible for this section: N.Mokhov, F.Jackson

Overview

The baseline collimation design for the ILC BDS is an adaptation of the NLC scheme [LCC52,LCC111]. The design implies a betatron collimation section followed by energy collimators. Octupole pairs are located to perform beam-tail folding which relaxes collimation requirements {GDE#35}. The collimation system [Mokhov_Conf_05_154] consists of spoiler/absorber pairs, arranged to survive impact from errant bunches which escape the machine protection fast extraction system. Additional protection collimators are located elsewhere in the BDS providing local protection of components and absorption of scattered halo particles, while synchrotron radiation masks in the immediate vicinity of the interaction point (IP) protect the collider detector components. The BDS crossing angle does not greatly affect the design of the collimation system, but may impact issues affecting the collimation depth.

Betatron and Energy Collimation Scheme

In the baseline design, degraded energy particles originating from betatron collimation section (and not absorbed there) may be collected by the energy collimator {GDE#37}.

Rev. March 12, 2007

Beam Delivery 15

Alternative ordering of betatron and energy collimation sections in this design has not been studied. However, a comparison has been done of the collimation designs of NLC and TESLA, which have opposite ordering of energy collimation and betatron collimation. In this study [LCC111] the NLC collimation performance was found to be superior.

Existing Status of R & D

The spoilers are 0.5 to 1 X0 (radiation length) thick, absorbers and synchrotron radiation masks are 30 X0, and protection collimators (PC) are 15 X0. The betatron collimation with 'survivable' spoilers included into baseline has advantage that these spoilers can withstand hit of two bunches at 250 GeV/beam, matching the emergency extraction design goal [Keller_20041015]. They can survive one bunch at 500 GeV. The survivable spoilers are more demanding for the optics (more difficult tuning, tighter tolerances). The alternative is to use consumable spoilers which ease the optics, but require more R & D in terms of renewable spoiler design, development of damage detection, study of MPS issues, etc.

The studies of the dynamic heat load show that most of energy among the beam line elements is deposited in the protection collimators PC1, PC5, PC8 and PC9. The residual activation and radiation damage in the magnets downstream of those is in excess of the limits if the length of the PCs is kept 15 X0 as originally proposed. For example, an averaged residual dose on-contact after 30 days of irradiation and one day of cooling on the front surface of the first quadrupole downstream of PC1 is as high as 7.7 mSv/hr compared to the limit of 1 mSv/hr. The absorbed dose in quadrupole coils reaches 300 MGy/yr compared to the coil insulation limit of about 4 MGy, meaning a lifetime of only a few days. This forced us to increase the PC's length to 45 X0 (about 60 cm of copper) resulting in the coil lifetime of at least several years. Dynamic heat load distribution obtained with the MARS15 code after the collimator optimization gives acceptable loads for the magnets, about 50 W/m for spoilers, and about 10 kW/m for the protection collimators PC1, PC5, PC8 and PC9. (The loss and radiation numbers correspond to conservative assumption of  0.1% for the beam halo population.)

Collimation depths (based on halo synchtrotron radiation clearance through IR apertures) have been studied for the current 20mrad and 2mrad final doublet designs  [Jackson_20050816,Carter_20050816]. Although the aperture constraints are different in each case (extraction quadrupoles for 20mrad, beam calorimeter for 2 mrad), the collimation depths are the roughly the same in both (~10 sigx in the x plane and ~60-80 sigy in the y plane). In the 20 mrad deck these correspond to betatron spoiler gaps of ~1 mm in the x plane, ~ 0.5 mm in the y plane.

The performance of the 20 mrad collimation system has been studied in halo tracking simulations [Jackson_20050816,Carter_20050816,Drozhdin_BDIR2005]. These demonstrate a reasonable performance of the collimation system. Some halo repopulation outside the collimation depth is evident, particularly in the x-plane, which can be remediated by reducing the spoilers x-apertures. The 2 mrad lattice has not yet been studied, or optimized for collimation yet.

Rev. March 12, 2007

Beam Delivery 16

Future R & D Required

The most important R & D subjects affecting the collimation baseline are as follows:

Development of a fast extraction system (tune-up extraction beamline) to provide the necessary protection for the collimators. Experimental studies of collimator survivability may be performed via material damage beam tests at the SLAC end-station-A [Woods_20050302].

Optimization of the current optics designs to minimize non-linear transport between the betatron spoilers and the final doublet.

Reduction of radiation loads at normal operation and beam accidents. Further study of the use of octupole tail-folding (examined previously for the NLC

[LCC118]), possibly with beam tests to verify the principle. (Initial studies of tail folding octupoles for ILC done, see [Seryi_20051102]).

Evaluation of wakefield effects in the baseline design(s). For previous BDS designs, predicted vertical jitter amplification and emittance dilution were found to be significant [TRC2_p340]. In addition, wakefield measurements are planned at the SLAC end-station-A to optimize collimator geometry, and improve analytical methods of estimating collimator wakefields [Collwake_beamtest_request]. (Initial studies of collimator wakefield in ILC done, see [Jackson_20051102]).

Benchmarking of simulation codes. To see how much our understanding of collimation performance is affected by simulation conventions and approximations.

Muon flux attenuation studies (see next section).

R&D for alternative with consumable spoilers    -- renewable spoiler design and prototyping     -- development of damage detection    -- study of MPS issues

Reduction of background fluxes at detectors

Muons

A lot of muons are generated - predominantly as Bethe-Heitler pairs - in electromagnetic showers induced in the collimators and other BDS elements during the beam halo cleaning. Fluxes of these muons accompanied by other secondary particles, could exceed the tolerable levels at the detector by a few orders of magnitude [Mokhov_Conf_05_154]. Calculated with the MARS code muon flux equals to 4.1 1/cm^2 s^-1, or 7600 muons in the tunnel aperture for 150 bunches from one beam (with halo population 0.1%). This is to be compared to a few muons allowed in such a sensitivity window. The mean energy of these muons is about 27~GeV. About 700000 photons and 200000 electrons accompany these muons at the detector. The fluxes doubles for energetic muons for two beams.

Magnetized spoilers sealing the tunnel would reduce muon fluxes substantially. MARS calculations were performed for two iron spoilers 9 and 18 m thick at 648 and 331 m from the

Rev. March 12, 2007

Beam Delivery 17

IP, respectively [Mokhov_Conf_05_154]. The square spoilers are extended by 0.6 m in the tunnel walls and dirt on each side. The field of 1.5 T is used in opposite polarity on left and right sides to compensate it at the beam pipe center. Central gaps are 10 cm wide and 1 m high with 0.8 T field. The gap between the parts is as the beam pipe. This set of spoilers reduces muon background load on the detector from 7600 to 2.2, i.e. to about an acceptable level. Other particle fluxes coming from the tunnel are also down in about the same proportion.

At the initial installation of the ILC, only the shortened 5 m muon walls will be installed in place of the 18 m wall, and the 9 m wall will not be installed. The 5 m wall also serves as a Personal Protection System, reducing the muon dose rate to people who are accessing the IR area while beam is delivered to the upstream tune-up dump or other IR. The caverns for all muon spoilers will be built with full dimensions, allowing future upgrade of the muon spoilers from 5 m to 18 m or the addition of the 9 m muon spoilers, if much higher halo population and flux of muons would be measured during ILC operation. It is estimated that such an upgrade would require a shutdown of about three months.

Two alternatives to the muon tunnel spoilers need to be investigated: muon attenuator (about 120m long collar at with 1T field, 0.6m OD) and wide-aperture magnets.

Future R & D Required

The most important R & D subjects in this area are as follows:

Build consistent, realistic BDS+detector integrated models, with detailed magnetic field maps, tunnel and experimental halls. Add engineering realism wherever possible.

Generate/refine sub-detector tolerance tables on three levels: (1) pile-up; (2) pattern recognition; (3) radiation damage.

Study backgrounds for 3 detector concepts and 2 crossing angles: (1) sensitivity windows with respect to tagged origin an bunch crossing; (2) tag origin of backgrounds for all particle types; (3) further explore mitigation methods such as better collimation performance, low-Z masks, muon tunnel spoilers, attenuators or wide-aperture magnets.

Code/model benchmarking; interfacing; BDIR code/model/map depository.

Extraction

(Text of this subsection may need to be moved to extraction section).MARS simulations confirm that synchrotron photons produced from the beam core and halo upstream of the IP are collimated by the photon masks and - with an appropriate design - their contribution to backgrounds and radiation loads to extraction line components is negligible. Same with beamstrahlung photons which form a very narrow beam. e+e- pairs and synchrotron photons generated by disrupted beam remain the main source of the IP backgrounds and radiation loads to detector, final focus and extraction components.

Rev. March 12, 2007

Beam Delivery 18

At high luminosity and 120-nm vertical offset total radiation load in extraction beam line is 13.3 kW with a 600 W/m peak. Without a vertical offset, these numbers are a factor of ten lower. This is to be compared to estimated tolerance levels of 10 W/m for superconducting magnets and a few hundred W/m for conventional magnets.

IR and IR magnets for 20mrad

Section updated December 2, 2005, with IR magnets description by B.Parker. Need to expand IR description. Primary authors responsible for this section: B.Parker, Y.Nosochkov, K.Tsuchiya, T.Mihara

The compact SC quads provide possibility to focus the incoming and outgoing beam independently, maximizing the luminosity performance and minimizing the losses of the disrupted beam. Expanded description of compact superconducting magnets for 20mrad baseline IR (or for ) is given in [Parker_20050915]. This includes description of BNL direct wind technology [BNL_direct_wind], details of engineering design, characteristics of the quads for incoming and outgoing beamlines, plans for stability study of such magnet, etc. A short 38cm prototype of SC QD0 has been built and successfully tested [Parker_20050816,Parker_20050915], exceeding the expectation for the design gradient.    

A self-shielded design of QD0 has been recently suggested, which allows reduction of crossing angle [Parker_20050818,Parker_20050823,Parker_20050913], and will be used for development of intermediate crossing angle alternatives. The cancellation of the external field with a shield coil has been successfully demonstrated in the recent tests at BNL.    

The 14mrad IR based on self shielded compact quad was designed, see section Intermediate crossing angle 14mrad.

Intermediate crossing angle 14 mrad   

Design of intermediate crossing angle (14mrad at L*=3.51m) has been presented in mid October at Nanobeam 2005. The design includes optics of incoming and extraction line, see [Nanobeam_Markiewicz_20051019] and [SLAC_PUB_11591], design of FD magnets [Nanobeam_Parker_20051018, Nanobeam_Parker_20051021], IR optics optimization and background simulations [Nanobeam_Seryi_IR_20051019], as well as considerations for civil engineering upgrade scenarios from single 14mrad IR to two IRs [Nanobeam_Seryi_Civil_20051019] with small or larger crossing angle of the second IR. The IR final quads are based on self-shielding concept [Parker_20050818, Parker_20050823] which eliminates the field interference between beamlines. The IR magnets are shown in Fig.1 and Fig.2 (taken from B.Parker summary [Nanobeam_Parker_20051021]) and are based on tested prototype, see Fig.3. The 14mrad crossing angle reduces the SR effects, allowing greater flexibility in optimization of vertical orbit and background, which now can be as low as in 2mrad IR, see Fig.4 and [Nanobeam_Markiewicz_20051019, Nanobeam_Seryi_IR_20051019]. 

Rev. March 12, 2007

Beam Delivery 19

20mrad Extraction Line

Section updated September 25, 2005. Primary authors responsible for this section: Y.Nosochkov

The 20 mrad extraction design [Nosochkov_20050817_20mrad] is based on the independent beamline for the spent beams, without shared FF magnets. The outgoing primary e+(e-) beam and the beamstrahlung photons are transported through the same extraction magnets to a shared 18.3 MW dump. The optics consists of the initial DFDF quadrupole system followed by the two vertical diagnostic chicanes and two protection collimators before the dump. The first four quadrupoles after IP will be superconducting followed by warm magnets downstream. A 2 m space after the last SC quadrupole is reserved to accommodate the crab-cavity on the incoming line. The extraction apertures and quadrupole focusing are optimized for a large energy acceptance to minimize the disrupted beam loss caused by overfocusing in the low energy tail. The magnet apertures are sufficient for a photon beam with up to ±1.25 mrad angle at IP. The optics provides the 2nd focal point with the required beam size of < 100 µm for the Compton polarimeter diagnostics [Moffeit_SLACPUB11322]. The extraction magnets are compatible with up to 1 TeV CM beams.

The first SC extraction quadrupole QDEX1 is placed side-by-side with the first SC incoming quadrupole QD0 at L* = 3.51 m after IP. This choice is based on the SC compact quadrupole design [Parker_20050915] which makes it possible to have independent SC coils with a small transverse separation. The QDEX1 has a low strength to limit its residual field on the incoming line and built-in correction coils to compensate for QD0 residual field on the extraction line. The first diagnostic chicane will serve as the energy spectrometer. It will include wiggler magnets to produce synchrotron radiation at ±2 mrad directions to measure the average beam energy using SR stripe detectors. The polarization measurement will be performed by a Compton polarimeter, with the Compton IP located at the 2nd focus at center of the second chicane. The horizontal angular amplification term R22 from the IP to the Compton IP is adjusted close to -0.5 for maximum sensitivity to the measured effects. A long 170 m drift is included before the dump to increase the undisrupted beam size. Further increase of this size is required for a realistic dump design [Walz_20050817]. This can be achieved by a longer drift and/or the use of beam rastering. The two collimators before the dump serve to limit the size of the disrupted beam and the beamstrahlung photons to the 15×15 cm size of the dump window [Walz_20050817].

It has been shown that the effect of detector solenoid on the extraction optics and beam loss can be compensated using dipole and quadrupole correcting coils on the first SC quadrupoles [Nosochkov_20050510]. The particle tracking [Nosochkov_20050817_20mrad, Drozhdin_20mr_extr, Ferrari_20050831] showed that the primary beam loss in magnets is acceptable in the 0.5 TeV and 1 TeV CM nominal luminosity options. The power loss in the 0.5 TeV CM high luminosity option and the two alternative 1 TeV CM high luminosity options [Seryi_20050817] may be acceptable, provided that the loss of ~500 W/m in warm magnets is acceptable. The 0.5 TeV CM high luminosity option will require a larger aperture in the 3rd and 4th SC quadrupoles to reduce loss to 2 W/m in these quadrupoles. The

Rev. March 12, 2007

Beam Delivery 20

maximum loss of the charged and photon beams on the collimators is 5 kW and 200 kW (per collimator) in the nominal and high luminosity options, respectively.

R&D (magnets): 1. SC compact quadrupole with correcting coils.2. Large aperture warm chicane bends.

R&D (optics): to be updated! Some work done and presented at Nanobeam, see

[Nanobeam_Markiewicz_20051019]1. Design extraction optics for 14 mrad crossing angle and position of the first extraction quadrupole at 6 m after IP. (DONE)2. Increase free space for the SC crab-cavity. (DONE)3. Include BPMs and correctors where needed.4. Evaluate effects of magnet field and alignment errors on beam loss and diagnostics, specify tolerances, provide correction.5. Specify realistic magnet parameters: field and length for required aperture.6. Include solenoid, anti-solenoid and DID fields, provide correction.7. Include diagnostic wigglers and weak bends.8. Consider protection collimators for magnets with large power loss.

2mrad Extraction Line

Section updated September 19, 2005. Minor updates November 9. Primary authors responsible for this section: D.Angal-Kalinin, R.Appleby, K. Kubo, N.Mokhov

The 2 mrad extraction optics [Angal-Kalinin_20050816] is based on large aperture superconducting QD0 magnet (the technology of this magnet is described in BCD for 2 mrad IR magnets) and a warm pocket coil quadrupole QF1 [Spencer_Parker_20050524]. Very large aperture superconducting sextupoles are required to provide focussing to the low energy tail particles, the studies of such large bore sextupoles have already started [Kashikhin_20050419]. L* of 4.5m has been considered in this case to accommodate the superconducting large aperture magnets. The final doublet is optimized for both the incoming as well as the outgoing beam. The collimation depth for this doublet is sufficient [Jackson_private_comm]. The large aperture special magnets; either warm Panofsky or the superconducting super septum [Spencer_Parker_20050524] starting at a distance of ~36 m from the IP contain the extracted beam and the beamstralung cone. Separate warm magnets start from ~47m. The optics contains energy cleanup chicane in the vertical plane, followed by the energy spectrometry and polarimeter chicanes [Moffeit_20050422,Nosochkov_20050817]. The beam is made parallel to the IP at the second focus for polarimetry. A long drift space after the polarimetry chicane provides enough separation (~3.5m) between the incoming and extracted beams for the beam dump. This drift also helps to increase the beam size of undisrupted beam at the window. However, additional sweeping/rastering mechanism needs to be included in this design to achieve the beam sizes as required for the beam dump and beam window [Maslov_TESLA2001_07,Walz_20050809]. Throughout the extraction line, dedicated

Rev. March 12, 2007

Beam Delivery 21

collimation sections have been provided to control beam loss. The optics design has been optimized to localize beam losses on the collimators and minimize them on magnets [Drozhdin_20mr_extr]. Estimated beam loss rates on the extraction line magnets can reach tens to hundreds of Watts per meter [Mokhov_private_comm]. A table of tolerable beam losses and radiation loads on superconducting and conventional magnets of types envisioned in the extraction line needs to be generated [Mokhov_private_comm]. Detector Integrated Dipole is not required for 2 mrad case and the orbit and angle correction will be done by moving the final doublet.    

The possible locations of the IP feedback BPMs need to be identified, taking into account also the background conditions [Hartin_20050818]. For 2mrad, the feedback BPM may need to be located in front of the FD, where the envelope of outgoing beam is still small. The outgoing beam comes to this BPM with an offset and is separated in time from incoming beam. One need to study the possibility to detect the BPM signal in this situation. Preliminary study show that using directional stripline BPMs one can achieve sufficient separation between signals [BDS_mtg_smith_20050329].    

For the IP feedback, a kicker need to be included into FD as close as possible to QD0 (to maximize the feedback capture range) [BDS_mtg_seryi_20051004]. For 2mrad case, the kicker aperture should be about 180mm. Preliminary studies show that the kicker is feasible [BDS_mtg_smith_20051101].

The optics studies indicate that it is possible to get a good bandwidth for the final focus optics for the long doublet used in this case [Seryi_20050811]. The preliminary studies also indicate that it is possible to use the 2 mrad extraction scheme for the e-e- operation [Seryi_20050812].

In 2mrad FD, energy deposition in SC QD0 due to radiative Bhabhas need to be mitigated. These particles are deposited along the QD0, primarily in the horizontal plane, in a narrow vertical stripe. The energy density exceeds the safe limit of 0.5mW/g by about a factor of three, for the nominal 500GeV CM parameters. For 1TeV case the density is ~2 higher. This issue can be solved with use of 3mm thin tungsten liner, which reduces the energy density by about a factor of ten [BDS_mtg_keller_20050726].

R&D, prototypes required: 1. Super septum and Panofsky quadrupoles2. Large aperture sextupoles [final doublet quadrupoles including pocket coil quadrupole need to be mentioned in IR magnets section of the BCD]3. Detector integration of large aperture quadrupoles and sextupoles within the detector [based on the feedback on detector opening procedure].4. Study possibility to integrate feedback BPM into FD and detect its signal in presence of large offset and of the incoming beam. 5. Integration of the large aperture feedback kicker into FD.6. Design tungsten liner for QD0, to reduces energy density due to radiative Bhabhas [this

Rev. March 12, 2007

Beam Delivery 22

should appear in IR magnet section] 

Further Studies : 1. Generate a table of tolerable beam losses and radiation loads on superconducting and conventional magnets2. Optimization of the extraction optics and collimator designs to minimize the losses on the extraction line magnets3. Possible range of 3.5m<L*<5m 4. Orbit and angle correction at the IP5. Include solenoid field in to the simulations6. Design final doublet with QD0 gradient of 250T/m based on the Nb3Ti technology7. Mitigation of radiation loads on incoming and outgoing beam SC magnets

Crab cavity system

Section updated September 20, 2005. Minor updates November 9. Primary authors responsible for this section: G.Burt, P.Goudket, H.Padamsee, L.Bellatoni

Overview: For a large crossing angle at the IP a crab cavity system is required in the baseline to increase the luminosity by horizontally rotating the bunch by half the crossing angle, without deflecting the bunch or greatly increasing the beam emittance. In the 20 mrad case the luminosity loss without a crab cavity is 80% [Burt_EuroTeV]. However, parameters can be reoptimized so that the luminosity loss without crab cavity is minimized. This is in general achieved by shortening the bunch (e.g. to 150microns) and focusing tighter in Y (e.g. with betaY=0.1mm). In this case, there is also a freedom to increase horizontal (to BetaX=40mm) size to reduce beamstrahlung energy spread to the same level as nominal. As a result, with nominal emittances, the luminosity can be about 57% for 20mrad and 68% for 14mrad. If smaller y-emittance is possible (3E-8m), luminosity could be up to 80% or higher.

In 2mrad case, the loss is 10-15% with nominal parameters [Burt_EuroTeV] and the loss can reach about 30% for large sigma Y parameter set [Seryi_20050310]. Luminosity loss in 2mrad can be avoided with use of dispersion at the IP, however it requires introducing additional correlated energy spread and causes higher energy bias (100-200ppm) of the luminosity spectrum [Seryi_20050310]. Therefore, crab cavity considered essential for 2mrad as well.    

The crab-cavity will be placed near FD in 20mrad case, and approximately in the region of QF5 quadrupole (about 400m from IP) in 2mrad case (for the reason of transverse separation with another beam, see [BDS_mtg_seryi_20041207]). To avoid deflecting the bunch the phase of the cavities must be highly stable.   

Rev. March 12, 2007

Beam Delivery 23

The ILC bunches require, at 500GeV and for a 20mrad crossing angle, a transverse kick equivalent to approximately 6.5MV for 3.9GHz cavities (19.5MV for 1.3GHz). The phase jitter between crab cavities should be less than 0.07 degrees for a 3.9GHz system or 0.02 degrees for a 1.3GHz system, with a 20mrad crossing angle [Burt_EuroTeV]. 

Cavity baseline:Fermilab 3.9GHz CKM superconductive cavity [McAshan_TM2144,Solyak_LINAC04] three/four 9-cell cavities running at 6MV/m placed near the Final Doublet. The total length of this cavity should be about 4m long and has an aperture diameter of 30mm.    Issues: This cavity may not be rigid enough to keep microphonics to the required level. Lower order mode coupler lies 11mm from centre of the beampipe. ILC application requires redesign of coupler to open the aperture.    Justification: This cavity has been under development for several years, and its research program is at an advanced stage.    R&D required: Cryostat design including cavity tuner, x-y tuner and roll tuner. Effective damping of LOM, damping of modes (SOM) within pass-band of primary deflecting mode and reliable separation of unwanted polarization of deflecting mode, beam pipe HOM coupler for high frequency modes above beam pipe cut-off frequency, search for trapped modes, beam loading. Multipacting simulations in cavity and coupler regions for the deflecting mode operation. Microphonics/rigidity. Evaluation of losses in the crab cavity region as a function of aperture. Evaluation of crab cavity length on extraction line design.    Alternatives: 13 cell CKM, CKM with larger aperture (~5m), NC (normal conducting) cavity (4.5metres) [Adolphsen_20050301], 1.3GHz cavity (~9 meters), 3.9GHz redesigned lower loss shape (~3.5 meters), stiffening rings. 

Phase control and Distribution baseline: Fast phase control based on FPGA (field-programmable gate array) with vector-sum phase control driven by two 3kW klystrons with one modulator.    Issues: The use of two klystrons can induce jitter due to the klystrons phase stability.     Justification: This is a well-developed and tested fast phase control system.    R&D required: Working system built at the operating frequency. Piezo tuners. Distribution test; the full length of line should be tested for phase jitter. Cavity phase stability tests: The phase jitter between two or more cavities should be tested both with and without a beam at full power.    Alternatives: High power ferrite phase shifters driven by a single high power klystron, 2nd klystron phase control at high power, individual cavity phase control, RF reference distribution using fiber optics [Naito_PAC2001], use of IOTs (inductive output tube) instead of klystrons if the frequency is appropriate.

Feedback system

The fast beam-beam intra-train feedback is a must and is baseline, as well as slower train-by-train feedback. The alternative is to consider additional intra-train feedback loops at the

Rev. March 12, 2007

Beam Delivery 24

entrance to BDS and throughout the linac. Baseline configuration of final doublet does not have any active means of its mechanical stabilization, but includes vibration measurement of FD as part of the baseline. This would give diagnostic for fast beam-beam feedback, may reduce commissioning time and may improve feedback convergence time. Possible implementations include accelerometer on cryostat, interferometry from cold mass to cryostat and interferometry from the cold mass to an external reference (invasive but by far the best). One of the most challenging tasks in design of fast feedback is its integration into IR, taking into account background conditions [Hartin_20050818], realistic constraints on kicker location and apertures [BDS_mtg_seryi_20051004, BDS_mtg_smith_20051101] and all engineering details of FD.    

The needed r&d includes prototype of the intra-train feedback, being done  at ATF and ATF2 [Nanobeam_Burrows_20051021], beam-test of BPMs in realistic background conditions planned to be performed at ESA, design work on incorporating kicker and BPM hardware into the final doublet, developments of methods to monitor FD stability, e.g. by means of interferometry (tests ongoing at UBC and planned at ATF). For the alternative configuration when additional feedbacks are used, feasibility studies and optimization will be necessary. 

BDS stability

The issue which define whether ILC can work with baseline feedback configuration (intratrain feedback at IP only) depends on the achieved ILC stability. The need for a stability specifications for ILC was one of the actions items identified at Snowmass, at the joint WG1/WG4 meeting. The ILC stability goals were discussed at Nanobeam 2005 workshop on October 17-21, see [Nanobeam_Seryi_20051017]. Suggested stability goals for the beam jitter, in a brief summary, are the following: up to 50% sigma at the end of linac (or in BDS diagnostics); up to 100% sigma at the end of BDS, before FD; several sigmas at IP. These goals will be consistent with baseline feedback configuration. The beam stability goals translates to required stability of the tunnel floor and give limit to additional noise due to beamline components, namely:     -- Tunnel floor stability (site + noise of nearby ILC equipment):                 Linac area: up to ground motion model K or C                 BDS area: up to ground motion B*3 or gm C/3 and     -- Additional noise due to beamline components:                Linac area: up to 30 nm                BDS area : up to 10 nm                FD : up to ~100nmExceeding these stability limits will likely require using multiple intratrain feedbacks along the machine.  These stability requirements need to to be discussed and set consistently by the DR, Linac and BDS groups.     Stability and background vibration were studied at many sites, including most of the sample sites under considerations. Data can be found in [CERN_SL_94_41] for LEP tunnel data, [PRSTAB_e031001]  for data near Fermilab. Data for DESY site as well as recent comparison of different sites can be found in [Amirikas_Nanobeam05].

Rev. March 12, 2007

Beam Delivery 25

R&D required :    -- studies of stability of sites (preliminary data exist for most of the sample sites)    -- investigation and minimization of vibration produced by conventional facility, power supplies, and other in- and near-tunnel equipment    -- studies and minimization of noises generated at or amplified by the beamline elements in linac and BDS, in particular linac quads in the cryostats and BDS magnets    -- studies and optimization of stability properties of final doublet. 

Beam dump system

Section updated December 2, 2005. Primary authors responsible for this section: R.Appleby, D.Walz, R.Sugahara/S.Ban

The baseline beam dump is based on water vortex design rated for 18MW beam [Walz_20050817, Appleby_20050914]. The choice of a water dump for the baseline has many advantages: the water dump has been studied in detail for accelerator projects and a lower power (2MW) beam dump was used at the SLC [Walz_IEEE1965]. Furthermore the problems of the larger dump design have been noted, and the studies indicate there are no “show-stoppers”. The water dump for the TESLA project was studied in detail at DESY [Bialowons_TESLA200104]. The alternatives suggested in the ACD require more R+D to be sufficiently mature for a baseline design. The beam dump is common for charged beam and photons for the 20mrad interaction region layout, while for 2mrad layout the photon dump is separate (rated to about 1MW).

There are separate beam dumps rated for full power for all beam lines including tune-up lines, for a total of six beam dumps in the baseline. The tune-up dumps are sufficiently remote from the IP, so that the collider halls can be accessed for detector maintenance while the linac is being tuned, and full beam sent to tune-up dump.     

Discussion of pro/cons of separate full power tune-up dumps have started [BDS_mtg_20051129]. Technically, the elimination of two full power tune-up dumps should be possible, there will be impact on availability which may be partly mitigated by reduced power tune-up dumps (~0.5MW), the cost saving need to be further evaluated, detailed design would need to be made.   

The water flow is sufficient to avoid volume boiling of the water (for disrupted beam). The vacuum/water boundary consists of a suitably thin window. The cyclic stress of the window is controlled by limiting the temperature rise per pulse of the water system. The undisrupted beam size is allowed to increase in the extraction line, to avoid damaging the window, which required lengthening the extraction line by a hundred meters or so. Preventing water boiling due to undisrupted beam has to be done by sweeping the beam in the final part of the extraction line, as it require large increase of the beam size, which cannot be realistically done with lengthening the drift length. The water circuit consists of two closed loops and an external water circuit. The inner water loop is pressurized to 10bar and has a volume of around 18 cubic meters. The length of the dump, including all shielding, is about 25m

Rev. March 12, 2007

Beam Delivery 26

longitudinally and about 15m transversely.

The required R+D items for the baseline are a study of window survivability, and the corresponding computation of the displacement per atom (DPA). A window replacement procedure and schedule can then be developed. A prototype of the window and a beam test are also necessary. The required test beam must give similar energy densities in the window as the full ILC machine. Furthermore, some studies of pressure wave formation maybe necessary.

The alternative beam dump design is the gas dump [Leuschner_20030916]. This consists of about 1km of a noble gas (Ar looks the most promising) enclosed in a water cooled iron jacket. The gas core acts as a scattering target, blowing the beam up and distributing the energy into the surrounding iron. This gas dump design may ease some issues such as radiolysis and tritium production, and a gas profile can be exploited to produce a uniform energy deposition along the length of the dump. However, other issues arise such as particle beam heating of the gas and ionization effects [Agapov_20050914]. Further studies needed to understand feasibility and benefits of the gas dump. A further possibility is a gas/water hybrid dump, involving the use of a shorter gas dump as a passive beam expander, followed by a small water dump. This option also required further study. A further possibility is for a rotating solid dump immersed in water, or a dump based on some kind of liquid metal.

The required R+D items for the alternative design are studies of gas heating, including ionization effects, and a study of radiation and activation effects. A study of the gas dump windows is also required. A smaller scale prototype of the dump, and some test beam, would also be required. 

Beam Energy Measurements

Section updated September 19, 2005. Primary authors responsible for this section: E.Torrence, S.Boogert

Overview

Precision absolute beam energy measurements are required by the ILC physics program to set the absolute energy scale for measurements of particle masses. In both direct reconstruction and threshold scans, the center-of-mass collision energy needs to be measured to an absolute accuracy below 200 ppm to keep collision energy uncertainties from dominating experimental uncertainties of 50 MeV for both the Higgs and top quark masses. To reach this goal, a target uncertainty of 100 ppm on the absolute beam energy has been set.

In addition to the absolute beam energy measurement, relative measurements of the beam energy pulse-to-pulse along the train at the 100 ppm level are seen to be critically important to keep the variation in beam energy at an acceptable level and mitigate the impact of correlations between beam energy variations and other beam parameters such as luminosity or polarization. Relative measurements of the disrupted energy spectrum made downstream

Rev. March 12, 2007

Beam Delivery 27

of the interaction point are also seen as useful to provide direct information about the collision process and provide data to validate models of beam-beam effects in the collision process.

To achieve the challenging goal of a 100 ppm absolute beam energy measurement, two independent and complimentary detectors are planned for each beam. Upstream from the IP, a spectrometer with a four-magnet chicane and precision RF-BPMs will be used. This device, an evolution of the LEP-II energy spectrometer [cite], is designed to be capable of making high-precision bunch-to-bunch relative measurements in addition to measuring the absolute beam energy scale. Downstream from the IP, a spectrometer using synchrotron radiation in the style of the SLC WISRD [cite] is planned. This device can also monitor the energy spectrum of the disrupted beams during collisions.

Upstream Measurements

This section may be edited further.

The upstream energy spectrometer is an evolution of the LEP-II spectrometer design. A four-magnet chicane in the instrumentation region (figure link) provides a point of dispersion which can be measured using triplets of high-precision RF BPMs. To avoid emmittance dilution, the maximum bend angle for this chicane is expected to be less than 100 micro-Radians. With a characteristic spectrometer length of about 10 meters, the maximum displacement of the beam will be a few millimeters. To achieve the required spectrometer performance, then, this displacement must be measured to an accuracy and precision below 100 nanometers.

Rather than require the BPMs to achieve this accuracy over a large dynamic range, the design includes precision actuators and movers to keep the beam nearly centered in the BPMs at all times. These actuators, then provide the bulk of the position information, which only small corrections being provided by the BPMs themselves. One key aspect of this design is the ability to calibrate the straight-line reference line on a time scale which is comparable to the mechanical and electrical stability of the system. This straight-line reference can be derived by ramping the chicane magnets and either making measurements with zero field, or reversing the field polarity an making position measurements at positive and negative deflection angles. The details of this calibration procedure and requirements on the magnet system still need to be specified in detail.

The nanoBPM project has demonstrated RF BPM precision and stability below 50 nm for several hours [cite], although with somewhat reduced aperture as would be foreseen in the upstream spectrometer. Using this experience, designs for the spectrometer BPMs are currently being developed. It is important to investigate the sensitivity of these devices to the details of beam tilt, beam size, bunch length, backgrounds, etc. Beam tests of these devices are foreseen to begin at SLAC ESA in the end of 2005.

Downstream Measurements

Rev. March 12, 2007

Beam Delivery 28

The downstream energy spectrometer is an evolution of the SLC WISRD design. A three-magnet chicane in the extraction line provides the necessary beam deflection, while the trajectory of the beam in the chicane is measured using synchrotron radiation produced in wiggler magnets imaged ~70 meters downstream at a secondary focus near the polarimeter chicane.

The layout of the instrumentation chicanes and detectors for both the energy spectrometer and polarimeter is indicated for 2 mRad and 20 mRad extraction lines. With a 3 mRad bend angle in the energy chicane and about 70 meters between the chicane and the synchrotron radiation detector plane, the effective single-beam dispersion at the detector plane is nearly 1 GeV/mm.

To achieve a 100 ppm accuracy, a position-sensitive detector for synchrotron radiation with a pitch of order 100 microns is necessary. While the WISRD used a fine wire array to achieve similar performance, the Oregon group (Torrence) is pursuing R&D into using the Cherenkov radiation produced by secondary electrons in quartz fibers. The advantages of this approach over wires are simplicity in readout, speed, radiation hardness, and potentially reduced cross-talk between channels. Any radiation hard position sensitive detector (like diamond strips) could potentially be used.

Future R&D Required

The immediate R&D required for the spectrometers is listed here:

Specification of upstream spectrometer calibration and operational procedure. Design of spectrometer-specific RF BPM. Underway by several groups. Stability and performance tests for BPMs and associated readout electronics. Related

activity underway at ATF as part of the nBPM program. BPM tests scheduled to start at SLAC ESA at the end of 2005.

Detailed simulation of downstream spectrometer performance including realistic magnetic fields and backgrounds.

Design and evaluation of large-wavelength, large-aperture wiggler magnets. Test of downstream spectrometer quartz fiber readout scheme, scheduled for SLAC

ESA starting in Fall 2005.

For the longer term, a full-scale prototype of both a BPM-based and synchrotron radiation-based spectrometer are foreseen for SLAC ESA.

References need to be fixed.

Luminosity Measurements

Section updated September 30, 2005. Primary authors responsible for this section: W.Lohmann, M.Woods

Rev. March 12, 2007

Beam Delivery 29

The luminosity of a linear collider will be measured with a precision of 10-3 or better by measuring the Bhabha rate in the polar region from 30-90 mrad in the LUMICAL detector [Stahl_LCDET200504].  At 500 GeV center-of-mass energy, the expected rate in this region is ~10 Bhabhas per bunch train.  At smaller polar angles of 5-30 mrad the rate or energy deposition of beamstrahlung e+e- pairs can be measured in the BEAMCAL detector for a fast luminosity diagnostic.  The expected rate in BEAMCAL is 15,000 pairs (and 50 TeV energy deposition) per bunch crossing. 

BEAMCAL is expected to be useful for machine tuning and can be used for the fast IP feedback [Burrows_LCWS05] planned to stabilize the colliding beams.  In addition to its total energy or rate signals for a luminosity diagnostic, the spatial distributions of pairs in BEAMCAL may be useful for determining some of the beam collision parameters such as spotsizes and bunch lengths [White_Stahl_Yamamoto].

LumiCal and BeamCal are positioned inside the Detector just in front of the first quadrupole magnets.[Moenig_200408]  LUMICAL is planned to be a segmented silicon-tungsten calorimeter.  BEAMCAL must be very radiation hard and a finely segmented diamond-tungsten calorimeter is planned, though other technologies are also studied.  BEAMCAL should also provide good hermeticity and efficiency for detecting high energy electrons; these are needed to suppress brackgrounds from copious 2-photon events in new particle searches (ex. SUSY).  A possible layout of the very forward region of an ILC detector [Abramowicz_NSCI2004], in this case designed for head-on collisions or a small crossing angle, is shown in Figure 1.

R&D Required

LUMICAL:  compare precision achievable for different crossing angle geometries and study effects of non-uniformities in the azymuthal distributions of the Bhabha signal and backgrounds from pairs

LUMICAL:  adequate event rate for 10-3 precision can be achieved by going to larger polar angles where there is less effect from a finite crossing angle, but the theoretical uncertainties in the Bhabha rates can be larger at high polar angles.  10-4 relative precision is desired for the Giga-Z option; pair backgrounds for for Giga-Z will be significantly reduced but still need to understand azymuthal effects for the Bhabha signal rate systematics.  These issues need further study.

BEAMCAL:  study possible systematics for its use determining beam parameters and how this is affected by a crossing angle

BEAMCAL:  check that optimizing the pair signal also optimizes the e+e- luminosity for different aberration tuning scans

BEAMCAL:  detector requirements and capabilities for electron identification given the high flux and energy deposition of beamsstrahlung pairs; also effect from crossing angle on hermeticity and electron id.

BEAMCAL:   detailed detector design and prototype:  radiation-hard, hermetic, electron id

Rev. March 12, 2007

Beam Delivery 30

Other Detectors:  radiative Bhabha [Napoly_CERN2000135] and wide-angle beamsstrahlung photon detectors [Bonvicini_PHREV1999]

Beam Polarization Measurements

Section updated September 30, 2005. Primary authors responsible for this section: K.C. Moffeit, K. Moenig, K.P. Schuler, M. Woods

Overview

            Precise polarimetry with 0.25% accuracy is needed.[Moortgat-Pick_hep-ph0507011]  Compton polarimeters are being designed to achieve this and have been included in the baseline beam delivery design.[Moffeit_SLACPUB11322,Meyners_LCWS05]  Preliminary designs for polarimeter diagnostic chicanes are included upstream and downstream of the IP for both the 2mrad and 20mrad IR designs. Detailed studies are underway to refine these designs and evaluate their performance capabilities.  To achieve the best accuracy for polarimetry and to aid in the alignment of the spin vector, it is desirable to implement polarimeters both upstream and downstream of the IR.  The beam optics need to be designed such that the polarization vector can be fully longitudinal simultaneously at the collider IP and the two polarimeter IPs.

            The upstream polarimeter measures the undisturbed beam during collisions. The relatively clean environment allows a laser system that measures every single bunch in the train and a large lever arm in analyzing power for a multi-channel polarimeter, which facilitates internal systematic checks.

            The downstream polarimeter measures a priory the polarization of the outgoing beam after collision. The average depolarization for colliding beams is 0.3%, and for the outgoing beam 1%.  Due to a clever choice of the extraction line optics the beam can, however, be focused such that its polarization is very similar to the luminosity-weighted polarization. The polarization of the undisturbed beam can be measured as well with non-colliding beams. The much higher background requires a high power laser that can only probe one or a few bunches per train and the lever arm in analyzing power is smaller.

Upstream Measurements

            The upstream polarimeters are located ~1400 meters before the e+e- IP. The design has evolved from an earlier study for the TESLA machine.[Gharibyan_LCDET200147]  Most major aspects of this work, except for the spectrometer configuration, remain valid for the ILC. In particular it is foreseen to use a similar laser as will be used for the electron source.  A prototype for such a laser was developed by Max-Born Institute,[Schreiber_NIM2000] for the TTF injector, and is well adapted to the ILC pulse structure.

             Dedicated 4-magnet chicane spectrometers will be employed, similar to those at the extraction line polarimeters. This will eliminate some of the operational shortcomings

Rev. March 12, 2007

Beam Delivery 31

inherent in the original TESLA design that relied on beamline magnets in the existing BDS lattice. A schematic layout of the chicane spectrometer is shown in Figure 1.The horizontal width of the good field region of the individual dipoles is chosen to accommodate a maximum dispersion of 11 cm for the lowest expected beam energy of 45.6 GeV for the Giga-Z option. The laser beam enters and exits between the inner two dipoles, which must be separated by some 8 meters for a vertical beam crossing of 10 mrad. A possible optical arrangement was given at LCWS-05.[Meyners_LCWS05]

            Compton electrons generated at the laser IP at mid-chicane will propagate essentially along the electron beam direction . The third dipole D3 will fan out the Compton electron spectrum, while the fourth dipole can be used to restore the angular direction, if it has sufficient width. The Compton electrons are detected behind the last dipole in a gas Cerenkov hodoscope with 20 identical channels.[Meyners_LCWS05]

Downstream Measurements

The layout for the 20 mrad crossing angle interaction region shown in Figure 2 has the Compton interaction point approximately 142 meters downstream from the e+e- Interaction Point. All bends are in the vertical plane. The extraction line apertures are designed to accommodate the ±0.75mrad cone of beamsstrahlung photons produced in the e+e- interaction and the low energy disrupted electrons. The 2 mrad crossing angle extraction line first moves the extracted beam away from the incoming beam line and then bends the beam back to the direction it had at the e+e- interaction point . This is done in the horizontal plane as shown in Figure 3. The Compton polarimeter is located 226 meters downstream from the 2 mrad crossing angle e+e- interaction point and the polarimeter chicane bends in the vertical plane.

The Compton interaction point is located at a secondary focus in the middle of a chicane with 20 mm dispersion, but with no net bend angle with respect to the primary IP.[Moffeit_SLACPUB10669] At the middle of the chicane the Compton scattering occurs and the scattered electron is confined to a cone having a half-angle of 2µrad and is effectively collinear with the initial electron direction. The beam-beam depolarization effects are measured in the extraction line polarimeter directly by comparing beams in and out of collision. Also, spin precession effects due to the final focus optics and beam-beam deflections can be studied by correlating the polarization and Interaction Point beam position monitor measurements.

             A 532-nm (2.33eV) circularly polarized laser beam collides with the electron beam in the middle of the Polarimeter Chicane .  Compton-scattered electrons near the kinematic edge at 25.1 GeV are detected in segmented detectors near the last chicane magnet.

 R&D Required for Baseline Design Study of extraction line polarimeters for 14/14 baseline and comparison of extraction

line polarimeters for 2mrad and 20mrad IRs alternative     - backgrounds (synchrotron radiation, disrupted electron beam, beamsstrahlung,

Rev. March 12, 2007

Beam Delivery 32

radiative Bhabhas)     - sensitivity to misalignments of spin orientation, collision offsets     - compatibility with energy spectrometer     - requirements and feasibility for chicane magnets

spin alignment procedures and uncertainties; sensitivity to crossing angle and DID develop detailed Compton laser, IP and detector designs

R&D Required for Alternative Configuration

evaluate use of a Fabry-Perot cavity for the laser at the Compton IP at the upstream polarimeter [Variola_LCWS05]; this technology may have better applications for the laser systems for a Compton-based polarized positron source or a gamma-gamma collider

Multi-TeV issues

The ILC BDS is being designed to be optimal up to 500GeV CM (1TeV CM in upgrade).     Realizing that the question of multi-TeV upgrade goes much beyond the scope of the working group, the group suggests that serious consideration need to be given by the whole community to studying the advantages and disadvantages of not precluding the multi-TeV compatibility.

Technical implications and cost impact due to multi-TeV constraints will have to be studied. Such constraints were discussed in the recently (November 10, 2005) published report [CERN_Open_2005_024] and includes:    -- crossing angle about 20mrad required (it is believed that parameters of multi-TeV collider can be optimized so that it works also with 14mrad crossing angle [D.Schulte, private communication])    -- horizontal bend between high energy end on the linac and beam delivery should be less than 2mrad, zero for vertical    -- strongly prefer laser straight linac tunnels    -- provision to add tunnel alcoves every 600m to house a drive beam return loop and 2MW drive beam dump    -- strongly prefer ground motion to be no worse than model A or B    -- surface space 1200x250m in IP region to house the drive beam generation complex    -- provision to connect to power grid with capacity 450MW    -- main beam dumps for 20MW, very similar to ILC

Several of these constraints are related to BDS. In particular, the linacs need to be pointing to the larger crossing angle IR. In the former baseline with two IRs with 20/2mrad, or present baseline with 14/14mrad, satisfying this constraint does not affect the performance and does not necessarily affects the cost. This layout choice, however, may affect site selection, for example because the orientation of linacs at angle widens the site, and this could make it

Rev. March 12, 2007

Beam Delivery 33

more difficult to place the collider under a straight power-line (may be partly mitigated by horizontal back kinks at the middle of the linac). The linacs-at-an-angle layout may complicate, but would not make impossible moving the IP many km downstream, if such possibility would ever be needed.  The laser straight tunnel may have cost impact in comparison with tunnel which follows earth curvature closer to  surface. These and other configuration issues, technical and cost implications need to be studied seriously and the impact will be clearly strongly site dependent.

Number of sections were not yet described. These sections and authors name are listed in [WG4_20050825_BCD]:

IR & IR magnets for 20 mrad and alternative 10-12mrad (partly written)IR & IR magnets for 2 mrad and alternative 0mradBeam switchyard (partly written)Diagnostics section (partly written)Detector integration (DID, magnets, assembly, support, push-pull issues)Detector performance (Backgrounds)Final focus optics (performance, chromaticity compensation, antisolenoids, DID)Magnets nomenclature, stability, etc.Standard components (BPMs, current monitors, loss monitors, vacuum, ) Tolerances & tuning Gamma-gamma & e-e-Summary of technical risk, R&D needed for baseline & alternatives (partly written) 

Detector Assembly

The assembly of ILC detectors will be done on-surface, in a manner similar as was done for CMS detector at CERN. The on-surface assembly allows saving 2-2.5 years of time, because detector assembly on surface can start earlier than underground. Both pure “CMS assembly” and “modified CMS assembly” are discussed. The details for the schedule, hall sizes, capacity of cranes, etc. are being worked out by BDS, CF&S and Detector Concept groups.

References

1. [Pre_Snowmass_BCD] Pre-Snowmass draft BDS BCD. http://www-project.slac.stanford.edu/ilc/acceldev/beamdelivery/bds_bcd.htm

2. [WG4_Snowmass_Agenda] WG4 Agenda at Snowmass 2005. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/agenda

3. [WG4_Snowmass_Summary] WG4 Summary of first week at Snowmass, August 19, 2005. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug19_wg4_1st_week_summary.ppt

Rev. March 12, 2007

Beam Delivery 34

4. [Snowmass_decision_list] Snowmass decision spreadsheet, assembled by T.Himel, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/GG3/tom_himel20050824165242.xls

5. [WG4_20050823] WG4 plenary discussion of responses to GDE questions, August 23, 2005. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug23_wg4_gde_questions.ppt

6. [WG4_20050825] WG4 responses to GDE questions, August 25, 2005. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug25_wg4_gde_questions.ppt

7. [WG4_20050825_BCD] TOC of BDS BCD/ACD and list of authors, as discussed at Snowmass. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/BDS_BCD_ACD.docsee also the modified listhttp://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/modified_BDS_BCD_ACD_sept8.doc

8. [Raubenheimer_20050228] T.Raubenheimer, ILC Suggested Beam Parameters Range,  http://www-project.slac.stanford.edu/ilc/acceldev/beampar/Suggested%20ILC%20Beam%20Parameter%20Space.pdf, Feb.28, 2005.

9. [Seryi_20050817] Alternative 1TeV High Luminosity parameters, A.Seryi, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/GG1/aug17_seryi_high_lumi_pars.ppt

10. [Kozanecki_20050824]  Normalized backgrounds in ILC tracking detectors, W.Kozanecki, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug24_WK_BgdTol_Snow05.ppt

11. [ILCSC_scope] ILCSC scope document, September 2003, http://www.fnal.gov/directorate/icfa/LC_parameters.pdf . Updated, Novermber 2006, http://www.linearcollider.org/newsline/pdfs/20061207_LC_Parameters_Novfinal.pdf .

12. [BDS_optics] Optics files for BDS baseline: Incoming beamlines: http://www.slac.stanford.edu/~mdw/ILC/Extraction beamlines: http://www.slac.stanford.edu/~yuri/ILC/Decks/

13. [BDS_IR_models] Geant models of IR regions for BDS baseline: link? 14. [Parker_20050816] Compact SC quads, B.Parker, Snowmass 2005, 

http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug16_wg4_parker_CompactQD0.pdf

15. [ILC_TRC2] Report of 2nd ILC Technical Review Committee, SLAC-R-606, Feb 2003.

16. [Iwashita_20050818] Update on rf kicker 0-degree scheme, Y.Iwashita, Snowmass 2005,   http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_RFKickerUpdate1.pdf

17. [Keller_20050818] Zero-degree extraction with an electrostatic separator, L.Keller, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_ElectrostaticSeparator.ppt

Rev. March 12, 2007

Beam Delivery 35

18. [Bellantoni_20050818] FNAL 3.9GHz deflecting cavity, L.Bellantoni, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_CRAB_intro_bellantoni.pdf

19. [Burt_20050818] Crab cavity system design, G.Burt, P.Goudket, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_burt_Crabsnow.ppt

20. [Walz_20050817] Beam dump concept for ILC, Dieter Walz, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug17_beam_dump_d_walz.pdf

21. [LCC52] http://www-project.slac.stanford.edu/lc/ilc/TechNotes/LCCNotes/PDF/lcc0052.pdf

22. [LCC111] http://www-project.slac.stanford.edu/lc/ilc/TechNotes/LCCNotes/PDF/lcc-0111.pdf

23. [Mokhov_Conf_05_154] http://www-ap.fnal.gov/users/mokhov/papers/2005/Conf-05-154.pdf

24. [Keller_20041015] L.Keller, ILC-Americas workshop, October 15, 2004,  http://www-project.slac.stanford.edu/ilc/meetings/workshops/US-ILCWorkshop/talks/ILC_MPS.ppt

25. [Jackson_20050816] http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug16_collimation_summary_jackson_snow05.ppt

26. [Carter_20050816] http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug16_carter_Snowmass05_CollDepth.pdf

27. [Drozhdin_BDIR2005] A.Drozhdin et al., STRUCT Modeling of Collimation and Extraction System Performance,  June 21, 2005, BDIR Workshop, London,  https://ilcsupport.desy.de/cdsagenda/askArchive.php?base=agenda&categ=a0522&id=a0522s14t7/transparencies

28. [Woods_20050302] M.Woods, Update on ESA beam tests, March 2, 2005,  http://www.slac.stanford.edu/xorg/lcd/ipbi/monthlymeetings/02mar2005/woods/ESA_beamtests.pdf

29. [LCC118] LCC-118 http://www-project.slac.stanford.edu/lc/ilc/TechNotes/LCCNotes/lcc_notes_index.htm

30. [TRC2_p340] ILC TRC, Second report, Section 7.3.2.6, p340, http://lcdev.kek.jp/TRCII/PAPERS/TRC03C7.PDF

31. [Collwake_beamtest_request] Collimation wakefield beam test request, April 25, 2005, http://www-project.slac.stanford.edu/ilc/testfac/ESA/files/ColWake_TestBeamRequest.pdf

32. [Angal-Kalinin_20050816] D.Angal-Kalinin, Status of Optics, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug16_snowmass_optics_summary.ppt

33. [Spencer_Parker_20050524] C. Spencer, B.Parker, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-05-24/index.htm

Rev. March 12, 2007

Beam Delivery 36

34. [Kashikhin_20050419] V. Kashikhin, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-04-19/index.htm

35. [Jackson_private_comm] F.Jackson, private communication 36. [Moffeit_20050422] K. Moffeit,

http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-04-22/index.htm

37. [Nosochkov_20050817] Y. Nosochkov, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug17_nosochkov_extraction2mrad.pdf

38. [Maslov_TESLA2001_07] M. Maslov, et al, TESLA report 2001-07 39. [Walz_20050809] D. Walz,

http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-08-09/index.htm

40. [Drozhdin_20mr_extr] A. Drozhdin, http://www-ap.fnal.gov/users/drozhdin/prdriver/pap_20extr_yuri.pdf or pap_20extr_yuri.pdf

41. [Mokhov_private_comm] N. Mokhov, private communication 42. [Seryi_20050811] A. Seryi, FF Optics for 2mrad with L*=4.5m, http://www-

project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-09-20/index.htm 43. [Seryi_20050812] A. Seryi, e-e- in 2mrad optics, http://www-

project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-09-20/index.htm 44. [Parker_20050915] P.Parker, Draft 20 mr IR Magnet Description, 15-Sep-05, (original

Word file)http://www-project.slac.stanford.edu/ilc/acceldev/beamdelivery/bcd_docs/ILC_MagnetDescription_15Sept05.htm

45. [Parker_20050818] B.Parker, Thoughts on minimum crossing angle with compact SC quads, Snowmass 2005, August 18,  http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_wg4_parker_XingAngle.pdf

46. [Parker_20050823] B.Parker, Update on SC quad with self-compensating fringe field, Snowmass 2005, August 23, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug23_parker_comp.pdf

47. [Parker_20050913] B.Parker, Compact SC quad for intermediate crossing angle, BDS meeting September 13, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-09-13/index.htm

48. [Burt_EuroTeV] G.Burt, P.Goudket, et al., Forthcoming EuroTeV note on crab-cavity system (to be written).

49. [McAshan_TM2144] M. McAshan, R. Wanzenberg, RF Design of a Transverse Mode Cavity for Kaon Separation, FERMILAB-TM-2144, May 2001

50. [Solyak_LINAC04] N. Solyak et al., Test Results of the 3.9GHz Cavity at Fermilab, Proc. LINAC04, THP85

Rev. March 12, 2007

Beam Delivery 37

51. [Adolphsen_20050301] C. Adolphsen, Options for ILC Crab Cavity, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-03-01/index.htm

52. [Naito_PAC2001] T. Naito et al., RF Reference Distribution Using Fibre Optic Links at KEKB Accelerator, Proc. PAC2001, Chicago

53. [Appleby_20050914] Overview of ILC dump issues, Rob Appleby, LC-ABD dump meetings, 14th September 2005, http://www.astec.ac.uk/lc-abd/ral.14.09.05/ral.14.09.05.htm

54. [Walz_IEEE1965] Walz et al, in IEEE Trans.Nucl.Sci.12:367-371, 1965 (issue No.3) 55. [Bialowons_TESLA200104] Bialowons et al, TESLA note 2001-04 56. [Leuschner_20030916] Another idea of a LC dump, Albrecht Leuschner, 16th

September 2003, http://tesla.desy.de/new_pages/hamburg_meeting_9_2003/pdf/wg2/gasdump_talk_16sep03.pdf

57. [Agapov_20050914] Some thoughts about gas dumps, Ilya Agapov, LC-ABD dump meeting, 14th September 2005, http://www.astec.ac.uk/lc-abd/ral.14.09.05/Gas-Dump-14-09-05.pdf

58. [Stahl_LCDET200504]  Luminosity Measurement via Bhabha Scattering Precision Requirements for the Luminosity Calorimeter, A. Stahl, LC-DET-2005-004, 2005

59. [Burrows_LCWS05]  Interaction Point Feedback System R&D, Philip Burrows, presentation at LCWS05.

60. [White_Stahl_Yamamoto]  Reconstruction of IP Beam Parameters at the ILC from Beamstrahlung, G. White, SLAC-PUB-11341, 2005;       Diagnostics of Colliding Bunches from Pair Production and Beam Strahlung at the IP, A. Stahl, LC-DET-2005-003, 2005.  Beam Profile Monitor, H. Yamamoto, presentation at SLAC MDI Workshop, 2005. 

61. [Moenig_200408]  K. Moenig, talk at the Workshop "Instrumentation of the Forward Region of a Linear Collider Detector", August 2004, DESY-Zeuthen, Germany.

62. [Abramowicz_NSCI2004]  H. Abramowicz et al., IEEE Trans. Nucl. Sci. 51, 2983 (2004).

63. [Napoly_CERN2000135]  Luminosity Monitor Options for TESLA, O. Napoly and D. Schulte, CERN-OPEN-2000-135, 2000. 

64. [Bonvicini_PHREV1999]  Measurement of Colliding Beam Parameters with Wide Angle Beamstrahlung, G. Bonvicini, D. Cinabro, and E. Luckwald, Phys.Rev.E59:4584-4593,1999. e-Print Archive: physics/9812020.   Diagnostics with Beamstrahlung Electrons and Photons, C. Grah, presentation at SLAC MDI Workshop, 2005.

65. [Moortgat-Pick_hep-ph0507011]  The role of polarized positrons and electrons in revealing fundamental interactions at the Linear Collider,      G. Moortgat-Pick et al.  e-Print Archive: hep-ph/0507011 (submitted to Physics Reports).

66. [Moffeit_SLACPUB11322]  Polarization Setup and Polarimetry for 2 IRs, and Status of Downstream Polarimeter Designs,  K.C Moffeit, M. Woods, Y. Nosochkov , K.P. Schuler, K. Moenig, and W. Oliver,. SLAC-PUB-11322, 2005.

67. [Meyners_LCWS05]  Upstream Polarimetry with 4-magnet Chicane, N. Meyners, V. Gharibyan, K.P. Schuler, presentation at LCWS05.

Rev. March 12, 2007

Beam Delivery 38

68. [Gharibyan_LCDET200147]  The TESLA Compton Polarimeter, V. Gharibyan, N. Meyners, K.P. Schüler, LC-DET-2001-047, http://www.desy.de/~lcnotes

69. [Schreiber_NIM2000]  Running Experience with the Laser System for the rf Gun Based Injector at the TESLA Test Facility Linac, S. Schreiber et al., Nucl. Instr. Meth. A 445 (2000) 427.

70. [Moffeit_SLACPUB10669]  Studies for a Downstream Compton Polarimeter at the ILC, K.C. Moffeit and M. Woods, SLAC-PUB-10669, 2004.

71. [Variola_LCWS05]  A Fabry-Perot Cavity for FLC Polarimetry, A. Variola, presentation at LCWS05.

72. [Nosochkov_20050817_20mrad] Y. Nosochkov, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug17_nosochkov_extraction20mrad.pdf

73. [Nosochkov_20050510] Y. Nosochkov, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-05-10/index.htm

74. [Ferrari_20050831] A. Ferrari, http://www-project.slac.stanford.edu/ilc/acceldev/beamdelivery/bcd_docs/Ferrari-050831-update.pdf

75. [Nanobeam_Seryi_20051017] A.Seryi, Issues of stability and ground motion in ILC, Nanobeam workshop, October 17, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-P-03.ppt and http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-10-25/index.htm

76. [Nanobeam_Markiewicz_20051019] T.Markiewicz, T.Maruyama, Y.Nosochkov, Status of 14mrad extraction line, Nanobeam workshop, October 19, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-03.ppt

77. [Nanobeam_Seryi_IR_20051019] B.Parker, A.Seryi, IR optics optimization, DID and anti-DID, Nanobeam workshop, October 19, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-04.ppt

78. [Nanobeam_Seryi_Civil_20051019] A.Seryi, BDS civil layouts and upgrade path from single IR to two IRs, Nanobeam workshop, October 19, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-05.ppt

79. [Nanobeam_Markiewicz_IR_20051019] T.Markiewicz, IR Design Issues Requiring More Discussion, Nanobeam workshop, October 19, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-08.ppt

80. [Nanobeam_Parker_20051018] B.Parker, Recent Progress Designing Compact Superconducting Final Focus Magnets for the ILC, Nanobeam workshop, October 18, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2d-01.pdf

81. [Nanobeam_Jain_20051018] A.Jain, et al., Vibration Measurements in a RHIC Quadrupole at Cryogenic Temperatures, Nanobeam workshop, October 18, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2d-05.ppt (This is PDF file, not PPT. Save it, rename to pdf and open in Acrobat).

82. [Nanobeam_Parker_RnD_20051018] B.Parker, A.Jain, R&D Activities Regarding ILC Compact Superconducting Final Focus Magnets, Nanobeam workshop, October 18, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2d-07.pdf

Rev. March 12, 2007

Beam Delivery 39

83. [Nanobeam_Angal_Kalinin_20051019] D.Angal-Kalinin, Update on 2mrad design, Nanobeam workshop, October 19, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-11.ppt

84. [Nanobeam_Mattison_20051020] T.Mattison, ILC fast abort systems, Nanobeam workshop, October 20, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-12.pdf

85. [Nanobeam_Seryi_20051021] A.Seryi, Summary of working group on BDS and IR, Nanobeam workshop, October 21, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-S-02.ppt

86. [Nanobeam_Burrows_20051021] P.Burrows, Summary of working group on stabilization and feedback, Nanobeam workshop, October 21, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-S-03.ppt

87. [Nanobeam_Parker_20051021] B.Parker, Summary of working group on magnets, Nanobeam workshop, October 21, 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-S-05.pdf 

88. [Blair_Snowmass_20050817] G.Blair, ILC Laser wires, Snowmass 2005, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/GG2/grahame_blair20050818043011.ppt

89. [Woodley_20050927] M.Woodley, BSY emittance, diagnostic section, BDS meeting September 27, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-09-27/index.htm

90. [Nanobeam_Blair_20051018] G.Blair, Simulation of ILC Laser Wire, Nanobeam, October 18 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG1-02.ppt

91. [Nanobeam_AngalKalinin_20051018] D.Angal-Kalinin, Laser-wire location in ILC Diagnostics section, Nanobeam, October 18 2005, http://atfweb.kek.jp/nanobeam/files/presen//presen-WG1-01.ppt

92. [Seryi_20050310] A.Seryi, Test of using IR dispersion to recover luminosity loss in 2mrad, March 10, 2005, bcd_docs/ip_disp_2mrad_20050310_seryi.pdf

93. [BDS_mtg_20051101] Discussion of e+ source location at BDS meeting on November 1, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-11-01/index.htm

94. [Jackson_20051102] F. Jackson, Collimation studies,  Meeting at Daresbury devoted to Collimation for ILC, November 2-3, 2005,  http://www.astec.ac.uk/ap/collider/collimmeet02Nov05/fjackson_collim_mtg_02Nov05_BCD.ppt

95. [Seryi_20051102] A. Seryi, Collimation optics issues,  Meeting at Daresbury devoted to Collimation for ILC, November 2-3, 2005, http://www.astec.ac.uk/ap/collider/collimmeet02Nov05/seryi_ilc_collimation_2Nov05.ppt

96. [BDS_mtg_seryi_20041207] A.Seryi, Crab cavity location for 2-mrad x-ing, BDS meeting, December 7, 2004, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2004-12-07-Nosochkov,Pivi,Markiewicz,Seryi/index.htm

Rev. March 12, 2007

Beam Delivery 40

97. [BDS_mtg_smith_20050329] S.Smith, Independent Measurement of Two Beams in an IP Feedback BPM in 2mrad, BDS meeting, March 29, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-03-29/index.htm

98. [BDS_mtg_smith_20051101] S.Smith, Kicker for fast feedback, BDS meeting, November 1, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-11-01/index.htm

99. [Hartin_20050818] T.Hartin, IP FB BPM background environment and BDS design, Snowmass 2005, August 18, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug18_hartin_fbk_bpm.ppt

100. [BDS_mtg_seryi_20051004] A.Seryi, Comments on feedback kicker location, BDS meeting, October 4, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-10-04/index.htm

101. [CERN_Open_2005_024]  H.H. Braun, J.-P. Delahaye, D. Schulte, ILC Compatibility with Possible Multi-TeV Upgrade, CLIC Note 644, November 2005, http://doc.cern.ch/archive/electronic/cern/preprints/open/open-2005-024.pdf

102. [WG4_mtg_20051110] Phone meeting of WG4 representatives, November 10, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-11-10/index.htm

103. [BDS_mtg_keller_20050726] L.Keller, T.Maruyama, Energy deposition in SC QD0 for 2mrad, BDS meeting, July 26, 2005, http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-07-26/index.htm

104. [CERN_SL_94_41]  V. Juravlev, et al., Seismic vibration studies for future linear colliders, CERN-SL-94-41, 1994, http://www.slac.stanford.edu/~seryi/gm/gm_papers/as_lep.html

105. [PRSTAB_e031001] B. Baklakov, et al., Ground vibration measurements for Fermilab future collider projects, 1998, http://prst-ab.aps.org/abstract/PRSTAB/v1/i3/e031001

106. [Amirikas_Nanobeam05] R. Amirikas, A. Bertolini, W. Bialowons, H. Ehrlichmann, Ground Vibration Measurements and Site Comparison, Nanobeam 2005, http://atfweb.kek.jp/nanobeam/files/proc//proc-WG2b-01.pdf and also see http://vibration.desy.de

107. [SLAC_PUB_11591] Y.Nosochkovy , T.Markiewicz, T.Maruyama, A.Seryi, B.Parker, ILC Extraction Line for 14 mrad Crossing Angle, Nanobeam 2005, SLAC-PUB-11591, bcd_docs/nanobeam_14mrad.pdf

108. [Tauchi_20050816] T.Tauchi, Summary of Detector concept responses to MDI questions, Snowmass 2005,  http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG4/aug17_tauchi_machine.parameter.pdf 

109. [BDS_mtg_20051129] Discussion of pro/cons of separate tune-up dumps. BDS meeting Nov.29, 2005.

Rev. March 12, 2007

Beam Delivery 41

http://www-project.slac.stanford.edu/lc/bdir/Meetings/beamdelivery/2005-11-29/index.htm

110. [BNL_direct_wind] BNL direct wind technology. See http://www.bnl.gov/magnets/Linear_Collider/Magnet_Construction.asp for examples of ILC magnets prototypes and http://www.bnl.gov/magnets/BEPCII/Overview.asp where BEPC-II magnets built with direct wind technique are described.  

111. [Smith_Wands_2005] Richard P. Smith and Robert H. Wands, A Five Tesla Solenoid with Detector Integrated Dipole for the Silicon Detector at the International Linear Collider, 2005, http://ilc.fnal.gov/detector/rd/solenoid/MT_19_Paper.pdf

112. [Vancouver1414] BDS area leaders, BDS report, Vancouver Linear Collider Workshop/GDE Meeting, 19-22 July 2006, http://ilcagenda.linearcollider.org/contributionDisplay.py?contribId=75&sessionId=2&confId=316

 

 

refs not used yet:[X] WG4 Post-Snowmass Summary, draft.

http://www-project.slac.stanford.edu/ilc/acceldev/beamdelivery/snowmass_wg4_summary.htm

Rev. March 12, 2007

Cost and Engineering 1

8. Cost and Engineering.

Convenors: Robert Kephart, Wilhelm Bialowons, Tetsuo Shidara

Session 1: Establishing ILC Project Standards; Tuesday, August 16, 2005.

Presentations

US Project Standards Harry Carter (FNAL) European Project Standards Lars Hagge (DESY)Asian Project Standards Hitoshi Hayano (KEK)LHC experience Jean Pierre Delahaye (CERN)

Speakers should address:1) Engineering standards2) Safety codes, reviews, and standards 3) CAE software (mechanical, RF, accelerator modeling, etc)4) CAD software packages & drawing exchange5) Software license issues6) Project information exchange (EDMS, etc)7) Change control in a multi-regional project

We heard 4 interesting talks by representative from 3 regions and LHC concerning “Project Standards” which stimulated exciting discussion.

Harry Carter stressed the importance of project standards since these have a strong influence on how we work and communicate. Necessity for standard ILC Terminology was also pointed out expecting that a lot of people are/will work on ILC and they must speak the same “language” ; this doesn’t mean speaking English, German, Japanese, etc., but rather “ILC”ese. GDE & the WG’s should create a “dictionary” of ILC terms and post it on to the web that define crisply what is meant by RDR, BCD, ACD, TDR, change control, example civil site, core cost estimate, “operating” cavity gradient, etc.

We discussed engineering standards issues in some detail. Engineering standards are also important and significant issues that must be resolved to realize a big international project. These issues are largely orthogonal from the issues of cost and schedule methodology standards, cost modeling, etc. and probably they don’t have a big impact on the costs. It is highly desirable to create a more specialized group which focuses on Engineering Standards issues.

Next we discussed “codes” issues. Standards are captured as “codes” in the various regions. Although we don’t know the ILC site for some time, therefore, we don’t know which

v. Dec. 12, 2005

Cost and Engineering 2

“regional” standards we should adopt for ILC project, it is important that the ILC make choices early on the establishment of internal “ILC Project Standards”. In order to realize this, we need to make a matrix of codes versus regions; pressure vessel codes, electrical codes, life safety & fire codes, radiation protection, units, drawing standards, etc. Considering the impact of any choice on each region, we will pick something as the “Internal Project Standards”. Since there are some differences of “codes” between regions, everyone has to make compromises, but in general the regions and industry will adapt them. When the site is known, ILC “Project” will make specific adjustments to insure we obey relevant regional codes and laws.

Hitoshi Hayano provoked the necessity to define the “depth” of ILC standardization. According to his definition, there are 3 categories in the depth of ILC standardization.1) Component Standardization: (Build-to-print, all items interchangeable regardless of where they are made)2) Module level Standardization: (Modules with well specified interfaces, but internals may vary depending on who makes it)3) Machine Segment: (Regions take responsibility for all components in entire sections of the machine)

Although it will be worth discussing 3) as an ultimate case of international collaboration, it might make sense when we choose the boundary somewhere between 1) and 2). At the end of the session, we heard an impressive talk “CERN system for LHC” by Jean Pierre Delahaye from CERN. CERN has many experiences on international collaboration which requires project standards. We are fortunate since we have examples to follow not only CERN system but also other international project like ITER.

Recommendations:

Establish ILC internal Project Standards

Choose an EDMS system (Electronic Document Management System)Appoint a group to collect requirements for ILC document management Survey available systems Make a recommendation to GDE very soon (already the GDE plans)

Establish an engineering standards group

Choose common engineering tools e.g. CAD systemCollect requirements for ILC Standard CAD systemsUse 3-D CAD modeling for all drawings including CivilEstablish drawing standards (including units and language)Survey existing CAD software, including interoperability across regionsRecommend a standard ILC system to GDE

v. Dec. 12, 2005

Cost and Engineering 3

Study other large international projects and learnExamine the CERN LHC system and its applicability to ILCExamine the ITER case

Session 2: Cost and Schedule Methodology; Wednesday, August 17, 2005.

A Conceptual Design Report (CDR) including a reliable cost estimate and schedule is to be written for the International Linear Collider (ILC) by the end of 2006. The methodology for cost estimates is different in the three regions North America, Europe and Asia. Thus the main topic of this session was to understand these differences and to arrive at a common approach to cost and schedule estimates for the ILC. Because of the similarities of the ILC and the European XFEL we first heard a talk by Reinhard Brinkman on the cost and schedule methodology for that project. Next, in the following three talks the representative cost estimate methodology in the three regions were presented.

A year ago, the International Technology Recommendation Panel (ITRP) recommended that the linear collider be based on superconducting RF technology. The ITRP report added an important comment regarding this technology: “The construction of the superconducting XFEL free electron laser will provide prototypes and test many aspects of the linac.”3 The superconducting linac of the European XFEL at DESY in Hamburg has about 5 % of the size of the ILC. The length of the machine is 1.7 km, the gradient 23 MeV/m, the energy 20 GeV, the repetition rate 10 Hz and the beam power 650 kW. The construction of the XFEL will start at least two years earlier than the construction of the ILC. For that reason the ILC project can hope to learn at lot about cost and schedule from this project.

Reinhard Brinkmann from DESY gave first a talk on “XFEL Cost and Uncertainties/risk Analysis”. He pointed out that the construction cost estimate is based on the XFEL design described in the October 2002 supplement to the TESLA TDR. The estimate has been revised to the year 2005 escalating with 1.5 % per year from the original year 2000 basis. An update of the cost estimate is ongoing and will be completed by the end of 2005. This estimate will accompany a new TDR which takes into account design changes, different site, more detailed analysis of some of the sub-systems, etc. Personnel costs are estimated on the basis of salaries at DESY in the year 2005 and include overhead to cover basic central services and administration. This is different from the 2002 TDR supplement. The project schedule assumes that final project approval and funding at the European level will be given approximately mid 2006. The plan approval procedure (Planfeststellungsverfahren) with the formal approval for construction and operation and the preparation for placing civil construction orders the will be finished before official project start. The actual construction of the European XFEL could begin about 2008 with beam operation in 2012 and SASE operation in 2014. The estimated total project cost including manpower are 793 M€, year 2005 basis, not including project preparation and escalation over construction period. The

3 ILC – Global Design Effort. The Directors Corner: July 20, 2005.http://www.interactions.org/linearcollider/gde/bbdc20050720.html

v. Dec. 12, 2005

Cost and Engineering 4

capital cost for the superconducting linac, and the civil construction cost are each about 20 % of the total project cost. The capital cost for infrastructure is an additional 10 %. The starting point for the cost estimate was the TESLA Technical Design Report (TDR) published in 2001. Industrial studies were carried out for this report including estimates for the production and treatment of about 20 000 cavities and for the assembly of the cryo-modules. In October 2002 the studies were updated for a smaller number of components and scaling rules were used for single components like tuners and RF couplers. In contrast to the TDR, no large reduction factor assumed for the relative small number of RF system components. Consistency checks with the TESLA linac, present prices, and experience from projects like HERA and TTF were carried out. One different from the TESLA TDR cost estimate is that the capital cost for the manpower was calculated and included the XFEL cost estimate. On average the cost of labor at DESY is 77 k€ per Full Time Equivalent (FTE) year on basis of year 2005 including overhead. For the entire European XFEL project a detailed uncertainty analysis is in progress. The methodology was described for the linac, which is with about 27 % of the total cost (including manpower) and represents the largest Work Package group. A statistical analysis of cost probabilities was performed using a set of uncertainty categories for the costed items. The result was a cost probability distribution in which the estimated cost had a probability of being achieved of 35 %. An additional 10 % risk budget increases the probability close to above 90 %. The cost risk of delays was determined by multiplying the delay with the personnel cost per unit time. i.e. a delay of 6 months is approximately 2 % of the total project cost, which is about 15 M€. The total operational costs are estimated at 85 M€/year. This number includes manpower for operations, cost of electricity, maintenance, refurbishment, and additional R & D and additional overhead.

Next Wilhelm Bialowons from DESY explained the “Methodology of TESLA Project Costs and Schedule and Possible Applications to the International Linear Collider”. The Project Cost and Schedule is described in the TESLA Technical Design Report (TDR).4 Some general remarks were mentioned as a prologue. Cost issues are very delicate and very important for successful construction of ILC (Norihiko Ozaki, LC Forum). It is very important both for project approval and ultimate success that the International Linear Collider management limit the overall project cost. The guideline recommended by Wilhelm is that the ILC management should choose the minimal solution for the ILC and not always the safest solution. Wilhelm also pointed out that cost control is as important as the cost estimate. A further remark was that coupling a new large accelerator complex physically to an existing laboratory site is advantageous from the point of view of investment cost and construction time. The basis of the TESLA TDR cost estimate was explained. The legal basis of the TESLA project is a law, in the form of a state treaty. This law covers the planning for the construction and operation of a linear accelerator. In1998 the two federal German states ‘Freie und Hansestadt Hamburg’ and ‘Land Schleswig-Holstein’, which seek to host the International Linear Collider, ratified the law. The treaty defines e.g. the codes, standards, safety zones, and preliminary investigations. In particular, the law requires that the linear accelerator may be only constructed and operated after the plan approval procedure has been

4 TESLA Technical Design Report (TDR) Part II The Accelerator Chapter 10http://tesla.desy.de/new_pages/TDR_CD/PartII/chapter10/chapter10.pdf

v. Dec. 12, 2005

Cost and Engineering 5

finished. A baseline design was necessary for a reliable cost estimate. The capital cost included all components necessary for the baseline design of TESLA, as described in the Technical Design Report: Part II The Accelerator, Chapter 3 to 9. Not included are the costs for the High Energy Physics detector. The cost estimate complies with the Budgetary Regulations of Germany, which are similar to those of Europe as a whole. The cost of capital equipment was estimated assuming the purchasing power of January 2000. No additional contingency was added. The Value Added Tax (VAT) is not included in the estimation. It is assumed that the manpower required for the various stages of the project will be supplied by the existing manpower in the collaborating institutes. The manpower is quoted separately, and is not included in the total cost. A planning group has been continuously reviewing the technical layout of the system and the cost evaluations. The cost estimates for all major components have been obtained from studies made by industry, and are based on a single manufacturer supplying the total number of a given component. Using this basis the cost for the 500 GeV linear collider baseline design with one interaction region without detector is·3136 M€ (January 2000). The cost estimates for all major components were obtained from studies made by industry. For example: The cost estimate for the niobium for the superconducting cavities was taken from quotes from Wah Chang and Cabot Corporation both in the United States. The estimates for cavity fabrication were obtained from an industrial study by Babcock Noell in Germany. Similar estimations were done for the other main components of TESLA. The price of the vacuum vessel for the cryomodules came from a study by E. Zanon SpA in Italy and estimates for the cavity preparation came again from a study by Babcock Noell. Thomson (now Thales) in France made a mass production study that was used for the klystron cost estimate.

The manpower required for the different stages of the project (design, procurement, fabrication and assembly, testing, installation and commissioning) was estimated mainly using experience from TTF and from other large projects like HERA. It is assumed that collaborating institutes will supply all required manpower. TESLA estimated that a total of 6 933 man years will be required.

The total cost for operations was estimated at 120 Million Euro per year. This includes the electrical power consumption, the regular replacement or refurbishing of klystrons, and helium losses. The numbers are determined assuming current prices and an annual operation time of 5 000 h. Costs for general maintenance and repair have been estimated assuming 2 % per year of the original total investment costs which corresponds to the DESY experience.The construction time of TESLA was estimated to be 8 years. This evaluation is based on the following assumptions: At HERA the experience was an average tunneling speed of 10 m per day. Thus the TESLA tunnel construction could be completed in 3.5 years using 4 tunneling machines. Two years after the start of the civil construction on the DESY site, the tunneling machine will have reached the shaft for the next service hall 5 km away. At that time installation can begin in the first tunnel section. Installation of the first cryo plant, water-cooling systems and other infrastructure into the service hall on the DESY site will start after 2.5 years. Orders for major components will be placed at the same time as the civil construction starts. Between 2 and 3.5 years will be needed to set up the production facilities. Full production rate will be reached after one additional year. The first cryomodules will be

v. Dec. 12, 2005

Cost and Engineering 6

assembled and ready for tunnel installation 4 years after the start of the civil construction. After 5 years the production and installation of all components would proceed at full design rate. The first 16.5 km tunnel section of the linear collider will be completed after 6.5 years. The positron site of the collider will be completed after 8 years. We expect that no more than one year will be required (between financial approval and the beginning of construction) for the bidding and awarding of contracts. Different from the TDR the plan approval procedure with the formal approval for construction and operation and the preparation for placing civil construction orders is assumed to be finished before official project start.Site-specific costs were also considered. Assuming an identical baseline configuration was built (i.e. site at an existing lab, comparable tunnel depth, etc.) at a site different from the DESY site and world market prices apply, the capital costs to build the project elsewhere should be the same in a first order approximation. Because of regional differences, the ILC costs should be estimated without VAT (Value Added Tax), escalation, and contingency. Costs should be estimated for a fixed date (e.g. December 31, 2005). Additonal regional differences in the actual ILC cost estimate could come from differences in geology, codes, laws, labor rates, availability of electrical power and cooling water, or as a result of specific requirements of funding agencies that supply the resources. An example of the latter would the if a funding agency required components to be produced in a region even though that did not result in the lowest cost to the project. The cost driver for the civil engineering is the tunnel construction costs. Nearly fully automatically tunnel boring machines are available that will work in glacial deposits as well as in hard rock. The costs depend mainly on the tunnel diameter and not on the geology if the speed of the machines is approximately the same. Some of differences in geology may be compensated by other factors (e.g. a tunnel in sand below the water table may need no additional air conditioning; tunnels and halls in hard rock are likely stable without reinforced concrete, etc.). Availability of electrical power including the distance to the main overhead line and the availability of cooling water will have an influence on the costs. Typically this is not the case if the site is connected to an existing HEP laboratory.In summary the TESLA TDR costs are a good basis for an International Linear Collider site-specific cost study especially for the superconducting accelerator and the RF system. To make additional progress on costs estimates a new baseline design is necessary. Cost drivers like a twin tunnel, a deep tunnel, and spare length should be avoided. The capital costs should be estimated. These costs should be optimized and not the costs over the lifetime of the machine. Contingency and escalation can be estimated after the site decision using the techniques demanded by the funding agencies.

The next talk was from Rich Stanek from Fermilab who reported on “Cost & Schedule, a US Perspective”. In the opening remarks he pointed out that his talk is more of a guideline to stimulate further discussion rather than a tutorial for how to do it. Globally, there are differences in the cost and schedule methodology, especially in the assumptions, how detailed an estimate is required, basis of estimate, software used. Often the general emphasis of the process is different. It is necessary to agree on the approach for the ILC cost and schedule exercise. Rich recommended that an all-inclusive estimate should be done with nothing hidden or assumed and a more global techniques for estimates is required that acknowledges different approaches in the three regions. Then he reported his observations. If you know the

v. Dec. 12, 2005

Cost and Engineering 7

answer before you perform the cost and schedule exercise then problems can be expected, that means cost overruns, schedule slippage and technical compromises have to be expected which results in a “Blame game”. Set the rules and develop the tools before asking the people to fill in the numbers. Cost estimates for project are now taken much more seriously than in the past. The general methodology was explained with the following block diagram:

Before creating the Work Breakdown Structure (WBS) the rules have to be established and the tools set up. This must be done in parallel to establishing the requirements and the design. Only then can the cost and schedule be reliably estimated. Some other issues have to be resolved. The beginning of the project has to be defined (i.e. when does R&D and planning end and the project begin). How are the expenses associated with assembly and test facilities, scientist salaries, overhead (varies widely across institutions), space, floor or utility charges, and learning curves to be handled? The basis of estimate can be made up by catalog prices, vendor quotes, and comparison to similar projects, time in motion study, parametric analysis and physicist and engineer estimates. The next questions are how to structure the WBS including how deep to go and at what level to allocate costs. It is important to define the risk level (i.e. 50 % probability rule that the estimated cost will be enough). The rest of the talk described the required standards and software tools.

At last Tetsuo Shidara from KEK presented the “Cost Estimates in Japan”. First he explained the organization of ILC-Asia costing. The methodology used for cost estimations was different for the different sections of the linear collider. The injector costs were estimated by cost survey and scaling, the costs for the main linac components like klystrons, modulator, and waveguides were taken from R & D and a resumed reduction factor, the main linac structures cost evaluation based on a production model, the price for vacuum pumps and power supplies are estimated by the survey of commercial data and the costs for the conventional facilities are coming from an industrial estimate. In the Japanese industry cost estimation contingency, escalation, and quality check costs are implicitly included. These are the direct costs. Then overhead cost (general administrative cost) has to be added. The design and management costs are included implicitly in the industrial estimates. Salaries for the

v. Dec. 12, 2005

Cost and Engineering 8

laboratory staff are not included; the salaries for other staff including labor in industry are included. R&D is charged separately and not to the project.

In Japan two assessments tasks where made to estimate the ILC cost. In the first task a technical assessment of the original TESLA 500 design and a cost was done using KEK experience. Items with small technical risks or with global standards of cost resulted in small cost differences between the TDR and the KEK evaluation. Example: infrastructure (excluding cryoplants), Niobium material, surface treatment, klystrons, etc. Some technical items specific to the TESLA scheme showed big cost differences. Examples: cavity fabrication, RF power coupler, cryomodule fabrication and assembling, klystron modulator.There were large differences in the Civil engineering estimates vs expected costs in Japan. ( Japan is more expensive) Standard costs like RF power distribution systems, HV cables, cryoplants are also presently high in Japan. In a second task, a technical assessment and a cost evaluation for 21 000 9-cell superconducting cavities was performed.

Summary of the second assessment task.

This effort concluded that there is a possibility for a drastic reduction in the fabrication cost of the 9-cell cavities including the material cost from the first assessment. Further studies are needed for RF couplers, cryomodules, and klystron modulators.

In the above talks, the speakers were asked to address the following series of questions:1. What is the current practice and methodology for project cost and schedule estimates?

United States: Cost and schedule methods for DOE projects are a direct result of DOE guidelines (DOE Order 413.3) and the Lehman reviews. Cost Estimate includes contingency, basis of estimate, resource loaded schedule, milestones and a way to track them.

2. How do we process for establishing a set of “rules” for ILC cost and schedule estimates?Europe: First we need the “tripod” for the cost estimate of the ILC: legal basis, baseline concept or design and the capital cost estimate basis. The sites in the three regions must be comparable (i.e. coupling to a lab, deep or shallow tunnel, etc.).United States: Need agreement as to which cost and schedule system will be used. Need to have some level of “stable” design. Data needs to be globally accessible.Japan: Do we need “rules”? We really need BCD (Baseline Concept Document), time schedule and sufficiently detailed WBS, as well as guidelines for inter-regional cooperation, information sharing and cooperation with the industrial sectors.

v. Dec. 12, 2005

Cost and Engineering 9

3. How do we handle contingency, overheads, “in kind” contributions, lab or university contributions etc. Europe: The suggestion is to separate the capital cost, the additional manpower, the overheads, the contingency and the escalation (for the last three if really necessary).United States: Contingency should be managed by the project. Cost everything as if it were being done in one country.Japan: Respect the regional difference. Contingency and escalation are implicitly included, and salaries for laboratory staff are not included in Japanese projects. “In kind” contributions might be a baseline for an international collaboration.

4. Include actual estimates for industrial work in a public cost estimate?Europe: They should be included in a confidential way.United States: Yes. Should keep the name of the vendor classified and not release confidential back up calculations or analysis.Japan: Probably no.

5. What is the correct methodology to include profit in estimates for industrial work? Europe: Cost for industrial work must be estimated by industrial studies. The profit must be included in the industrial studies and should not be separated.United States: Needs to be included in a bottom up estimate. Percentage depends on the specific work.Japan: Ranging from 10 – 30 % in KEK. 20 % is nominal.

6. Should ILC commission industrial cost studies of ILC in all 3 regions?Europe: No, the call for tender will by worldwide.United States: Independent cost studies are a good way to certify the accuracy of the cost estimate. The need to do it in all three regions is a function of the belief that the results would be substantially different.Japan: Yes, as long as assuming the construction of ILC by sharing three regions.

7. How do we develop a cost model for ILC?Europe: With the help of industry and international organizations like ITA International Tunneling Association. (I.e. ITA investigated world wide the “Legal and Administrative Issues in Underground Space Use”). Make use of a “Construction Set”.United States: The civil planning group can develop an accurate civil construction cost estimate based on representative sites in each region. We don’t know the “bottom line” cost of building a cryomodule until a factory is set up to produce them and you run it like an assembly line. Industrial studies help. Costs for cryogenic plants and distribution should be able to be estimated based on similar projects and an accepted cost function relationship (cost versus capacity).

During the session it was notice by Maury Tigner that the ITER cost estimate could be a good model for the ILC. Thus on Monday, August 23 an additional talk on “ITER Cost Methodology” by Robert Aymar, CERN Director General and former ITER Director was organized. A brief summary is available under the International Linear Collider Webpage.

“On Monday morning CERN Director-General Robert Aymar addressed Global Group 5 – Cost and Engineering to share his experiences with ITER, an international project that many ILC scientists are using as a model. Aymar described the twenty-year cost estimate and

v. Dec. 12, 2005

Cost and Engineering 10

planning process for ITER. Global Group 5 members had the opportunity to ask Aymar questions that ranged from ‘Will the ILC need an international treaty?’ to ‘Will the ILC take twenty years to plan like ITER?’ Aymar warned about such complications as exchange rates over a period of ten years. He explained that the Japan to U.S. exchange rate varied by more than 50 % over ten years. ‘The estimates start side by side and end up with very different costs because you are not working with the world market,’ said Aymar. He also advised scientists to keep in mind that the government’s timescale is very different from the scientific timescale. ‘International cooperation is a good way to slow down everything,’ he said. ‘As soon as you get through to the diplomats to get an international agreement, you have to follow their timescale, not the technical timescale.’ Aymar’s final words of advice to Global Group 5 were to define a goal. ‘Our goal for ITER was to provide each party with an understanding of an equitable contribution,’ he said. ‘Presenting the cost estimate for the ILC is totally different. You have to put in very strong terms what the goal is for the costing estimate.’ ”5

Recommendations: The essential technical basis is the BCD. For the cost estimate it would be very helpful to standardize the technical solution. That means to use one concept with a small number of options. But just as important is the legal basis. One should aim to get a similar basis as at ITER. Need to agree on a standard method for the “Basis of estimate” for the ILC project. Use this method to identify the “core cost” of the project. The core cost (value) is the best guess of the cost of all materials, labor, and resources the project will need to be built on an agreed upon schedule (50 % probability that this cost will be this number). It should be an all-inclusive estimate, which is public with nothing hidden or assumed. The industrial costs included, but encrypted in a way that the numbers will not be useful for “rigging” subsequent bids. No hidden contingency or risk management funds should be included. The regions can then interpret these numbers in the way the regional funding agencies are used to seeing.

Session 3: Industrial Issues; Thursday, August 18, 2005.

Presentations:

US Industrial Forum: Tony Favale (AES) Euro Industrial Forum: Michael Peininger (ACCEL) Japanese LC Forum: Norihiko Ozaki

Speakers should address: 1) Principle issues in industry for a multi-regional project?2) Can companies deal with different engineering and design standards?3) Technical information sharing 4) Revealing industrial costs estimates vs competition for contracts5) Intellectual property rights for industrial processes 6) Design drawings, CAD packages, etc.

5 CERN Director-General Shares Advice about International Projects and Costinghttp://www.linearcollider.org/cms/?pid=1000062

v. Dec. 12, 2005

Cost and Engineering 11

7) Build to performance specification vs build to print?8) Infrastructure Issues 9) Industrial studies.

Again we heard 3 very interesting talks by representative from Industry that provoked a lot of discussion. Some particular points that were made by Tony Favale were that from an industrial point of view “ ILC is a Project, not a business”. This is true because there is no obvious follow-on business in SCRF cavities at the scale of the ILC. This conclusion has significant impacts on what industry would be willing to do in creating the required infrastructure in advance of and ILC project. Micheal Peininger made the point that Big business in Europe and the US will not necessarily be interested in the ILC as a project. He pointed out that Big industry typically has a Business Plan and stock holder demand a plan for 5-7 years in the future. Uncertain, one-time projects do not fit well in these sorts of plans. This is probably not true for Big business in Japan nor is it true for A&E firms who might do the civil work. (for them, civil work on big projects is a business). As a result, it was suggested that ILC may have to depend on small and medium scale industry. This will have implications on cost & schedule

We discussed infrastructure issues in some detail. As stated above, for industry to invest in the large infrastructure required for ILC there would have to be a follow-on business or market è that the ILC project should plan on paying for all the required infrastructure wherever it is built. We were told that it is unlikely that industry would invest in cryomodule infrastructure before the project receives final approval è it may take 1.5-2 years to build this up before production can begin. These issues need to be included in the ILC cost and schedule estimates. We discussed possible ideas for getting around these problems. One suggestion is that a region might be able to buy the infrastructure beforehand and then industry could bid to use it.

We also discussed issues related to Intellectual Property rights. The general conclusion seemed to be that this was not to be a significant issue for industry. There are several arguments as to why this might be so. First it is not clear that there will be all that many marketable ideas in ILC. Second, they are used to dealing with these issues. The solution is to license this technology. Similarly, transregional information exchange is not an issue.

Next we discussed the issue of open industrial costs estimates. This also seems not to be aproblem for industry and is common in US project cost estimates where it does not seem to cause difficulities. The group recommends that TESLA reevaluate its position on releasing industrial cost estimates for use in the ILC cost estimate.

We also discussed standardization issues. Industry is used to making products for other regions and told us that they could deal with whatever engineering standards the ILC project established. However, the strongly suggested that ILC adopt such standards early in the project.

Several speakers advocated industrial studies. The arguments are that it is important to

v. Dec. 12, 2005

Cost and Engineering 12

understand industrial costs and to examine potential cost reductions. Such studies will cost money. Thus the ILC management will need to think carefully about what studies are needed and when. Its clear that such studies should focus on the cost drivers and items for which there is significant technical risk to the project. It is these places where information is required to make sensible initial cost estimates. The GDE should establish an official point-of-contact for industrial studies in all three regions.

We next discussed risks and how they relate to industrial cost estimates. The message from industry is that we should let them make their mind on risks. Although the ILC management will push for improved performance at lower prices, perceived risks in industry will influence the actual cost of contracts. Cavity Gradient was singled out. ILC should be careful not to set this too high.

Ozaki-san discussed the concept of In Kind Contributions. In this model various regions would provide large components or sub systems as their contributions to ILC. He suggested we follow the ITER system where components categorized as “Key” or “conventional” and both Key and conventional components are allocated fairly across the 3 regions. It was assumed for ITER that the host country would assume the entire cost of infrastructure & civil engineering.

We discussed the cost methodology of ITER. For the cost estimate, ITER uses estimates from all the regions to estimate the “value” of a system or component. This includes material and Services and estimated labor in MY. Costs are those for which there is a 50/50 chance the system can be obtained for that cost. (ie no contingency or escalation included.) Each region can then interpret these core costs for its regional funding agency in the manner to which they are accustomed. The ILC then compute the “equivalent value” all in kind contributions in arbitrary units. (ITER uses 1986 US dollars)

Ozaki-san also discussed cost reduction in industry. He cautioned that the number of components for ILC is small vs industrial scales and that ILC should be careful about assumed cost reductions vs quantity. This is especially true if production is shared among several regions and/or several companies. He suggested that cost reductions might be achieved via joint facilities shared by the regions.

v. Dec. 12, 2005

Cost and Engineering 13

Recommendations:

1) Select a committee to recommend an Electronic Data Management System for ILC. Implement this system ASAP.

2) Select a committee to recommend Cost and Schedule system for ILC. Implement this system ASAP.

3) Form a new working group (distinct from GG5 whose focus is cost & schedule) and charge is this group with establishing engineering standards for the ILC project.

4) Choose common engineering tools e.g. CAD systemCollect requirements for ILC Standard CAD systemsUse 3-D CAD modeling for all drawings including CivilEstablish drawing standards (including units and language)Survey existing CAD software, including interoperability across regionsRecommend a standard ILC CAD system to GDE

5) Create a glossary of ILC terms & definitions and post it on the web. This could expand with time to become the WBS dictionary.

6) GG5 should create a WBS dictionary for the ILC project. Level 1,2,3 managers should be appointed by the GDE such that every major line in the WBS has a person associated with it who is responsible for the cost and schedule estimates for those tasks.

7) GG5 should continue to investigate the ITER Model for cost estimates. ILC should consider adopting it. If so, the GDE must recommend and the ILC project must agree on the ILC unit of measuring cost.

8) It was not obvious why industrial cost estimates cannot be used in an open ILC cost estimate. We recommend all regions consider releasing these cost estimates for use by ILC.

9) As the RDR develops the ILC project should fund industrial cost studies to validate the industrial cost estimates in the project.

v. Dec. 12, 2005

Conventional Facilities & Siting 1

9. Conventional Facilities & Siting

9.0 Overview, principles, general parameter, criteria

9.1.1 Objective

Global Group #4, Conventional Facilities and Siting (CFS) is comprised of members from the Americas, Asian, and European regions. Global group #4 has been charged by the GDE to select, independently by region, a sample site(s) to be used for the development of a regional Reference Design Report. Under the guidance of their home institutions and regional directors, Global Group #4 has completed this task and presents its finding in this Baseline Configuration Document.

These sample sites have been selected so that a study of the construction of the ILC tunnels, conventional facilities and layout of accelerator elements in order to understand how distinctive site features affect the design, performance and cost. The sample sites will not necessary be potential candidates for the final site. This process is intended to assess and identify the sample site to be used for the basis of a credible design and cost estimate for incorporation into the Reference Design Report (RDR) scheduled for completion in late 2006.

9.1.2 Methodology -Assessment Matrix

Each region identified possible sites that allow for the major construction parameters of the Linear Collider. Once sample sites were identified, a rational method employing a matrix to assess the candidate sites was developed and utilized. The Assessment matrix has nine major headings: 1) Site Impacts on Critical Science Parameters, 2) Scientific /Institutional Support Base, 3) Land Acquisition, 4) Environmental Impacts, 5) Construction Cost Impacts, 6) Operational Cost Impacts, 7) Environmental, Safety and Health Issues, 8) regional Infrastructure Support, and 9) Risk Factors. There are approximately one hundred sub-headings that were addressed using available data. The matrix is comprehensive, addressing both global and regional concerns. Not all variables are weighted the same in each region. The weighting of variable and the scoring of the matrix was accomplished within the framework of rules set by each region. 9.1.3 Conventional Facilities Siting Requirements

Many considerations were examined while assessing various sites in each region. Sites were assessed to accept a range of machine parameters developed from past studies.For reference:

(1) Accelerator energy: 0.5 TeV cm Initial, upgradeable to 1 TeV cm (2) Accelerator gradient: 31.5 MV/m Initial (500 GeV CM), 36 MV/m Final, each w/75%

fill factor(3) Accelerator length: ~26 km Initial, including BC1, BC2, undulator, diagnostics, etc.

Rev. March 16, 2006

Conventional Facilities & Siting 2

(4) Crossing angles: 20mrad & 2mrad, length of BDS between wyes 3.84km (5) Damping ring length: 3 @ 6.12 km circumference each, round, 2 on positron side are

in single enclosure.(6) Linac elevation; continuously curved follow earth’s curvature (level)(7) Number of tunnels: 2, with periodic surface buildings(8) Beam line alignment: Follow earth’s curvature, laser straight Beam Delivery (9) Number of interaction regions: 2 Final. One IR hall will be fully costed initially. (10)Overall Length of the Tunnel: ~30 km initial; the site should accommodate ~50 km.

9.2.0. Americas Sample Site

9.2.1 Sample Site DescriptionThere is an emphasis on siting the ILC in Americas at a location which takes advantage of an existing DOE High Energy Physics infrastructure/campus and support buildings.  To that end, five (5) sites were identified for comparative purposes within a 5 county region, in Northern Illinois area near FermiLab.  The search for locations near Fermilab is consistent with general guidance for the Americas region.

The sample site that has the best attributes for constructing the ILC in Northeastern Illinois is situated in the Galena Platteville Rock Formation, with a north-south alignment using a deep-tunnel solution. The sample site is indicated by the shaded area in the figure below.

Rev. March 16, 2006

Conventional Facilities & Siting 3

Figure 2.1a Americas Sample Plan / Section

The general geology of northeastern Illinois is glacially derived deposits overlaying bedrock. The rock layers that are present in the sample site area, from top to bottom, are the Silurian dolomite, Maquoketa dolomitic shale, and the Galena-Platteville dolomites. The dolomite layers of the Silurian are the uppermost rock stratum in the area near Fermilab. These are massive, pure to slightly clayey fine-grained dolomites. The underlying Maquoketa shale is interbedded with clayey dolomite. The dolomite cement in the shale makes this unit strong and resists weathering. The Galena-Platteville is massive layered limestone dolomites. This layer is approximately 100 meters thick and about 100 meters below Fermilab’s ground surface and rises to 30 meters below ground surface at the western edge of the sample site area. Water bearing sandstone lies below the Galena-Platteville dolomites.

The rock layers generally slopes down to the east-southeast about 5 to 7 meters per km. The upper Silurian dolomite found at the Fermilab site is absent several miles to the west and the Maquoketa Scale is absent several additional miles to the west. The only fault in the area is the inactive Sandwich Fault zone which is located to the extreme southwest. The Maquoketa aquitard and the Galena-Platteville dolomites are very tight and inhibit water movement. The Galena Platteville is characterized as a fine to medium grained dolomite that is cherty. Throughout the proposed siting the Galena Platteville varies from 100 to 125 meters in thickness. The Maquoketa shales overlaying the Galena-Platteville dolomite have a low hydraulic conductivity that will act as a hydrologic barrier between upper aquifers and the dolomite. These geologic conditions should provide a relatively dry tunnel, both during construction and during operations. The Galena-Platteville is the most structurally sound rock in the area and, in general, should not require any extraordinary rock support methods. The North-South orientation provides consistently flat sub-surface geological features with proven favorable rock conditions. Extensive experience and knowledge of the geological strata in Northern Illinois provides a level of confidence that a cost effective tunnel can be constructed.

Rev. March 16, 2006

Conventional Facilities & Siting 4

The area under consideration for the Americas sample site is bounded on the east by the Fermilab site and on the west by the geology limiting feature of the absence of the Maquoketa Scales. Limiting features for the north / south direction is less defined both from geologic and surface feature restraints. The western half of the sample site is less populated and is considered mostly Greenfield. The eastern half of the sample site area, and especially routes that are under the Fermilab site are more populated, but there are specific sitings that avoid placing surface features in developed areas.

The Americas sample site is located approximately thirty-five miles west of downtown Chicago. The area surrounding Fermilab is comprised of residences, research facilities, light industry, commercial areas, and farmland. Higher population densities are found to the east with more rural and farm communities to the west. The towns and villages around Fermilab vary in population size from ten thousand to over one hundred thousand people. The surrounding communities have established schools, hospitals, infrastructure support functions and local governments. Major collector roads connect Fermilab to the Illinois toll road system within two miles of its gates. Access to O’Hare International Airport and Midway Airport are via the Illinois toll way system with travel times to these airports less than one hour. Steel mills and other heavy industry are located both in Illinois and its neighboring states. Commonwealth Edison services the Northern Illinois area with a capacity of 22006 MVA. This capacity is made available through both fossil fuel and nuclear power generating stations. In addition, the local power grid is capable of tying to other national power generating sources.

9.3 Asian Sample Site

9.3.1 Sample Site

A large number of site candidates can be conceived for constructing ILC in Asia, since this region encompasses a huge area on the globe, most notably the Eurasian Continent. However, in consideration of the ACFA statement issued at Kolkata in Nov., 2004[1] and the present status of site studies in Asian countries, the Global Group 4 from the Asian Region puts forward a sample site located in Japan for purposes of Reference Design studies. It is noted that use of a Japanese instance as the Asian sample site has been endorsed by the 4th ILC-Asia meeting that was held in November, 2005,

The following requirements were imposed for the sample site:- Existence of firm and uniform geology - Availability of large enough area to accommodate the straight tunnels spanning over

50km.- Absence of active dislocations in the neighborhood.- Absence of wide faults in the neighborhood.- Absence of epicenters of earthquakes exceeding M6 within 50km from anywhere in the

site since AD1500.- Terrain uniformity to maintain the ILC Tunnel depths less than 600m anywhere.

Rev. March 16, 2006

Conventional Facilities & Siting 5

- Availability of large enough electric power to operate ILC.

Of particular importance are the requirements that are related to stable beam operation at the interaction region where the beam sizes are squeezed to the level of a few hundred few nanometers. This led the Asian GG4 to focus its sample site search in stable hard rock areas with small ground motions and ground vibrations, rather than in flat plains which are prone to ground motions and surface activities. Japan is abundant with suitable hard rock areas, since its terrain is dominated by mountain ranges. Civil experts in Japan point out that an excellent accumulation of civil engineering experiences and technologies should allow them to build suitable underground tunnels and experimental halls within the time and budget that are consistent with the construction scenario of ILC. In case of uniform granite, according to past instances in Japan, reinforcement by rock bolts or concrete linings could be often omitted during excavation of underground tunnels and halls, leading to a substantial cost reduction.

The Asian sample site was picked up from a selection which was compiled with assistance from geology experts, while taking into account the considerations above. It satisfies all the scientific requirements for ILC. However, the Asian GG4 takes this opportunity to emphasize that this is merely a sample site (i.e. one instance) which is utilized for studies of a realistic facility design, a specific construction scenario, and the associated project cost. Consequently, in order not to give false impressions of being a site candidate (which is not) when extracted and quoted out of context, the Asian GG4 refrains from offering any maps of this sample site except the schematic cross section view as shown in Figure 3-1-1.

According to published reports and field studies, most part of our sample site is expected to be located in a uniform, solid rock area. No problematic active dislocations or faults have been reported. However, the details of the geologies are known to affect the construction schedule and the cost, in addition to the performance of the accelerator. Therefore, the Asian GG4 recognizes the need for more detailed site field studies including investigations with pilot boring.

The surface condition which is relevant to the construction and performance of ILC is summarized as follows: The ILC tunnels are to be built underground whose surface is covered mostly by woods and forests. Otherwise, the neighborhood is dominated by an agricultural area which is crossed by occasional local paved roads. Only a few local residences exist along the tunnel route. Big streets with heavy traffic or large river systems are absent, leading to very few sources of natural or human-made sources of vibrations. An adequate amount of flat surface areas exist to accommodate surface facilities.

In this sample site the ILC tunnels are excavated in a moderate plateau area (altitude variation within 600m). The access tunnels are arranged every 5km along the ILC tunnels as inclined tunnels. The average length of the access tunnels is approximately 1km. The surface routes to these tunnel access entries will be extended from existing local paved roads. The tunnel boring machines will be introduced through access tunnels (10km apart) whose inclination is shallower than 10%. Most of other construction equipment such as electric and mechanical machines will be delivered through this route, as well as the accelerator components during

Rev. March 16, 2006

Conventional Facilities & Siting 6

installation. Inclination angles of other access tunnels may reach maximum 20%. Figure 3-1-2 shows a conceptual view of the tunnel structure in the neighborhood of an access point. The thickness of the ground above the tunnel ranges from 40 to 600m. The tunnel is about 150m deep at the IP.

Figure 3-1-1

　　Figure 3-1-2

Availability of stable electric power is a critical element in ensuring reliable operation of the cryogenic systems of ILC. It is noted that electric power companies in Japan has an excellent track record in that regard. In 2002, the frequency of power outage for a customer averaged over the entire nation was 0.17 times/year, the average duration of the power outage was 13 minutes.

Reference [1]: http://ccwww.kek.jp/acfa/document/3rdLC.html

Rev. March 16, 2006

Conventional Facilities & Siting 7

9.4 European Sample Sites

9.4.1 Sample Sites

For the European region, it has been decided, at BCD stage, to consider two sample sites. Because of the great benefits for such sites being located close to existing Particles Physics Research Laboratories, one site will be situated close to CERN, near Geneva on the French-Swiss border, and the other one close to DESY, near Hamburg in Northern Germany.The assets and infrastructure connections for each site are briefly described in the following paragraphs.

9.4.2 CERN sample site

9.4.2.1 General considerations

The site benefits from the proximity of: CERN existing site with offices, workshops, laboratories and storage space, including

a 400 KV electrical supply. The city of Geneva with its international airport. Close highway and railway network connections. Headquarters or branches of firms active in fields of interest for ILC construction and

operation (Consultants, Work Contractors, Suppliers). Both Switzerland and France being experienced in accommodating international organizations, this site would also benefit from professionalism and flexibility of their Authorities in dealing with settlement of large scale new entities. However a number of environmental related issues will need to be treated and solved through an Environmental Impact Study, building permits, discussions, agreements, etc… with the local and central Authorities in both countries. This applies in particular to land acquisition (free for CERN projects) and environmental protection measures linked to such a large scale project as ILC. The positioning of the shafts and related surface buildings will need to take into account the existing protected villages and vineyards situation on both sides of the border, particularly in Switzerland.

In terms of geology, this site will benefit from the stability and water tightness of the bedrock called “molasse” (sandstone of variable characteristics), which stretches between Jura Mountains on one side and Lake Leman on the other.Most of the faults in the bedrock are non-active ones but underground depressions (or fossil valleys) are to be avoided as far as possible by deepening the tunnels. Of course thorough in situ site investigation will need to be conducted before any final decision on the layout is taken. In particular there is a risk of encountering natural hydrocarbons and/or gas in such rock, which will also need to be investigated. Because of this configuration, earthquake and man-made vibrations do not represent so much of a problem in the specific case of this site.However, the geology dictates a deep tunnel solution, in order to stay in stable and watertight rock.

Rev. March 16, 2006

Conventional Facilities & Siting 8

Layout and sections

The longitudinal section and typical cross section are presented on figures Nos 4.2a and 4.2b hereafter.

The tunnel will stretch over Département de l’Ain in France, Cantons of Geneva and Vaud in Switzerland. It will be sited roughly on a parallel line to the Jura Chain and the Swiss side of Lake Leman. Its South West end would be located in France close to the Rhône river bed, and its opposite end North East of the town of Gland in Canton de Vaud, Switzerland, in a 50 km overall length configuration. In the baseline configuration, the two 6 km long damping rings will be excavated in the same “molasse” bedrock at the same depth as the linac tunnel to which they will be connected. An angle of 20 Mrad has been taken into account between the linac tunnels on each side of the intersection point. The central experimental area will be grouped with the LHC area of Point 6 to take advantage of the available facilities. According to baseline options, an earth curvature profile has been adopted for the Linac tunnels while the Beam Delivery System will be housed in a 5 km laser straight tunnel. In this configuration, the average tunnel’s depth will be 120 m with variation of ± 15 m due to land shape.

As an alternative, a fully laser straight profile will be proposed, which will be tilted by about 2 mrad. In this case, the tunnels will be positioned as closely as possible to the surface, while passing under the Gland depression. The tunnel depth would then vary from about 30 m on the Rhône river end, to about 130 m on the opposite end, thus producing cost savings in terms of civil engineering (some shafts being shorter) and of services (electrical distribution, cooling and ventilation, etc.). However this alternative may give rise to some cost increases in terms of cryogenic equipment, which will be compared to these cost savings during RDR stage.

Figure 4.2a – Longitudinal solution of the CERN sample site project

Rev. March 16, 2006

Conventional Facilities & Siting 9

A two tunnel configuration has been retained for the linacs, according to baseline options. This configuration improves reliability and safety of the linacs equipment but final dimensions will need to be carefully adjusted to satisfy all requirements while keeping the minimum possible dimensions for economical reasons. The proposed diameters are 4 m for the service tunnel and 3.2 m for the beam tunnel, to be confirmed by equipment designers.

Fig.4.2b – Typical cross sections of the European sample sites projects

9.4.3 DESY Sample Site

9.4.3.1 General considerations

The site benefits from the proximity of: DESY existing site : offices, workshops, laboratories and office space including

110/220 kV electrical supplies, The city of Hamburg with its international airport and seaport, Close highway and railway network connections Headquarters or branches of firms active in fields of interest for the ILC construction

and operation (Consultants, Work Contractors, Suppliers).The proposed layout being at the same location and in the same shallow position for two thirds of its length as for the TESLA project, it will benefit from negotiations already carried out with the two concerned states of Hamburg and Schleswig Holstein, who granted planning permission agreed. However, as in the case of the CERN site, an Environmental Impact Study will be necessary, to be followed by building permit procedures, should the site be selected. As in the case of the CERN site, the necessary plots of land for surface infrastructure around the shafts will be purchased by German administrations, before being put at the disposal of the project. The potential difficulties of this sample site are linked to the fact that the Linac tunnel will be placed in a shallow

Rev. March 16, 2006

Conventional Facilities & Siting 10

position, implying protection measures with respect to the existing surface structures, including a 18th century old church near the village of Kellingen.In terms of geology, the major part of the tunnel will be bored in quaternary sand and a smaller part in marl. The tunnel will be situated below the ground water table over nearly its entire length. However thorough in situ site investigations will need to be conducted before any final decision on the layout is taken. Investigation will also need to be carried out as concerns the compatibility of tunnel deformations due to hydro-mechanical variations and ILC dimensional tolerances. Northern Germany is a low level seismic zone but potential sensitivity of the ILC equipment to some existing cultural sources of vibrations would need to be assessed. A shallow type positioning of the linac tunnels and other underground infrastructures of the project is highly recommended for this sample site in order to avoid having to cope with high water pressures which carry cost and risk implications.

9.4.3.2 Layout and sections

The ILC layout will follow closely the TESLA layout on the first 32.8 km, and then be extended to 50 km in the same specific direction as indicated on Figure 4.3a. The TESLA route was actually investigated up to a length of 70 km and could be extended to this length including shafts and surface building at about every 5 kms. For reference, the layout and longitudinal section of the TESLA project are presented on figures 4.3a and 4.3b hereafter. From DESY close to the border of Hamburg, the ILC tunnel will cross the whole county of Pinneberg, then the county of Steinburg, following the North-North-West direction. It will be sited under a rather flat agricultural ground scattered with a number of villages and small towns. No fractures or faults are expected to be encountered, thanks to the rather homogeneous geology of the site.An angle of 20 Mrad has been taken into account between the two linac tunnels.The location of the central experimental area will of course depend on the total final length adopted for the project as the starting point will remain the HERA accelerator. It is likely to be sited near the village of Vosslock, some 24.3 km away from the Desy Site.

Rev. March 16, 2006

Conventional Facilities & Siting 11

Figure 4.3a – General layout of the TESLA project site

The depth of the tunnels varies from 7 m up to some 30 m at the deepest point. The tunnels will be horizontal in each of their locations, which means that it will follow earth curvature except in the BDS area where they will adopt a laser straight profile. As for the CERN sample site, there will be a total of 11 access shafts created including the central one leading to the experimental cavern. In the same way, the possibility of having one of these shafts replaced by an inclined access tunnel will be investigated.

Figure4.3b – Longitudinal section of the TESLA projectA two tunnel configuration for the linacs has been retained as in the case of the CERN sample site (see fig. 4.2.b)

Rev. March 16, 2006

Conventional Facilities & Siting 12

9.5 Conventional FacilitiesThe conventional facilities for the ILC project will consist of above and below grade structures, utilities and support facilities generally found at high-energy physics accelerator laboratories.

9.5.1 Facilities Design

9.5.1.1 Main Tunnels

Parallel twin bored tunnels are planned. Preliminary tunnel diameter, depth, type and method of excavation for each region are listed in the table below.

The beam tunnel enclosure will house main linac cryostats and beamline components. The support tunnel enclosure will be a parallel tunnel that houses klystrons, modulators, power supplies, instrumentation, controls, and miscellaneous support equipment. The support enclosure allows personnel to access critical components during beamline operation for maintenance and operations purposes. The klystrons distribute RF to the cryostats through “duct like” wave-guides that pass between the two tunnels in the range of 50 to 60 cm diameter penetrations spaced at 36 meters along the main linac. Alcoves and small size caverns for electrical and cryogenic equipment will be excavated by using roadheaders or rock breakers, then concreted in situ in a traditional way.

Preliminary Design Data Americas Region

Asian Region

European Region (CERN)

European Region (DESY)

Accelerator Tunnel Diameter

4 m 3.2 m 3.2 m 3.2 m

Service Tunnel Diameter 4 m 4 m 4 m 4 mDepth of Tunnel (Range) 120 m to

150 mSeveral tens

to 200 mBaseline 100 to 130mAlternative 30 to 130 m

7 m to 30 m

Tunnel Excavation Method

TBM TBM TBM with precast concrete

lining

Hydro shield TBM with precast

concrete liningType of Rock Galena

Platteville Dolomite

Granite Bedrock in mountain

Molasse sandstone rock

of Lemanic Plain

Quarternary sand for the majority,

and marl for small part

Length of site 50 Km or more

50 Km or more

50 Km or more*

50 Km or more*

*Possible extension of CERN and DESY sites up to 70 Km, provided that a decision as to the final length is taken before the start of the detailed design, the layouts for these two sites being fixed at one end.

Rev. March 16, 2006

Conventional Facilities & Siting 13

As decided, the civil engineering works for the linac (and service) tunnels will be carried out in two phases corresponding to 500 GeV and 1 TeV levels of energy. The aim of this is to reduce to a great extent the initial cost of the project. Phase two of these works will start once the necessary budgetary resources have been gathered, possibly years after phase one has been equipped, commissioned and utilized. This will imply that a number of precautions have to be taken to avoid disturbance to the existing running machine from the second phase works.It is to be noted that the total cost of civil engineering works (phases 1 and 2) will be significantly higher than if they have been carried out in one single phase.

AmericasThe final lining for 1/3 of the tunnel length is expected to be shotcrete over-laying a drainage mat to control water inflow. The remainder of the tunnel will be rock with a sealer applied.

AsianBecause of the rigid bedrock, a reinforcement of the tunnels by the rock bolts or concrete lining may not be required. Figure 5.1.1b shows the longitudinal view of the Main Tunnel.

EuropeanThe final lining of these tunnels for CERN site as well as for, DESY site, will be formed using precast watertight concrete segments laid in place by the TBMs during excavation.

Figure 5.1.1a Cross Sectional View of the tunnels (Americas Region)

Rev. March 16, 2006

Conventional Facilities & Siting 14

Figure 5.1.1b Longitudinal View of the tunnels (Asia Region)

Figure 5.1.1c Before and after construction View of the tunnels (Asian Region)

Rev. March 16, 2006

Conventional Facilities & Siting 15

9.5.1.2 Damping ring tunnels

According to baseline design, there will be two Damping Ring tunnels of 6 kms circumference each, situated in a tangent position to each end of the phase 1 linac tunnels.On the electrons side, the Damping Ring will be fitted with one line including bending magnets and other equipment, while on the positrons side, two lines could possibly be stacked in the same tunnel. Their inner diameter will be 3.2 m, to be confirmed by equipment designers. They will receive two dedicated access shafts, which will be also used for air treatment and services. Civil Engineering works for these tunnels will be carried out in the same way as for the main tunnels (see 5.1.1 above).In the case of the CERN and DESY Sites, cost saving alternatives will be considered using respectively the existing SPS and HERA tunnels to house the Damping rings.

9.5.1.3 Access Shafts

The access shafts will provide personnel and equipment access from the surface to the service/accelerator tunnel as well as space for utilities (cables, pipes, air ducts) and cryogenic lines that feed in the service/accelerator tunnel. The shafts will be used for a transfer of the boring machine and to bring out muck generated by tunnel boring during the construction period. The shafts are spaced at approximately 5 km. At least 2 to 4 of these shafts will be sized to accommodate both cryomodule and TBM access. The final numbers and shaft dimensions shall be confirmed at Reference Design stage.

AmericasAccess Tunnels will be of a vertical access shafts. At least two additional shafts will be needed for ventilation and other services to the damping rings.

AsianThe IP and one of the access points will be accessed by the vertical shaft. Other access tunnels will be of the horizontal sloping shafts to match the mountainous geological features and bored by NATM. Figure 5.1.1c shows the cross sections of the access tunnel during construction and after completion.

EuropeanAccess Tunnels will be vertical shafts for both sites. On CERN site, the excavation through the upper layers of moraines will be carried out under the protection of diaphragm walls or alternatively by the use of ground freezing techniques. Through the bedrock, top down excavation methods with rock breakers and possibly road headers, will be used, the concreting of these shafts will then be carried out with non stop shielding form works methods, which have proved fast and reliable for the concreting of the LHC shafts.On DESY Site, the excavation works of the shafts will be carried out in a traditional way under the protection of diaphragm walls. Concreting will be done in situ also in a traditional way.

Rev. March 16, 2006

Conventional Facilities & Siting 16

9.5.1.4 Access Halls/Caverns and Machine Rooms

A cavern space or halls to house cryogenic equipment, electrical equipment, equipment for air-conditioning, heat exchangers, cooling water drainage and space to allow for distribution of cables, pipes and ducts are to be constructed at the bottom of each shaft, at an interval of approximately 5 km. A link to the accelerator tunnel will be created at the same location to allow for transfer of services and pieces of equipment.

9.5.1.5 Detector Hall

Each of the two interaction zones will consist of an above grade assembly building and below grade collision hall. The current required dimensions for inside space for each IR region collision halls below grade is 33 meter (108 feet) wide by 62 meter (203 feet) long by 30 meter (100 feet) high. The grade-level assembly area will be for detector subassembly and will include a heavy crane. Steel multi storey mezzanines will provide areas for computer, control room and office facilities. Other utility areas will house process water systems, electrical power services and air handling equipment.

Americas Rock will be formed into a structural element with rock bolts. Excavation will use the drill and blast method.

Asian Rigid and thick granite bedrock will be able to hold such a large cavity without structural reinforcement. Excavation will use the drill and blast method. EuropeanAt CERN, the floor level of the two underground experimental halls is approximately 130 m below ground, hence these halls are fully situated within the stable molasse rock. For Europe-CERN sample site, drill and blast method of excavation will be avoided for environmental reasons. These halls will be excavated using rock breakers and roadheaders, then concreted in situ.For Europe-DESY site, the floor depth is only 20 meter below ground level. Therefore all works will be carried out from the surface on the whole extent of the hall, under the protection of a diaphragm wall put in place prior to any other works. Only one experimental hall (corresponding to 20 mrad) will be constructed in a first phase.The construction of the second hall, including its connecting tunnels will be carried out later, while taking a number of precautions to avoid disturbance to existing machine and detector structures and equipment.

Rev. March 16, 2006

Conventional Facilities & Siting 17

Figure 5.1.4a Interaction Region

9.5.1.6 Surface Facilities

Americas RegionFor the Americas region, a central campus will be located near the detector halls. Space at Fermilab will be used when feasible. The campus area will include office spaces and facilities to build, operate, service, and maintain the project. Additionally there will be surface buildings at each shaft to house the Cryogenic and LCW equipment.

Asian RegionThe research and administration building as well as various supporting facilities including the main substation will form the central campus and be located near the center of the linear accelerator. Other surface facilities include the Entrance Halls to the Access Tunnels and the Machine Rooms which house equipment for ventilating and air-conditioning system, cooling water system and drainage system, both located at an interval of approximately 5 km. The cooling water system will be located near the Machine Rooms. Also residential facilities must be provided for a large number of scientists, engineers and their families.

European Region-CERNA central campus will be located near an existing LHC site called Point 6, on the French side of the border (commune of Versonnex). It will thus benefit from existing facilities at Point 6, but even more from the CERN site at Prévessin and its surrounding land situated in France only 4.5 km away. The campus will include a control building for the detector, a cryogenic

Rev. March 16, 2006

Conventional Facilities & Siting 18

compressors building, a cooling and ventilation plant building, a set of cooling towers with pumping stations, and an electrical sub station. Structures containing noisy equipment (ventilation and cryo-equipment) will be made of concrete with an internal lining made of absorbing material in order to mitigate the propagation of noise to the environment. The site will be fitted with ancillary facilities such as fences, roads, car parks, gates with access control devices and plantations. Every adjacent surface area to the 10 other machine access shafts will be fitted in the same way, except for the detector control building.The plots of land needed for this purpose will be minimized in size in each case to mitigate the environmental impact and to keep the related costs low.Wherever possible, the spoil generated from the excavation works will be deposited on the surface within the areas made available for the project. However, it may be necessary to find external dumping areas. All areas will be landscaped in accordance with the local environmental requirements.

European Region-DESYA central campus will be located near the village of Vosslock as mentioned in 4.1.2.2. In addition to the detector hall, this campus will include similar buildings and ancillary facilities as listed for the CERN sample site. The same applies to the 10 other access shafts areas. As for the CERN site, most of the generated spoil will be used for landscaping as dictated by the local environmental requirements.

9.5.2 Power Distribution

Americas RegionUtility grid power will be supplied at no less than 230 kV from two or more separate transmission sources. The primary site distribution voltage will be not less than 34.5 kV supplied from two Main Substations located ~26 km apart. Cryogenic Substations will supply large motors at no less than 4.16 kV. RF power transformers will supply modulators at no less than 10.5 kV. Various auxiliary equipment will operate at 0.48 kV, 0.208 kV or 0.12 kV. Main substations, cryogenic substations and related supporting auxiliary substations will be primary and secondary selective with an availability of 99.99% or higher and be configured to maintain continuous power to critical loads while transformers and switchgear are de-energized for inspection, repair or maintenance. The total self cooled capacity of the Main Substations will be not less than 300 MVA initially. Site power for life safety systems will be supplied from standby generators.

Rev. March 16, 2006

Conventional Facilities & Siting 19

Figure 5.2a Primary Power Schematic

Asian RegionThe 275 kV power will be supplied from a power company and transformed to 66 kV power at the Main Substation built on the ground surface. The site power is assumed to be 322 MW. The power then distributed to 10 underground substations and lowered to 6.6 kV. The 6.6 kV power will be distributed to a large number of transforming units at various locations of the facilities to be further lowered to 400 V, 200 V and 100 V for driving the RF power sources and other equipment. Standby generators of 6.6 kV will be provided at each substation to supply emergency power to the smoke exhaust fans, the drainage pumps, the emergency lights and other fire fighting equipment.

European Region-CERN The power will be supplied from the French supplier EDF 400 kV grid. The 400 kV line feeding CERN at present has a minimum short circuit power of 7600 MVA and a transport capacity of more than 1000 MVA. The power requirements are assumed to be between 300 and 375 MW. A new 400 kV/66 kV substation will be needed. It should be situated next to the existing one on the CERN Prevessin site to be located at the end of the existing delivery 400 kV. A transport line with a rated voltage of 66 kV will be constructed along the tunnels. The line will be realized using XLPE extruded cables laid in cross country trenches. This line may be extended to a loop, if deemed advantageous. At each access area a 66kV/24 kV substation will be constructed, supplying the local needs. The distribution may be directly at the 24 kV level, for large individual users, or via 24 kV/ 0.4 kV transformers. A 24 kV cable link will be installed all along the tunnel to supply the tunnel alcoves, needed for service installations. These alcoves will have to be created to house the 230V/400V power centers, supplied from 24 kV ring main units via 24 kV/0.4 kV transformers. The distance between neighboring alcoves should not be more than 1 km. The necessary compressor motors for cryogenic systems and chilled water will be supplied from a 3.3 kV system for units over 350 kW and supplied from 66/3.3 kV transformers.A diesel generator set will be installed in each of the access areas. The loads that cannot accept power cuts in case of mains faults will be protected by UPS systems.In the central area a dedicated 24 kV system will be generated, either directly downstream of

Rev. March 16, 2006

Conventional Facilities & Siting 20

the 400 kV systems or downstream of the 66 kV systems, depending on the load and its distribution.A separation in a pulsed and a non-pulsed system may have to be envisaged to keep perturbations away from certain loads. Also reactive power compensation and harmonics filtering will probably be needed, depending on the non-linear load.Signaling will be at the 48 V DC level. Safety systems will be fed from double charger/battery systems with autonomy corresponding to the standards. All alarm and fault signals will be by potential free contacts, normally closed, open on fault. The network monitoring will be via an industrial SCADA system. All main equipment will communicate via serial bus using standard protocols. Essential safety signals will be forwarded in parallel hardwired.

European Region- DESY For the power distribution to the electrical substations the existing infrastructure of 110 kV and 220 kV power lines will be used. New distribution stations will be built under the power lines, equipped with high-to-medium (20 kV) voltage transformers. The power will be transported to the service stations by 20kV power cables buried in the ground along existing roads.

The choice of medium (20 kV) and low (400 V 3-phase, 230 V single phase) voltages follows the industrial standard in Germany. There will be a switchboard plant with 20 independent 20 kV to 400/230 V transformers. The 20 kV transformers are air cooled and installed outside next to the service buildings. An auxiliary supply from the public grid is also foreseen. The modulators will be supplied from 690 V switchboard plants installed in service rooms in the modulator halls. The helium compressors as single high-power consumers require a special switchboard at 6 kV.

Since the spacing between access shafts is about 5 km, power has to be transported into the tunnel over a maximum distance of 2.5 km. This is too long to be efficiently done at low voltage level. Therefore only the first ~ 500 m tunnels will be supplied with low voltage cables from each hall. For the remaining 2 km two medium to low voltage transformer stations are foreseen in the tunnel, supplied by 20 kV cables.In order to reduce the fire risk, fire retarded and non corrosive (FRNC) cables will be used. In addition, they will be covered with fire-proof paste. The current density in the cables is relatively low, which reduces the fire risk.

9.5.3 Ventilation and Air-conditioning System

The tunnels will be provided with ventilation and cooling to maintain suitable temperature, humidity, and adequate air movement. There will be air handlers at the surface that treat the incoming air to obtain the required temperature and humidity in the tunnel. There will be chilled water cooled fan-coils in the service tunnel to remove the large amount of heat generated by a number of equipment in that tunnel, whereas the accelerator tunnel will be cooled only by air from the surface air handlers. Due to different climatic conditions, there will be difference in the preliminary methods of providing this system for each region. These

Rev. March 16, 2006

Conventional Facilities & Siting 21

are depicted in various figures as follows.

Figure 5.3a Americas Region Ventilation Scheme

Figure

5.3b Asian Region Ventilation and Cooling Scheme

Rev. March 16, 2006

Conventional Facilities & Siting 22

Figure 5.3c – European scheme of the ventilation and air conditioning principle in the service tunnel

Figure 5.3d – European scheme of the ventilation and air conditioning principle in the beam tunnel

9.5.4 Cooling Water System

This system will cool water-cooled components such as Klystrons, Modulators, and fan-coils in the service tunnels. Detailed information such as the loads, type, locations, and number of equipment, as well as the required supply water temperature to this equipment are needed in order to proceed to the next level of design. The system will be designed for all-year operation. The heat rejection from the water system and method of design will vary per region, due to climatic differences, and system pressure requirements due to varying depth in different regions. The preliminary cooling water design is still evolving among the different regions and any discrepancies with regard to the extent of the scope of the system will be resolved and finalized in the next reference design level.

For the most part, the system will have cooling towers/hybrid towers/dry coolers and/or chillers in the surface plants located approximately 5 km apart, and heat exchangers, pumps either in the surface or the tunnel level. The various schemes and preliminary methods are as follows; For the Americas region, there will be a primary, secondary and tertiary water loop with final heat rejection to a hybrid tower or dry cooler supplemented by air-cooled chillers during peak season. For Asian region, heat rejection is cooling towers at the surface and heat exchangers and related equipment in the tunnel level. For European Sample site in CERN, the final heat rejection equipment is induced-draft evaporative cooling towers with plume abatement techniques. The European sample site in DESY will use a hybrid type to minimize

Rev. March 16, 2006

Conventional Facilities & Siting 23

water evaporation and avoidance of plume. These entire schemes will be further investigated in the Reference Design Stage. The following diagrams 5.4a, 5.4b and 5.3b illustrate some cooling systems in the different regions.

Figure 5.4a Americas Region Cooling Water Scheme

Figure 5.4a European Region Cooling Water Scheme

Rev. March 16, 2006

Conventional Facilities & Siting 24

9.5.5 Handling equipment

Whenever possible standard industrial transport and handling equipment shall be chosen for the installation and operation of the ILC.

Depending on the required lifting capacity and the installation depths either material lifts or Electrical Overhead Traveling (EOT) cranes shall be installed. As a general approach the following baseline should be considered:

Material lifts for installation depths of more than 50 m and for capacities up to 25t. EOT cranes for installation depths of less than 50 m of for capacities of more than

25t.For safety aspects and practical reasons material lifts are the preferred solution although EOT cranes offer higher flexibility in material handling. For example, the lowering of equipment that is longer than the shaft diameter – then the equipment can be inclined to pass via the shaft.The access for personnel to the underground area for vertical shafts shall be done with standard “long traveling” passenger lifts while the electric automobile will be utilized for the horizontal sloping tunnels. A combination of goods and passenger lifts may be envisaged. The following capacities and dimensions are given as indicative only and will need confirmation at Reference Design stage. Most of them depend on final choices to be made as regards the cryogenics RF modules. Equipment in the shafts :

o Experimental shaft : (diameter 20 m)- One 80 T capacity EOT crane

o Personnel access shaft to Experimental cavern (diameter 6 m)- One 6 persons capacity lift

o Access shafts for middle of each linac, diameter 20 m- Two 30 T EOT cranes- Two 6 person capacity lifts

o Other access shafts (diameter 15 m)- Eight 10 T capacity good lifts- Eight 6 person capacity lifts

Equipment in the underground structures- One 80 T EOT cranes in the experimental cavern (possibly 2)- Several personal lifts for access to various levels of the steel platforms- Electrical lifting hoists in other underground caverns or alcoves.

Equipment in surface buildings - EOT cranes mentioned here above for shafts are installed in the related covering

buildings- Cryogenic compressors building : 10 to 20 T EOT cranes- Pumping station for the cooling towers : 10 T EOT cranes- Cooling and ventilation plant building : 10 T EOT cranes

The layout of the horizontal tunnel transport equipment will mainly depend on the characteristics of the material to be transported and on the space restrictions. In principal, two

Rev. March 16, 2006

Conventional Facilities & Siting 25

types of transport systems are possible. Either floor transport equipment (with or without guiding system) or suspended transport equipment such as monorail systems.A possible configuration consists in the use of floor transport equipment in the beam tunnel and a monorail system in the service tunnel.The general principles listed above also apply to the DESY sample site, with some exceptions, like most of the proposed cranes and lifts will be slightly simpler owing to the smaller depth of the access shaftsIn the case of Asian Site, most of the access shafts are not vertical shafts but are horizontal sloped tunnel where the trucks or trolley with rails will be used during the construction and the electric automobiles will be utilized during the operation.

9.5.6 Safety and Fire Fighting Systems

The issue of life safety and exiting from a deep tunnel configuration for the ILC is difficult because accelerator tunnels of any kind do not neatly fit into any or the prescribed categories identified in any of the prevailing national building codes. Office buildings, restaurants, hospitals, etc all have written requirements for the number of, and distance to, exits for a specific building design. Even the most lenient of these requirements would be inappropriate and cost prohibitive for an ILC tunnel design. The use of computer modeling can be used as a tool to demonstrate the viability of exiting for a unique design configuration such as an accelerator tunnel(s).

This being the case, existing transportation tunnel configurations may be an avenue to be pursued as we are in the case of a twin tunnel configuration. Many of these existing traffic tunnels have been designed with connecting passages that can be used in the event of an emergency. A system of fire protected accesses between the tunnels could provide a safe exit path away from a fire or other hazard and to the closest available access shaft to the ground surface. This is much like a fire protected stairway leads to the outside of a high rise buildingIn either case, the issue of life safety and exiting configuration and its approval will need to be addressed early in the design process to insure a complete and adequate design is developed. At a minimum there will need to be cross connections between two tunnels at a spacing to be determined during the next design stage. The service tunnel, the beam tunnel, and the damping ring tunnel will be provided with emergency lighting system, smoke exhaust system, fire suppression sprinklers system, and smoke and heat detection. All tunnels will have Access Safety & Access Control System, Audible Emergency Evacuation system, Automatic Fire Detection & protection system, Oxygen Deficiency Hazard (ODH) detection system and Alarm Transmission system

9.5.7 Survey and alignment

The geodetic reference networkThe sites will be covered by a geodetic network. The reference points will be installed as close as possible to the shafts, and will be determined from GPS measurements. An additional

Rev. March 16, 2006

Conventional Facilities & Siting 26

leveling network will be set up in order to complete the GPS measurements which are not accurate enough in the vertical direction. The shape of the geoid will be also determined carefully.

Because most of the CAD software doesn’t take the real shape of the earth into account, the transfer functions between the local astronomical systems of co-ordinates (related to the vertical), the geodetical ones (related to the ellipsoid) and the Cartesian CAD system of co-ordinates will be written as soon as possible and introduced into the CAD software prior to the studies. The geometry will be transferred from the surface to the level of the tunnel through the shafts. Dedicated space will be preserved permanently free during the life of the project in the shafts, allowing the link to be possible at any moment for the first alignment of the accelerator but also for the maintenance of the alignment.

An underground reference network will be set up on the most stable part of the tunnel, its slab. The slab of the tunnel will be filled with concrete and will take the full benefit of the vault effect for a maximum stability. The reference points will be measured from point to point, and the traverse will be calculated between two consecutive shafts. The distance between the shafts will have to be optimized according to the accuracy requested first for the civil engineering works –this is likely to have an incidence on the size of the tunnel to be built- and second for the alignment of the components, compared to the possibilities to reach the accuracy at mid distance between two shafts. If for the civil engineering works, traverses based on gyroscopic measurements will probably provide the requested accuracy during the boring phase of the tunnel, studies will have to be performed in order to determine the method for aligning the components within the requested tolerances. Clearly, a solution has to be found for measuring long distances with a relative accuracy of about 10-7, and the combination of angles and wire offset measurements has to be carefully studied.

The geometrical network of the damping rings will be connected to the surface in at least three roughly equidistant places. Four places would be better, due to the curvature of the tunnel which limits strongly the length of sights. The alignment techniques used on the SPS (CERN) could be used as the size of the machine is similar. Nevertheless, studies will have to be carried out in order to adapt the methodology to the size of the tunnel and to the specific tolerances of alignment of the components. An enlarged tunnel in the curved parts of the rings would ease for the alignment.

Alignment of the elementsFrom the alignment point of view, the question of having ILC installed along the earth curvature or along a laser straight line is not a relevant. Nevertheless, the use of hydrostatic leveling system (if needed) will be easier in the earth curvature solution adopted for the linac tunnels.The reference points will provide the absolute position of the elements. As said above, additional links with the surface geodetic network are not excluded, and studies have to be carried out on this subject. The relative position of the elements will be obtained thanks to a

Rev. March 16, 2006

Conventional Facilities & Siting 27

smoothing process performed from batches of elements to batches of elements. This process can be based on the existing stretched wire techniques.As the alignment processes needs space kept free in the tunnel during the life of the project, studies will integrate the alignment constraints since the beginning of the project (shape of the slab of the tunnel, size of the tunnel, geometrical links to the surface…etc).

Final focus areasParticular attention will be paid to the alignment constraints in the interaction regions. The tunnels or the underground halls and the layout of the services will allow a direct and permanent view along the beam line through or aside the detector and as closely as possible to the most critical elements.The mechanical stability of the components is very critical. In particular, the ventilation will maintain the elements at constant temperature to minimize as far as possible the deformations of supports and mechanical elements. No variation will occur due to season’s effects.Slow drifts due to the variation of the gravity field-because of the ocean and atmospheric loads, variation of the ground water level, and earth tides- will be taken into consideration.Some substantial R&D needs to be carried out in order to cope with the alignment of the components in this area.

9.6 Summary

Global Group #4, Conventional Facilities and Siting (CFS) is organized to continue the efforts of the development of a Reference Design Report as an integrated group. The BCD requirement of each of the other GDE groups will be translated into a consistent set of conventional facilities criteria. The criteria and resulting design solutions will be reviewed with the “single points of contact” that have been established with the other GDE groups. Design solutions will be common to all regions where possible and site specific when needed. The understood charge is to produce an economical and functional design that is consistent with the BCD and produce a credible estimate and schedule.

Rev. March 16, 2006

Operations and Availability 1

10. Operations and AvailabilityAuthors: E.Elsen, T.Himel, N.Terunuma

Last revised: February 3, 2006

IntroductionThe purpose of the operations and availability group is to look globally at the ILC and make sure it is designed so it is producing luminosity a large fraction of the time and there are enough diagnostics and controls that it can be commissioned, operated and tuned efficiently. The most work so far has been done on availability. The work which has been done on the machine protection system (MPS) is described in the controls section of this document.

The main result of the work done for this section is to set extra requirements for various parts of the accelerator. It is important to make sure that these requirements are met by the systems described in the other sections of this document.

Availability considerations provide much of the impetus to have 2 tunnels so that the support equipment can be repaired without the need for entering the accelerator tunnel. They also show the need for a positron keep-alive source.

Summary of Key Decisions for the Baseline Design The linac will have two parallel tunnels so that the support equipment may be

accessed without entering the accelerator enclosure

There will be a keep-alive positron source that can provide positrons when the electron DR or linac is down.

Each region of the ILC will have sufficient beam stoppers and shielding so people can be in that region while beam is in another region. The PPS system will be designed to allow this. By region we mean injector, DR (except for the dog-bone version which is in the linac region), compressor, main linac, and BDS.

A large effort will be needed to make individual components reliable and/or redundant. More details are given below.

Availability

MethodologyThe ILC will be the most complex accelerator ever built. If careful attention isn’t paid to making the individual components extremely reliable and designing in redundant systems, it will be down all the time.

To quantify the problem a simulation was written which has the following features.

It has a count of the number of each type of component in each part of the accelerator. Examples of components are water cooled magnets, superconducting magnets, power supplies, power supply controllers, vacuum pumps, vacuum pump power supplies, BPMs, klystrons, modulators and pulse transformers.

Each component has a Mean Time Between Failure (MTBF) and Mean Time To Repair (MTTR). The starting values for these are taken from experience at existing accelerators where that was available.

Rev. Feb.4, 2006

Operations and Availability 2

When a component breaks, the performance of the accelerator is degraded in a component specific manner. Examples:

o A failed klystron in the linac degrades the energy headroom. Only if that headroom goes to zero is the accelerator considered to be down.

o A failed quadrupole in the main linac is assumed to degrade the luminosity by 1% after 2 hours at zero luminosity is spent tuning around it.

o A failed quadrupole in the DR is assumed to completely break the ILC.

Each component has one of 3 repair methods

o Hot repairable: It can be repaired without taking the accelerator down further. The canonical example of this is a klystron or modulator in the main linac.

o No access needed: The accelerator has to be down for the repair to be done, but no access is needed to the accelerator tunnel. An example of this the replacement of a failed AC breaker in the equipment tunnel or a support building.

o Access required: In this case one must go into the accelerator tunnel. A 1 hour cool-down and 1 hour startup period are added to the MTTR to get the total repair time for these components.

Hot repairable items get repairs started immediately after they break. The others only get repaired when the accelerator is down. When this happens, many items may get repaired in parallel.

The time to recover well tuned beams in a section of the accelerator is proportional to the time that section has been without beam.

There are many more features. More details can be found in Chapter 4 of the US Technology Options Study at www.slac.stanford.edu/xorg/accelops/ and in a talk given at Snowmass 2005 at http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/GG3/tom_himel20050812014007.ppt.

Given a set of inputs, the simulation computes the total downtime and how much of that time was due to each component. Repeated runs of the simulation were used to simulate different possible configurations of the ILC (e.g. 1 and 2 tunnels) and to adjust the MTBFs to achieve a desired uptime of the accelerator. Typically, the components which contributed the most to the downtime were adjusted to have better MTBFs as very little would be gained by improving components that did not contribute significantly to the downtime.

The goal set was to have the calculated downtime be only 15%. It is assumed that due to things that were left out of the simulation, or hardware design errors that there will be an extra 10% downtime for a total of 25%.

Specifications needed to achieve adequate availabilityThis section describes how the required calculated availability of 85% was achieved. As the design progresses, these may change, but this is a reasonable starting point.

There is a 3% energy overhead in each main linac that can be switched on without delay.

The roughly 5 GeV accelerators have a 5% energy overhead and the smaller linacs (e.g. the e+ acceleration to 250 MeV and crab cavities) have a 5% overhead plus an extra cavity.

Rev. Feb.4, 2006

Operations and Availability 3

There are hot spare klystrons/modulators with waveguide switches in all low energy linac regions (5 GeV booster, bunch compressor, crab cavities…). This allows a bad klystron or modulator to be immediately compensated for by switching in the hot spare. Note that one hot spare for each low energy linac will suffice. These are needed because the fractional energy change due to a klystron failure is very high in these regions so it is impractical to replace the energy lost due to a failed klystron with energy gain at a different longitudinal location. Without the hot spare switchable klystrons, the availability dropped 1.3%.

When klystrons are not in the accelerator tunnel, they can be hot swapped. This will require some type of valve in the output waveguide.

Most electronics modules not in the accelerator tunnel can be hot swapped. This means that if the module is in a crate it is not necessary to power down the crate to swap the module.

There are tune up dumps and shielding between each region of accelerator so that one region can be run while people are in another region. By region we mean injector, DR (except for the dog-bone version which is in the linac region), compressor, main linac, and BDS.

Problems with the overall site power will cause 0.5% downtime. Note that present day experience is that a quarter second power dip can bring an accelerator down for 8 to 24 hours. If this is the case for the ILC, then it is the 8-24 hours that counts towards the 0.5%, not the quarter second. Hence this is really a very stringent requirement on the reliability of the incoming power and onsite power distribution system. More engineering will be needed to see if this requirement is best met by keeping the rate of power dips very low or making all the equipment capable of running through short power dips. Of course, long power outages are even more deadly and need to be avoided. To set the scale, an IEEE survey summary at http://ieeexplore.ieee.org/iel4/6112/16344/00757619.pdf?tp=&arnumber=757619&isnumber=16344 says that with a single primary feed there is an average of 2.0 power dips per year while with dual feeds there are 0.16 such dips. If each dip causes 12 hours of downtime, these correspond to 0.27% (single feed) and 0.02% (dual feed). Even though the same surveys indicate that the single feed power systems are up 99.97% of the time, they may not be good enough for the ILC because they only include as downtime, that time for which there is no power, not the considerable recovery time required by the various ILC subsystems.6

The total downtime due to any cryo plant being down is 1%. If there are 6 cryo plants then each must be up 99.85% including outages due to their incoming utilities (electricity, house-air, cooling water, ventilation). (Hence all plants are up exactly 99% of the time.) This is 3-6 times better than the Fermilab and LEP cryo plants. Downtime accounted to the cryo system is all the time for which there cannot be full power beam due to the outage. If things warm up and have to be cooled back down, that counts as downtime to the cryo plants. Present accelerator experience is that around half of the cryo plant downtime is due to the incoming utilities. Hence achieving this goal will require both more reliable utilities and more reliable cryo

6 The IEEE survey summary indicates that the most sensitive of the on-site subsystems is motors, or more accurately, motor controllers. A standard addition of UPS (Uninterruptible Power System) power supplies to all motor control drives and to all process controllers would substantially improve overall power dip tolerance for the ILC. Engineering to identify other critically sensitive subsystems is necessary.

Rev. Feb.4, 2006

Operations and Availability 4

plants. Note that if there were a way to have N+1/N redundancy of the cryo plants then attaining an acceptable availability would be pretty easy.

Due to an oversight, no downtime has been budgeted for ventilation or global water cooling systems such as cooling towers. (Local water pumps are included under the heading of water systems.) In the RDR we will have to allot some downtime to these systems and make something else a bit more reliable.

The starting MTBF used for magnet power supplies of 200,000 hours is 4 times better than SLAC/Fermilab experience. This probably requires redundant regulators. The final MTBF given in Table 1 below is larger still.

The power coupler interlock electronics and sensors have MTBFs of 1 million hours due to redundancy. That is the individual channels may have worse MTBFs, but there is enough redundancy built in that the effective MTBF is at least 1 million.

Cavity tuner motors are a potential reliability concern. Its seriousness is unclear at this time. In the simulation they have a starting MTBF of 1 million hours, 2 times better than SLAC warm experience and much better than TTF cold experience. Even with this longer than realistic lifetime, 130 would break each year reducing the maximum CM energy by about 4 GeV. Repairs to regain this energy would require warming essentially the whole linac. The simulation assumes the motors must be used periodically to keep the cavities tuned to the correct frequency. This is the case at JLAB, but not TTF. Their frequency of use depends on Qext and the amount the helium pressure varies. If the ILC design results in the frequency being stable enough so tuning is unnecessary except immediately after cool down, there may be no problem. Otherwise, there will be and an improved tuner motor design may be necessary. Note that even with complete stability, the tuners must be used when the klystron is turned on or off.

There is a spare e+ target beam-line with 8 hour switch-over.

Failed superconducting linac quads can be tuned around in 2 hours. Careful attention should be paid to this particularly in the low energy parts of the linacs as the repair time for a SC quad is too long to be tolerated during the run.

Most failed correctors can be tuned around in 0.5 hours

There will be no vacuum leaks that force a cryomodule to be warmed up for a repair before the run can continue. Vacuum leaks that can be ameliorated with the installation of a turbo-pump are allowed.

Interlock systems are a particular concern. Even if they don’t actually break, giving many false trips can effectively inhibit stable operation. It is thus essential that all interlocks based on an analog value (e.g. water flow, temperature, ground fault current) must have that value read by the control system so that impending trips can be seen and prevented. The same number which is read should be digitally compared to a remotely settable limit to cause the actual trip. The cause of all trips must be recorded by the control system.

Table 1 gives the required MTBFs of the components which cause the most downtime. MTBFs are given for 3 cases to allow comparison of 1 vs. 2 tunnels (which will be described in a later section). The 2 tunnel case is shown in column 2. Note that many MTBFs need to be improved by a large factor over the present values. For example water cooled magnets need to improve by a factor of 20 to 20,000,000 hours = 2,200 years. This is not as impossible as it seems. The MTBF accounts for failures that occur during the run.

Rev. Feb.4, 2006

Operations and Availability 5

Preventive maintenance (such as flushing water passages) or even periodic replacement is allowed. Also, one set of magnets in HERA nearly achieved this MTBF. Small improvements in MTBFs can probably be accomplished with minor design improvements at little cost. Larger improvements (> 3) are likely to require careful redundant design and prototyping. Very large improvements (>10) probably carry significant technical risk and will need to be tested in large enough quantities to check the actual MTBF. Note that while we chose to improve the MTBFs in the simulation, applying some of the improvement to the MTTRs instead should result in similar improvements to the availability.

Note that Table 1 only shows the MTBFs for those components that the simulation showed contributed significantly to the downtime. It is important that all components be reliable. The ILC cannot afford to be cheap on items not listed or on seemingly simple areas like transport lines. Careful attention to availability will be essential everywhere. The pie chart in Figure 1 summarizes how much of the downtime comes from the various regions of the ILC. The chart in Figure 2 shows which systems are causing the downtime.

Rev. Feb.4, 2006

Operations and Availability 6

Table 1. This table shows the MTBFs that were used to obtain the desired 15% downtime for 3 different cases. Note that the desired MTBF is the product of the nominal MTBF and the improvement factor. The third column shows the percentage downtime caused by the devices for the situation given in the second column. These can be used to estimate the effect of not meeting one of the MTBF goals.

Rev. Feb.4, 2006

Device

Improvement factor

A that gives 17% downtime for 2

tunnel undulator e+ source

Downtime (%) due to these devices for 2 tunnel undulator

e+ source with strong keep_alive

Improvement

factor B for 1 tunnel undulator

e+ source, 6% energy overhead

Improvement

factor C for 1 tunnel undulator

e+ source, 3% energy overhead

Nominal MTBF (hours)

magnets - water cooled 20 0.4 20 20 1,000,000power supply controllers 10 0.6 50 50 100,000flow switches 10 0.5 10 10 250,000water instrumention near pump 10 0.2 10 30 30,000power supplies 5 0.2 5 5 200,000kicker pulser 5 0.3 5 5 100,000coupler interlock sensors 5 0.2 5 5 1,000,000collimators and beam stoppers 5 0.3 5 5 100,000all electronics modules 3 1.0 10 10 100,000AC breakers < 500 kW 0.8 10 10 360,000vacuum valve controllers 1.1 5 5 190,000regional MPS system 1.1 5 5 5,000power supply - corrector 0.9 3 3 400,000vacuum valves 0.8 3 3 1,000,000water pumps 0.4 3 3 120,000modulator 0.4 3 50,000klystron - linac 0.8 5 40,000coupler interlock electronics 0.4 5 1,000,000vacuum pumps 0.9 10,000,000controls backbone 0.8 300,000additional linac energy overhead 3% 3%

Operations and Availability 7

Total

IP0% e- DR

12%

e+ DR10%

e- BDS3%

e+ BDS4%

e- linac15%

e+ linac8%

e- source6%

e+ source14%

Site power6%

Cryo plants11%

e- compressor5%

e+ compressor4%

Global controls2%

Figure 1. This shows how the total downtime of 17% is distributed among the various regions of the ILC. The simulation had 2 tunnels with an undulator e+ source with a strong keep-alive source and the “A” MTBF improvement factors shown in column 2 of Table 1.

Rev. Feb. 4, 2005

Operations and Availability 8

Total

Cryo8%

Vacuum16%

Magnets5%

AC power7%

controls17% Diagnostic

0%

RF structure7%

Water system11%

PS + controllers17%

RF power sources12%

Figure 2. This shows how the total downtime is distributed among the various systems of the ILC. The simulation had 2 tunnels with an undulator e+ source with a strong keep-alive source and the “A” MTBF improvement factors shown in column 2 of Table 1. Note that the global system (site power, cryo plants, site-wide controls) are not shown in this chart.

Need for positron keep-alive sourceA priori an undulator source for positrons will have a lower availability than a conventional positron source. The reason is that the electron arm has to be up and fully tuned before positrons can be produced. Commissioning and machine development in the positron arm are thus affected. Simulations have been run to compare a conventional source and an undulator source. For a two tunnel configuration these yield availabilities of 80% and 69% respectively. This loss in availability for the undulator arises both from the strictly sequential retuning of the two accelerators after a downtime and the lack of machine development options for the positron arm while the electrons are not available.

This loss in operational efficiency can be largely mitigated by introducing a standby keep-alive positron source based on conventional technology. If such a source can be activated within two hours, the positron damping ring and linac can be tuned and machine development for positrons can be envisaged. The availability of 78% for such a system almost reaches the availability of the conventional source.

Rev. Feb. 4, 2005

Operations and Availability 9

The intensity of this keep-alive source must be sufficiently good so that BPMs can be used for all their normal purposes including beam based alignment and steering. If the intensity is low enough that BPM gains, offsets, or resolutions change significantly from their values at full beam intensity then there is very little improvement in the availability. The dynamic range of the BPMs can be optimized to some extent and the practical limits are not well understood at this time. A full intensity source with every second bunch filled could clearly fulfill the task and would be suitable to study essentially all intensity induced effects in the accelerator (e-clouds in DR etc.). A 20% intensity source will also serve the keep-alive function in many respects. A single bunch, 1% intensity at 5Hz repetition is likely too low to fulfill the requirement. The detailed requirements for such a keep-alive source still need to be developed. They depend almost completely on the lowest intensity at which BPMs work as well as described above. It should be noted that the WG3a (Sources) have sketched an attractive multiple source pre-accelerator scheme that would serve the high intensity requirements. It can be found at http://www-project.slac.stanford.edu/ilc/acceldev/eplus/Snowmass_files/Undulator_based/3_6_1_1_Keep_alive_overview.doc. The TESLA TDR implemented a lower intensity source which incorporates common use of the positron capture and acceleration part.

Need for 2 tunnels. Several tunnel configurations have been discussed by the parameters group. They range from a tunnel near the surface with a klystron gallery above it to a single deep tunnel to a pair of deep tunnels. A second tunnel or klystron gallery enables service access to the RF, modulators, power supplies, and electronics while the linac continues to operate in the main tunnel. With RF sources in the tunnel it is not clear which access restrictions apply for safety reasons. Such regulations may also depend on the country hosting the facility.

Four configurations have thus been compared in the simulation assuming an undulator positron source with high intensity keep-alive beam and the MTBFs given in column 2 of Table 1. The results are shown in Table 2.

Tunnel configuration Simulated % time integrating luminosity under normal running conditions

Simulated % time integrating luminosity when commissioning

a single tunnel without robotic repair

64% 25%

a single tunnel with robotic repair

73% Not simulated

two tunnels where the support tunnel is always accessible

78% 46%

Rev. Feb. 4, 2005

Operations and Availability 10

two tunnels where the support tunnel is only accessible when the RF is turned off

72% Not simulated

Table 2 The percentage of time spent integrating luminosity for various tunnel configurations. The third column gives results simulating an early commissioning period. To simulate that, the MTBFs are halved and the MD and tuning times are doubled.

Note the significant decrease in integrated luminosity when going from two tunnels to one. The third column of the table shows this is still more extreme for the early commissioning period when the accelerator is not operating as well.

One doesn’t have to settle for the lower uptime caused by a single tunnel. Instead, one can choose to improve the availability of the individual components to recover the lost uptime. Columns 4 and 5 of Table 1 show a couple of examples of how this could be done. For column 4, in addition to increasing some MTBFs, the energy overhead of the main linacs is increased by 3% to reduce the downtime due to insufficient energy. For column 5, MTBFs are improved still more and the energy overhead is not increased. For both cases, the MTBF increases are quite large. There is a significant risk that they won’t be achieved and uptime and integrated luminosity will suffer.

In addition to the availability problems, there are several other disadvantages to using a single tunnel.

1. The electronics will be in the accelerator tunnel and hence exposed to radiation. One must then design to avoid both problems from radiation damage and from Single Event Upsets.

2. Subtle electronics problems that need beam to diagnose will be very difficult to debug.

3. Installation, upgrading, and repairing of electronics cannot proceed when there is beam.

4. If the dog-bone DR is chosen, the pulse transformers of the linac RF system produce a magnetic field that disrupts the DR beam.

The only advantage of a single tunnel is that it is significantly cheaper.

Extra features needed for Commissioning

Phased commissioning

For the commissioning of the ILC the largest amount of time is expected to be spent in the damping rings. The low emittance out of the damping ring is the key to observing any further distorting effects downstream. It is thus evident that the most favorable construction scenario is one in which the damping rings are available early so that the commissioning can proceed

Rev. Feb. 4, 2005

Operations and Availability 11

in phases.

More generally, careful thought should be given to the construction schedule to allow as much commissioning time as possible while downstream construction proceeds.

Electron source and reversible e+ DR

Low emittance electron beams will be necessary to explore and understand the limitations of the positron damping rings. The damping ring polarity should hence be reversible and a low emittance electron source made available for injection into the positron damping ring. Even with this auxiliary electron source, the keep-alive positron source will be useful and should be available during the commissioning period.

Globally synchronized data acquisition for fault analysis

For commissioning (and normal running) the control system should be equipped to record globally synchronized data at the bunch level. Such a system should be available from the start so that correlations between effects can be properly traced offline with reduced need for specific experiments.

Bypass line to skip e- DR for early e- linac commissioning

It would be nice to be able to inject directly into the main linac for commissioning before the DR is complete or when it is broken. A transport line to bypass the damping rings will be needed to allow this. This is not a make or break item, so it should only be done if the design allows for a reasonably short, cheap bypass line.

Automated surveying system

Alignment questions will recur during commissioning. It will be very helpful to have an automated surveying system implemented that allows for alignment of accelerator components.

MPS

The Machine Protection System was discussed jointly between the operations and availability global group and the diagnostics and controls global group. Its description is in the controls section, not this one.

Operability

Not much work has been done on operability issues other than the availability simulations.

There are other aspects that do need to be considered.

The possible need for extra diagnostics (or better resolution) to track down where an obscure problem (vibration, wakefield, power supply ripple…) occurs

Needed control system features.

Rev. Feb. 4, 2005

Operations and Availability 12

Existence of enough control points and diagnostics to properly tune the beam

Specification of feedbacks and automated tuning procedures

Further work needed Tune MTBF requirements as engineering is done to try to minimize the cost while

maximizing the uptime.

Improve parts counts and add detail to the availability simulation as engineering is done (e.g. water pumps and AC distribution are very generic).

Calculate tradeoffs for other groups as questions arise (e.g. 16 channel chassis for corrector supplies vs. individual supplies)

Benchmark the simulation code with HERA.

Design, prototype, and test some parts that need large availability improvements.

Develop other requirements to produce an operable accelerator.

Rev. Feb. 4, 2005

Instrumentation and Controls 1

11. Instrumentation and Controls

Introduction The Instrumentation and Controls BCD document consists of thirteen sections, in addition to this introduction, as listed here: 11.1 Controls Standard Architecture 11.2 Timing System 11.3 Diagnostic Interlock Layer 11.4 Global Network 11.5 Machine Protection 11.6 Low level RF 11.7 Feedback 11.8 Integration with Instrumentation 11.9 Machine Detector Interface 11.10 Instrumentation – Beam position monitors 11.11 Instrumentation – Beam profile monitors (transverse) 11.12 Instrumentation – Longitudinal 11.13 Instrumentation – other (intensity, loss, ring)

It is anticipated additional sections will be needed as requirements are more carefully defined and understood.

Author list:

11.1 Claude Saunders, Andrew Johnson (ANL), Ray Larsen (SLAC), Matthias Clausen (DESY)

11.2 Frank Lenkszus (ANL)

11.3 Ray Larsen (SLAC)

11.4 Ferdinand Willeke (DESY), Margaret Votava (FNAL)

11.5 Marc Ross (SLAC)

11.6 Brian Chase (FNAL), Stefan Simrock (DESY)

11.7 John Carwardine (ANL)

11.8 Manfred Wendt (FNAL), John Carwardine (ANL)

11.9 TBA

11.10 Steve Smith (SLAC), Hans Braun (DESY)

11.11 Grahame Blair (RHUL) and Marc Ross (SLAC)

11.12 Marc Ross (SLAC)

11.13 Junji Urukawa (KEK)

11.14 Controls (to be added)

v. Dec.12, 2005

Instrumentation and Controls 2

11 Instrumentation

The ILC Instrumentation system functions both to provide diagnostic information to be used to correct substandard operation and as an integral part of the machine control system, providing input to the machine protection system and beam-based feedbacks. There are four types of basic monitors: position (BPM), intensity (toroid), profile and loss (BLM). These are supplemented by a system of special monitors to be 1) jointly used by the accelerator and detector, 2) to monitor other aspects of the beam – such as longitudinal profiles and correlations, beam timing, damping ring parameters, beam halo, and 3) feedback. The beam –based instrumentation system is further supplemented by hardware monitors: temperatures, field probes, radiation monitors and etc. The instrumentation for the ILC is challenging and much of it, although demonstrated in small test installations, has never been implemented on a large scale. From the point of view of instrumentation, the ILC is divided into two pieces, the ‘damped beam’ section (damping rings beam dumps), and the injector system (upstream of the damping rings, including injection into the rings). Typical beam sizes and required position monitor resolution in the damped beam systems are around 1 micron. In some cases, these can be much smaller (~0.1 microns). RD is needed to provide confidence in a given system design, especially for the BPM and profile monitor systems.

The most critical (and most expensive) instrumentation system is the BPM system. Experience at LEP, Tevatron, PEPII, SLC and many synchrotron light sources has shown the importance of having a well engineered, proven BPM system. The first instrumentation section of this chapter deals with BPM requirements and how these will be met, in large part by precision RF cavity BPM’s. There are 2 parts to the section, one for the injector and damping ring and the other for the downstream systems, linac and beam delivery. The second section describes the second critical system, the damped beam profile monitor system. It is this system that validates the performance of the low emittance transport. For the most part, these monitors will be based on ‘laser-wires’. A laser-wire consists of a 90 degree Compton scattering chamber where a finely focused, very high power pulsed laser is used to sample the particle beam density. Although laser-wires have been built and successfully tested in all three ILC regions, these systems are still very much in development and require constant handling by experts. It is useful to think of the laserwire system as providing an estimate of the luminosity, if the beams were brought into collision at that point. In that way, laserwires can be used to segment the low emittance transport. In sharp contrast to BPM’s, laserwires need their own section of beamline to function optimally and this has added cost. The beamline length needed depends on the surrounding components (e.g. collimation), typical beam sizes in the area and the expected performance of the laserwires. The fourth instrumentation section describes longitudinal diagnostics. The ILC longitudinal diagnostics will be used to measure the bunch length and the x z, y z and E z correlations. These devices are used to test the damping ring beam dynamics, the bunch compressor phase space rotation, the phase space distortion in the main linac, the wakefield kicks in the collimation system and the effects of poor optical matching and non-linear fields. Because the longitudinal phase

v. Dec.12, 2005

Instrumentation and Controls 3

space distribution is not expected to be Gaussian and small features in the distribution are important, these devices must have resolving power well beyond the characteristic bunch length scale. It is expected that a relatively small number will be needed, but, as with the laserwires, these

v. Dec.12, 2005

Instrumentation and Controls 4

devices need dedicated beam line space and hence have cost implications. Finally, the last instrumentation section deals with special monitors.

Table 1 summarizes ILC instrumentation requirements.

Monitors for intensity and transverse beam position

ILC

compone

nt

Required

resolution

/

precision

Requi

red

riseti

me

Technol

ogy

units

need

ed

total

both

sides

Cost

estimate/

unit

Information

from Remarks

R &D

requirem

ents

Injector Sigma/5 6MHz Stripline 600

4K

excludin

g

vacuum

hrdwre

Self

Reliabilit

y;

redundan

cy

Damping

ring

1um

narrow

band Roll

20 mrad

precision

100um.

Stability

1um

Slow; Button 600 4K exc.

Hrd.

Snowmass

WG3b

Stability,

roll

under

study

(CCLRC

)

ATF 1

pm-rad

Damping

ring

Special

for ffbk

fraction

sigma

also

injection

Bunch

spacin

g

Button 20 8KSnowmass

WG3b Ffbk RD

Ffbk

integrati

on

Damping

ring 1 µm / ?

L_w

/2

For

wiggler

sections;

vacuum

chamber

RD

Similar

to the

rest of

DR

v. Dec.12, 2005

Instrumentation and Controls 5

Linac

(BPM)

recomme

nded

sig/3, a

few at

sig/10 for

FFBK

Nom I.

scales

with I for

lower.

6MHz

separa

te. 10

%

increa

se in

noise

from

prev

bunch

Cavity ?

re-

entrant

?

800

10K incl

cavity –

more if

cleaning

is

included

M. Wendt

GG2 talk

WG4/1

common

session

Scale

factor;

integral

linearity,

from DS

0.5% for

absolute

gain over

200um

(needs

verificati

on)

Question

43

Calibrati

on

process,

analysis

from

nBPM

Linac

(inten) 1%

Whole

train

Ferrite

loaded

gap

4 5K self

Precision

intensity

– what is

needed

for 1

bunch

Linac

dark I

50nA/

1ms pulse 1ms

Resonan

t 010

mode/

? Olivier

Test dark

I meas at

TTF

Linac/

DR1e-4? 780ps 2

Parasitic

bunch

Single

photon

counting

?

Beam

delivery-

spectrom

eter

100 nm 36

Tesla

TDR/Snow

mass WG4

spectrom

eter 1

plane

Stability

200 nm

Beam

delivery-

IP

feedback

1 µm /

100 µm

Stripline

? 4 Tesla TDR

IP

feedback

Backgro

und

influence

(ESA)

Beam

delivery

8nm to

100um

Same

as

Cavity

for

500 10K

including

Woodley’s

table

‘normal’

Virt IP’s

3 or 4

types.

v. Dec.12, 2005

Instrumentation and Controls 6

– all else

s/10.

beam size

varies

from

85nm to

1.2mm.

linac hardest cavity counted?

Some

hard –

ATF2 IP

nBPM

Beam phase monitors

ILC

compone

nt

Requir

ed

resoluti

on /

precisi

on

Requir

ed

risetim

e

Technolo

gy

units

needed

Cost

estimate/u

nit

Informati

on from

Remar

ks

R &D

requireme

nts

Injector

– gun

system

0.1 deg Single

bunch Cavity 2 20K Self

Use

Haims

on

Test

required –

not used

for SHB

systems

Damping

ring 0.1 deg

Single

bunch Cavity 2 10K Self

From

main

RF

Bunch

Compres

sor

0.01de

g

Single

bunch Cavity 6 30K

WG1 BC

spec

Tighte

st

phase

monito

r req.

Linac 0.1Single

bunch

Part of

LL-RF ?

May be

integrated

in LL-RF,

no add’l.

cost

Self TTF, SNS

Beam

delivery

Collisi

on

overlap

Single

bunch

2 Integrat

ed with

crab

2M$

v. Dec.12, 2005

Instrumentation and Controls 7

–

s_z/10?

v. Dec.12, 2005

Instrumentation and Controls 8

Monitors for transverse profiles

ILC

component

Required

resolution

/

precision

Required

risetime Technology

units

needed

Cost

estimate/unit

Information

from Remarks

R &D

requirements

Injector Sigma/5 Single

bunch

Wire

scanner 30 30K Self

Damping

ring

10%

emittance

Multi-bunch

ok

XSR,

laserwire, ? 2 of

each/ring 250K

WG3b

Snowmass

ATF

performance

not quite

XSR RD

needed

Bunch

Compressor 10%e

Measurement

of single

bunch w/o

train

Laserwire

3sets/side

for 2

stage BC

250K/set WG1

Snowmass

Integration

with lattice

needed for

coupling

precision

ATF2 tests

Linac 10%e Same

Laserwire/

short warm 3 sets/

linac 250K/set

Question

29

Cryo warm

section

needs study

Beam

delivery 10%e Same

Laser wire

2 sets/

side 250K/set

WG4

Snowmass

Does not

include

secondary

waist

monitors,

extracted

beam

monitoring

IP area,

secondary

waists,

extraction

line

Beam

Delivery –

collimation

system

monitors

v. Dec.12, 2005

Instrumentation and Controls 9

Monitors for longitudinal profiles

ILC

compon

ent

Requir

ed

resolut

ion /

precisi

on

Required

risetime Technology

units

neede

d

Cost

estimate/

unit

Informa

tion

from

Remark

s

R &D

requirem

ents

Injector;

gun,

SHB

system,

e+

collectio

n,

booster

linacs

dE/s_z

dE

~0.01

% /

s_z ~

100um

Single

measure

ment

possible

w/o train

Wire

scanners and

LOLA Gun,

linac,

DR

entra

nce

30K

/wire &

300K/L

OLA

Self

Could be

tested at

SLAC/K

EK

Dampin

g ring

s_z

S_z/10

Single

bunch

w/o train

Streak

camera/defle

ction cavity

1per 500K

Dampin

g ring

dE

0.01%

XSR/visible

SR 2 350K

Bunch

Compre

ssor

dE ~

0.01%

/ s_z ~

30um

Laserwire/

wire scanner

& LOLA 2

30K

/wire &

300K/L

OLA

Linac –

dE at

end

dE

0.01%

2

Bunch

compre

ssor

monitor

s used

at input

Beam

delivery

-

correlati

2 Crab

system

see

above

v. Dec.12, 2005

Instrumentation and Controls 10

ons listing

v. Dec.12, 2005

Instrumentation and Controls 11

Special Monitors

ILC component

Type Require-ments

Technology

units needed

Cost estimate/unit

Information from

Remarks

R &D requirements

Injector Beam loss

1% remote handling limit – 1W/m- linearity for MPS sequence

Ion chamber

1/10 m + PLIC

0.5K

100x less sensitive than SNS

cost

Damping ring

Beam loss Same

Damping ring - wiggler

Beam loss Tighter 10x – neutrons?

Bunch Compressor

Beam loss Same as inj.

Linac Beam loss Same as inj.

Beam delivery

Beam loss Same as inj.

Beam delivery - Collimation

Beam loss Calorimetry?

Beam delivery

Luminosity

Beam delivery

Polarisation

v. Dec.12, 2005

Instrumentation and Controls 12

11.1 Control System Architecture

1. Overview The system topology is assumed to be two parallel tunnels with a central control room near the interaction point.

The baseline configuration (BC) shall use as a reference design existing packaging standards such as VME and VXI, and be similar to the model envisaged for the NLC and Tesla as well as modern machines such as LHC. The software standard will be a 3-tier architecture with established frameworks at each tier. This approach would minimize development effort.

However an alternative configuration (AC) is under consideration to develop a new architecture and packaging standard for the ILC, driven by the need for High Availability (HA) design of both hardware and software. This requires R&D evaluation of the technical and operational benefits of a significant HA investment to enhance the capabilities of both hardware and the 3-tier software at every level. HA systems use Intelligent Platform Management diagnostics and control which can also be extended to other electronics systems, including power electronics.

The BC can draw cost models from the NLC and TESLA models as well as newer machines. The AC requires additional R&D to evaluate and converge on a new incremental cost model with enhanced HA architectural features.

2. Baseline Configuration

a. Description The baseline design envisions a dual star network model for controls emanating from a central modular computer cluster (Figure 1).

Figure 1. Control Room Cluster & Dual Serial Networks

Dual star data links provide branch control to all sector nodes of the various machines

v. Dec.12, 2005

Instrumentation and Controls 13

(Figure 2).

Figure 2. Dual Links, Sector Node Processors, Front End Modules

The baseline software architecture utilizes a standard 3-tier approach: client tier, services tier, and real-time tier. This approach provides separation of concerns, re-use, load management, change management, and many other benefits. A significant portion of the logic that traditionally used to reside in the client tier is now provided as a service for use by many clients. Services provide a means to coordinate the activities of many applications, and also serve to integrate real-time and relational database data into a seamless API.

v. Dec.12, 2005

Instrumentation and Controls 14

Figure 3. Software Architecture

The performance requirements within and between the three tiers varies considerably, and therefore different protocols are required at different levels. See Figure 3 for a summary of the tiers, protocols, and relative performance timescales. The RTP (Real Time Protocol) provides a high-performance, narrow interface to channel (process variable) oriented data. The DOP (Distributed Object Protocol) provides a slower, wide interface to the services tier. The integral use of a relational database for maintaining an engineering model, physics model, controls model, and operational data should be noted.

b. Supporting Documentation [1] Johnson, A., Clausen M., “High Availability Architecture for ILC”, October 2005. http://docdb.fnal.gov/ILC/DocDB/0001/000106/001/10.1.R.1%20HA%20Software%20%20Architecture%20for%20ILC.doc [2] Saunders, C., Kraimer M., “ILC Controls Software Architecture”, October 2005. http://docdb.fnal.gov/ILC/DocDB/0001/000105/001/10.1.R.2%20Controls%20Software%20Architecture.doc [3] Larsen, R. “Controls and Instrumentation Standard Architecture Options”, July 2005. http://docdb.fnal.gov/ILC/DocDB/0001/000104/001/10.1.R.3%20Controls%20HA%20Standard%20Architecture.doc

c. Cost Estimation The NLC and TESLA cost estimates for controls can serve as a starting cost model. Costs or cost estimates of recent control systems of SNS, LHC and other newer accelerators will serve as useful reference points. An inter-regional collaboration cost team will be established to agree on the technical cost model, while actual estimates will follow a set of GDE cost rules yet to be defined. The goal is to define the technical model sufficiently to perform a preliminary bottom-up cost estimate, and compare this result parametrically with other machines. Cost development can progress alongside development of the technical model.

3. Alternative HA Configuration

a. Description Software and hardware engineering have to work hand in hand to achieve the required availability for the ILC. Depending on the time span of the breakdown of an individual component until the machine will go down we have to distinguish several different areas, which result in different approaches for mean time to automatically replace (MTTAR) (not repair) a component. These areas reflect the timescales of machine operation (from smallest to largest): bunch-by-bunch, macro pulse, process controls, static controls. The high availability solutions for each of these areas is potentially very different. For controls and instrumentation, the Advanced Computing and Telecommunications Architecture (ATCA) is a strong candidate for a base platform. ATCA is an open standard platform designed by a strong industry consortium called PICMG. The features include dual redundant communications processors and links, intelligent crate (shelf) diagnostics and management, and hot-swap-capable modules and sub-modules. See Figure 4 for an example of ATCA applied to an integrated beam instrumentation package.

v. Dec.12, 2005

Instrumentation and Controls 15

b. Benefits Current accelerators (other than light sources) have up-times ranging from 40-70%. The ILC will be ten times the size of any comparable existing linear machine so availability becomes a dominant issue. Recent machine availability simulations presented at Snowmass 2005 show the difficulty of reaching 80% even assuming much improved reliability of all subsystems.

Figure 4. Beam Instrumentation Packaging Concept

Because of the recognized difficulty, it is essential that the ILC controls, instrumentation and electronics systems, including power electronics, be designed with High Availability (HA) architectures. In addition, HA controls design with orders of magnitude increased reliability will make errant beam problems due to controls failures far less frequent and easier to diagnose, and may ease the design complexity of machine protection systems. HA architectures incorporate intelligent management features to negotiate around faults in the system. Fault management typically involves five stages:

1. Detection - the fault is found 2. Diagnosis - the cause of the fault is determined 3. Isolation - the rest of the system is protected from the fault 4. Recovery - the system is adjusted or restarted so it functions properly 5. Repair - faulty system components are replaced.

Operating, middleware and applications layers all are impacted by HA design. In the ultimate, critical applications can be designed with fault-avoidance auto-failover features through a combination of software design, hardware diagnostics and hot-swap serviceability. In principle, controls and instrumentation systems can be made very robust against single-point failures in communications or logic operations that typically plague large complex experimental accelerators and detectors.

c. Required Research and Development 1. Evaluate existing HA architecture standards for applicability to all ILC electronics

systems, subsystems, hardware, and software. 2. Evaluate incremental cost models for HA systems

v. Dec.12, 2005

Instrumentation and Controls 16

3. Demonstrate dual redundant backbone system. 4. Demonstrate proof-of-principle conceptual designs of typical applications. 5. Adopt hardware, software standards for ILC controls down to instrumentation

subsystem interfaces.

v. Dec.12, 2005

Instrumentation and Controls 17

11.2 Timing System

Overview Precision timing is needed throughout the machine to control RF phase and time-sampling beam instrumentation. The Timing System is envisaged to emulate the architecture of the control system with a centrally located dual redundant source phase-locked to a satellite reference, distributing redundant fiber signals to all machine sector nodes for further local distribution. Timing is phase locked to the RF system and stability at the point of RF measurement and control has to be ~10 picoseconds, and for the bunch compressor short-term stable to ~100 femtoseconds. The system will be designed to auto-failover in the event of a link failure between main control and a remote node. The fibers are temperature compensated to achieve phase stability by use of a shorter series section of fiber in a temperature controlled environment. The system builds on prior investigations done at NLC and uses commercial off-the-shelf components.

1. Baseline Configuration

1.1 Functions Global timing provides

a. Master oscillator distribution (1.3 GHz) b. 5 Hz timing fiducial distribution c. Programmable triggers for field hardware d. Mechanism to synchronize software processing to timing events e. Time fiducials for synchronized timestamps for software and hardware events.

1.2 Key parameters that influence Timing

Bunch Compressor Phase Tolerance

0.03 to 0.1 degrees at 1.3 GHz

Inter-linac timing tolerance 100 femto-seconds

1.3 Description The timing system will be fully redundant. The master oscillator and 5Hz timing fiducial will be distributed in a star configuration to sectors via redundant active phase stabilized fibers. Required timing triggers and other frequencies will be developed locally at sector locations. Figure 1 is a block diagram of the central timing distribution system.

v. Dec.12, 2005

Instrumentation and Controls 18

Figure 1. Block diagram of main timing showing star distribution of master oscillator and 5 Hz fiducial.

Figure 2 shows distribution of timing references to local controllers and the LLRF system.

Figure2. Sector Timing Distribution

2. Features Summary

2.1 Master 0scillator/ 5 Hz timing fiducial distribution The master oscillator will be distributed via dual redundant fiber optic links in a star configuration to each sector. Each link will use a phase stabilization scheme similar to that described in reference 2. The 5Hz timing fiducial will be encoded within the master oscillator waveform by a momentary phase shift.

2.2 Programmable triggers for field hardware The master oscillator and 5 Hz fiducial will be processed by a local module to

v. Dec.12, 2005

Instrumentation and Controls 19

produce programmable triggers for field device. Individual triggers will be produced by counting down the master oscillator or a lower derived frequency where appropriate to produce required triggers. Additional programmable delays will be provided where necessary. A graded approach to triggers will be used depending on the particular requirements;i.e., high precision (pico-second) timing will be provided for devices such as kickers, medium precision (nano-second) for devices such as septa and low precision (microsecond) for non time critical applications. Additional frequencies (3MHz for injector, damping Ring RF (TBD), damping ring revolution clock (TBD), 54 MHz (lasers)) will be generated where needed by synchronous digital dividers operating off the received master oscillator. These additional frequencies will be phased to the 5 Hz fiducial.

2.3 Synchronizing software processing to timing events A means of synchronizing software processing to timing events will be provided. This may be done by encoding events on the master oscillator distribution but this may affect overall received master oscillator stability. A separate event link may be used as is done at APS and SLS. Event system modules are commercially available. The event system rate could be synchronous to a subharmonic of the master oscillator if desired. A third method is to use a local module to count down a subharmonic of the master oscillator to generate interrupts at desired times.

2.4 Synchronized time stamps for software and hardware events. It is desirable for all remote processors, whether embedded or not to have the same concept of time; i.e., have their time-of-day clocks synchronized to a common source. This feature has been provided in the past by the timing event system. A similar capability is needed for the ILC. The time-of-day clock in conjunction with special timing hardware at each remote processor can provide synchronized time stamps to the microsecond or better resolution. A pulse ID number will be developed that identifies pulses within the 1 millisecond pulse train. This ID number will accompany data relating to individual pulses. The timing system also provides timing references to the MPS system to time stamp hardware events to aid in unraveling cascades of trips.

2.5 Other It is anticipated that global timing will be line-locked.

2.6 Present State of the art Reference 2 gives results of a fiber stabilization scheme that reported short term stabilities of +/- 750 femtoseconds and jitter of 250 femtoseconds RMS in a band of 10 seconds to 10 kHz. Reference 3 describes a scheme for delivering phase stable rf reference to antennas located as far as 15 km from the source. The frequencies involved are considerably higher (27 – 142 GHz) and the integrated phase noise measured at 81 GHz is 0.018 radians from 1kHz to 10 MHz offset. Timing event system hardware modeled on the APS system is commercially available from reference 4.

v. Dec.12, 2005

Instrumentation and Controls 20

2.7 Path to Specification The following items need to be resolved

a. Required timing tolerances, both drift and jitter, for each of the field devices. b. Required master oscillator stability (drift and jitter) for each RF device. c. The location and number of master oscillator receiver nodes. d. Interfaces between timing RF and Detectors need to be defined.

2.8 Required R&D a. The phase stabilization scheme described in reference 1 will be prototyped using newer high performance components. b. The bunch compressor (and perhaps devices at the interaction region) require a higher level of phase stability then the Linac RF modules. Methods to deliver such higher stability need to be studied. c. Evaluate cost/performance of different methods of synchronizing software processing to timing events.

2.9 References: 1. Larsen, R.S. , Technical Systems Configurations -Electrical Subsystem: Instrumentation – Timing, Rev.1, March 23, 2001 http://docdb.fnal.gov/ILC-public/DocDB/ShowDocument?docid=107 2. Frisch, J., Bernstein, D., Brown, D., Cisneros, E. “A High Stability, Low Noise RF Distribution System”, Proceedings of PAC2001, pp 816-818. http://docdb.fnal.gov/ILC/DocDB/0000/000035/001/PhaseAndTiming.pdf3. Shillue, B. “High Frequency Local Oscillator Transmission for the Atacama Large Millimeter Array (ALMA)”. 4. Micro-Research Finland Oy, http://www.mrf.fi

3. Cost Model The cost model will be developed from re-modeling of the previous bench prototype that was successfully tested for the NLC [Ref. 2]. A complete system design needs to be modeled with the machine sector model, numbers of distribution points from each node per sector, fibers and cables, and temperature – phase compensation. A bottom-up system cost model will be produced following cost estimating rules of the GDE.

4. Alternate Configuration No alternates under active consideration at this time.

v. Dec.12, 2005

Instrumentation and Controls 21

11.3 Diagnostic Interlock Layer

1. Overview The Diagnostic Interlock Layer is a proposed new control layer for High Availability (HA)design of the power systems in the ILC. Hard interlocks between power sources such as magnet power supplies and klystron modulators are typically a major cause of machine interruption. Typically interlocks for systems are summed into chains and often lack diagnostics to quickly isolate faults and return the machine to service. A new Diagnostic Interlock Layer (DIL) is envisaged to solve this problem in the ILC. All modules in a power system would be equipped with such a diagnostic with the capabilities if checking the health of the systems board-by-board and displaying summaries at any time in the control room. The DIL could have the local intelligence to take actions to avoid impending trips due to over-voltage, over-current or over-temperature. It also could read out critical switching waveforms for example to pinpoint problems before they cause a shutdown. Evasive actions could include lowering current or voltage in a module while others compensate, turning off a faulty unit and flagging maintenance to hot swap it before another failure interrupts the machine. Special network hardware, software, controls and displays are envisaged that are accessible to operators and maintenance technicians from any location in the accelerator complex.

2. DIL Design Principles The DIL main component is a small embedded card or chip that drops into a variety of power components. The idea is modeled after the commercial Intelligent Platform Management (IPM) system of the Telecom modular standard known as ATCA. In ATCA IPM performs the following tasks:

1. Senses faults in an individual carrier module or mezzanine module (subunit).

2. Removes primary power from the module (while retaining control power at all times)

3. Reports to main control that a module has failed, and lights a front panel LED to indicate it is safe to hot-swap.

4. After hot-swap, recognizes the new module type and power requirements.

5. Authorizes the delivery of primary power after checking that the Shelf total power will not be exceeded.

6. Monitors temperatures and power consumption of all modules in the Shelf.

7. Monitors for fan failures, and when one occurs, rearranges speed of the other fans (four total) to compensate, while reporting to main control the need to hot-swap one of the fans.

8. Other features can be programmed into the Shelf level at user discretion.

3. DIL Implementation in the ILC In the ILC the DIL will be part of a site-wide IPM system. Typically it would remotely set and monitor interlock trip limits, measure and display timing and trigger functions, and

v. Dec.12, 2005

Instrumentation and Controls 22

capture important waveforms. A response to impending or actual trips may be taken locally or by a higher controls layer. Operations and maintenance will routinely monitor the diagnostic layer to observe the health of running systems or to trace anomalous behavior. The following are some example DIL implementations:

1. Modulator Each module of a modulator contains a DIL sub-unit, such as an ATCA mezzanine board, which provides information constantly to the imbedded DIL unit manager, and secondarily to the master system at main control. See Fig. 1. Typical features are:

a. Disable a faulty module and restore total voltage output without machine turnoff..

b. Raise fan speed to compensate for increased module temperature.

c. Signal for replacement of failed modules or fans.

d. Send all diagnostic information to main control.

e. Capture important waveforms in memory on fault.

Quick repair of failed modules guarantees near 100% system availability and minimizes the number of spare units needed in the total modulator subsystem.

2. DC Power Supplies

DC power supplies designed as 1/n modular units gain similar advantages using a DIL as in the modulator example. The concept is shown in Fig. 2. A smaller multichannle supply can include a switcahbel spare channel as shown in Fig. 3. All units will contain DIL diagnsotics chips or small boards.

v. Dec.12, 2005

Instrumentation and Controls 23

Fig.3. Corrector Subunit Supplies on Carrier Module with Imbedded IPM/DIL System

3. Instrument Modules

The DIL layer in instrument modules will primarily be the ATCA IPM layer. All the hooks for basic platform management and hot swap will be part of a standard commercial chip. All higher level trip-avoidance strategies will be programmed at the central computer IOC level.

/DIL Dual Serial IPM Control Imbedded IPM Chips • Process RF stations under independent algorithm control at each station to minimize processing time as well as arcing.

• Track temperature of power components and reduce power or raise trips with local intelligence.

• Magnet over-temperature monitoring by analog means, such as resistance monitoring, in addition to klixon protection to detect, report normally hidden faults.

• Integrate flow switches to eliminate nuisance trips due to bubbles; monitor

v. Dec.12, 2005

Instrumentation and Controls 24

temperature of device as final protection trip source; schedule backflush, inspection on routine basis by remote valve control.

• Shut off LLRF level and Modulator source immediately on klystron or waveguide arc and compensate with standby station by MPS timing action.

• Develop quick diagnostic and recovery strategies for tuners that lose reference during operation; work around failed or stuck tuners; compensate for loss f beam power automatically for mistuned but still functioning station.

• Detect and report tripped off pumps, cable faults. Systems are redundant so loss of one pump is tolerable. Repair while machine operates except for cable end in machine tunnel.

4. Interlock Trip-Avoidance The system will be designed to avoid various kinds of trips by appropriate strategies. Examples are:

It’s unclear how to apply DIL to cryogenics instrumentation, which consists primarily of temperature and pressure monitoring along the machine, with constant feedback to the main cryogenic plants. Ideally, but not likely, cryogenic systems should have 1/n redundancy to work around a failed or compromised unit. Modular system design should be investigated.

5. DIL Initiative: Power System Control Board (PSCB) A DIL board with generic features for power systems monitoring is under development, to be used first on the Marx modulator. Provides independent trigger timing and pulse width control for each stage; fiber-optic isolation; IGBT temperature monitoring; transient waveform capture, monitor and set voltage references etc. See Ref. 1.

6. References 1.Diagnostic Interlock Layer, Draft, Ray Larsen. http://docdb.fnal.gov/ILC/DocDB/0001/000103/001/10.3.R.1%20DIL%20Draft%20110105.doc 2.Performance Specification for the Power System Control Board (PSCB), PS-390-000-01, Rev. August 15 2005, P. Bellomo et al. http://docdb.fnal.gov/ILC/DocDB/0001/000102/001/10.3.R.2%20PSCB%20Performance%20Specification%20R1.doc

7. Cost Estimation The DIL system consists of standard commercially available components, printed circuit assemblies, fiber networks and system controllers. A bottom-up system cost estimate will be made following GDE standard cost procedures.

8. R&D Path • Complete prototype hardware, software, interface to control system (FY06) • Test on Marx prototype modulator (FY07) • Develop HA software layer to test integrated features on HA platform (FY07) • Develop, test highly integrated version of DIL hardware (FY07-8)

v. Dec.12, 2005

Instrumentation and Controls 25

11.4 Global Network

1. Overview The ILC collaboration is truly an international collaboration with expertise in building, commissioning, and running the machine distributed on a global basis. It will require personnel with highly specialized skills in commissioning and running a machine with this level of complexity and sophistication. Given the length of construction and lifetime of the machine, concentrating all of that expertise in a single physical location is unrealistic. This means that installing, debugging, commissioning, trouble shooting, and performance tuning the machine will have significant contributions by (or under the direction of) remote experts of the collaborating institutions. The control system must support intense access and control of components, subsystem or even the entire facility from remote location(s).

2. Design & Implementation Features The sheer size of the ILC accelerator (40km) dictates that the control system be built on a hierarchy of network interfaces, building from the local machines and diagnostic segments into an operation center that manages the entire complex. Even the operation center that is closest to the machine itself is remote and can be considered as one [of a few] remote control centers, with no more authority or privilege than a center that is a continent away. Multiple operation centers inherently provide a failover for control if the primary center becomes inaccessible. Appropriate levels of secure access must to be designed into the system up front. This aspect may also have in impact on control system architecture, the choice of standards and platforms. In particular an extensive diagnostic layer of the hardware components and access by the control system need to be carefully designed into the system. A layered authorization and authentication system will be necessary to allow a well organized and efficient operation of the facility (where operation includes debugging, maintenance, trouble shooting, repair, tune-up, upgrades). Clear and effective inter-center communication is crucial and will be facilitated by advances in telecommunications and video links. Additional developments in commercial and freeware collaboration tools must be also explored with the option of in-house developed tools such as e-logs. The remote operation centers will also provide an important outreach opportunity for the public to see firsthand how research is progressing at the ILC. Visitors will be able to see current activities, and gain a better understanding of the involvement of their home institutions/countries/regions in a research effort on a global scale.

3. Needed R&D Layered Software: The R&D effort for tools and systems which support these needs should include the analysis and examination of the layered control system approach in the context of remote

v. Dec.12, 2005

Instrumentation and Controls 26

operation.

Access by Hardware Experts: Since it is expected that the accelerator complex will be accessed by a large number of hardware experts outside the active control center, appropriate management software must be designed which supports the chief of operation in coordinating these efforts during accelerator operation. The remote experts must be supported by portable audio, video capture systems and extensive use of virtual instruments. The components to implement these feature exist, but they must be modified and tailored to according to the needs of the ILC operation.

Examples: An R&D effort already started is the GANMVL project within EuroTev, which is developing a communication tool to support remote experts in participating in on-site efforts. http://ulisse.elettra.trieste.it/mvlgan/ The design of a remote operations center is currently underway at FNAL for remote access to the CMS experiment at CERN as well as an interface to remote viewing of the LHC accelerator controls system. It is envisioned that the personnel at the FNAL center will participate in the operational shift rotation schedule for CMS. http://home.fnal.gov/~eeg/remop.html

4. Cost Estimation The incremental costs of additional regional facilities needed for the Global Network access will be addressed at the appropriate time using the standard GDE cost estimating procedures.

v. Dec.12, 2005

Instrumentation and Controls 27

11.5 Machine Protection System

1. Overview The ILC Machine Protection System (MPS) is that collection of devices intended to keep the beam from damaging machine components. With nominal average beam power of 20 MW, consisting of 14000 bunches of 2e10 ppb each per second, and typical beam sizes near 10 x 1 micron, both the damage caused by a single bunch and the residual radiation or heating caused by small (fractional) losses of a many bunches are important for MPS. The MPS consists of 1) a single bunch damage mitigation system, 2) an average beam loss limiting system, 3) a series of abort kickers and dumps, 4) a restart ramp sequence, 5) a beam permit system, 6) a fault analysis recorder system, 7) a strategy for limiting the rate with which magnetic fields (and insertable device positions) can change, 8) a sequencing system that provides for the appropriate level of protection depending on machine mode or state, and 9) a protection collimator system. The systems listed must be tightly integrated in order to minimize time lost to aberrant beams and associated faults. Several of the systems listed below provide redundant protection mechanisms. The BCD recommendation is to adopt them all as listed. Alternate designs with less redundancy are proposed as ‘ACD’. 2. Baseline Configuration a. Single Pulse Damage Single bunch damage will be mitigated by systems that check the preparedness of the machine before the high power beam passes by. Single bunch damage control is only necessary in the ‘damped-beam’ section of the ILC, where the beam area is less than 50 micron^2 (2e10). Single bunch damage mitigation will be done using two basic subsystems: 1) a leading benign pilot bunch and 2) a beam permit system that surveys all appropriate devices before damping ring beam extraction begins and provides a permit if each device is in the proper state. In addition, some exceptional devices (damping ring RF and extraction kickers for example) will need fast monitoring systems and redundancy. The pilot bunch is one percent of nominal current and is spaced 10 usec ahead of the start of the nominal train. The pilot bunch must traverse the machine properly before the rest of the train is allowed to pass. Indeed, each bunch must traverse properly or the abort system will be triggered. Proper passage is sensed using the beam position monitors and beam intensity monitors as these are the only beam diagnostics with true single bunch response time. It is important to note the resolution requirements: BPM’s must have resolution and systematic offsets not more than 10 times worse at the low end of the intensity range 2e9 ppb to 2e10 ppb. (Availability and failed reading rates must also be very good). The actual required pilot beam BPM resolution may depend on the performance of the RTML/BDS collimation. If an errant trajectory is sensed, the nearest upstream abort system is triggered. Assuming the latency for detecting the fault is 500 ns, the upstream signal effective propagation speed is 0.7 c, and the abort kicker latency time is 1 us, the maximum kicker spacing should be 1000m. This ensures that the kicker can be turned on quickly enough to dump or spoil the high intensity bunches that follow the pilot bunch. Only those bunches extracted from the damping ring before the abort signal is sensed and received at the ring need to be dumped and the damping ring extraction sequence will be terminated, leaving what is left of the partially extracted beam train stored. Given that the

v. Dec.12, 2005

Instrumentation and Controls 28

time needed for the beam to go from the damping ring to the main beam dump is 67 us, in the worst case, (when the downstream most sensor detects a fault condition from the pilot), and the signal return time to the damping ring is another 100 us, roughly 450 bunches need to be dumped. The detailed timing will depend on the ring / linac transfer line geometry. Since there is more than one dump line, not all of these need to be dumped in one place. The injector complex must include systems that reliably generate the pilot bunch. Extraction from the ring should not begin unless the pilot is within allowed limits; its intensity should be high enough for the trajectory sensors to read and respond reliably yet below the single damage threshold, expected to be around 1% for bunches which are intended for the whole machine. There may also be a need for a benign pilot bunch of nominal intensity but much larger emittance. b. Average beam loss limiting system Average beam loss will be limited, throughout ILC, using a combination of radiation, thermal, beam intensity and other special sensors. This system will function in a manner similar to other machines, such as SLC, LHC, SNS and Tevatron. If exposure limits are exceeded at some point during the passage of the train, damping ring extraction or source production (e+/e-) will be stopped. For stability, it is important to keep as much of the machine operating at a nominal power level. This is best done by segmenting it into operational MPS regions. For the BCD, there will be 11 of these regions, as noted in table 1. Since the fault response can (and will) occur during the train, and since there will be 9 full power shut-off points, each with an extraction system and a full capacity dump, and we can expect to have trains of different lengths in the machine on any given pulse. The average beam loss MPS will be applied throughout the complex, including the source, damping ring injector and the damping ring itself.

Region name Begin End 1 e- injector Source (gun) e- Damping ring injection (before)

2 e- damping ring Ring injection e- Ring extraction (after) 3 e- RTML Ring extraction e- Linac injection (before) 4 e- linac Linac injection Undulator (before) 5 Undulator Undulator BD; e+ target 6 e- BDS BD start e- Main dump 7 e+ target e+ target e+ damping ring injection 8 e+ damping ring Ring injection e+ ring extraction 9 e+ RTML ring extraction e+ linac injection 10 e+ linac linac injection e+ BDS 11 e+ BDS e+ BDS e+ main dump

Table 1: beam shut off points. Each of these segmentation points is capable of handling the full beam power, i.e. both a kicker and dump are required. These systems also serve as fast abort locations for single bunch damage mitigation.

v. Dec.12, 2005

Instrumentation and Controls 29

c. Abort kickers and dumps Abort systems are needed to protect machine components, especially the superconducting cavities, from single bunch damage. It is expected that a single bunch impact on a niobium iris will leave a small hole, roughly the diameter of the beam, through which the helium will flow. The minimal abort system consists of a spoiler / collimator / absorber block (copper) and a kicker. The kicker rise time should be fast enough to produce a guaranteed displacement of more than the pipe radius in an inter-bunch interval. In any given fault, at most 450 bunches would then strike the copper block. It is expected that the upstream block surface would be marred with a sequence of small impact holes, but that the block would not fracture and would not require cooling. If the block is thick enough to absorb the full shower, the energy associated with 450 bunches should be less than 400kJ (250 GeV) and the block temperature will rise about 4 degrees. Since each abort precedes a cool down interval, a post-mortem evaluation of the fault, response to the fault and the restart sequence, (total time to recover may be close to a minute) the average power on the block should be very low. Even if the abort recovery sequence is made fast, care must be taken to avoid a rapid sequence of identical fault events. If this occurs, the block thermal interlock may trip. The block volume should be more than one cubic foot. RD is needed to validate this general concept. The beam aperture should be less than 20mm as defined by nearby protection collimators. In the baseline configuration five abort systems are needed on the electron side (four on the e+ side): 2 upstream of the linac, one upstream of the undulator and 2 in the beam delivery. An alternative is an additional abort per kilometer of linac. ACD RD will be required to determine if additional abort systems are required. This may depend on the linac straightness. The required kicker deflection is 10 mm, for the radius, and a relatively small additional amount for margin. With a kicker volume of 20 * 20 mm, about 25 MW of peak power would be required for a 50 m long kicker system [1]. RD is needed to reduce this requirement and to make a system with an appropriate safety factor. The total length associated with abort systems is 200 m per side (BCD), an additional 400 m / side for ACD. In the beam delivery and the RTML, 2 of the abort system can be integrated with the tune up dumps. The abort system can also be triggered during the train, if a serious trajectory distortion is detected. The kickers must be triggered as close as possible to the preceding bunch so that no bunch is kicked incompletely. d. Restart Ramp sequence

Depending on the beam dynamics of the long trains, it may be advisable to program short trains into a restart sequence. There may also be single bunch, intensity dependent effects that require an intensity ramp. In order to avoid relaxation oscillator performance of the average beam loss MPS, the system will be able to determine in advance if the beam loss expected at the next stage in the ramp sequence is acceptable. Given the number of stages and regions, the sequence controller must distribute its intentions so that all subsidiary controls can respond appropriately and data acquisition systems are properly aligned. The sequence may need to generate a ‘benign’ bunch sequence with the nominal intensity but large emittance. The initial stages of the sequence will be used to produce ‘diagnostic’ pulses to be used during commissioning, setup and testing.

v. Dec.12, 2005

Instrumentation and Controls 30

e. Beam Permit System

For all of the ILC, there will be a controls-based survey of critical subsystems, about 1 ms before the pulse begins, in order to determine readiness. All magnet systems, key RF and controls systems and collimators will be included in the survey. For the damped beam part of the system, this is redundant with the pilot bunch. f. Fault analysis recorder system

A post mortem analysis capability is required that will capture the state of the system at each trip. This should have enough information to allow the circumstances that led to the fault to be uncovered. Data to be recorded on each fault should include: bunch by bunch trajectories, loss monitor data, machine component states (magnets, temperature, RF, insertable device states), control system states (timing system, network status, diagnostic) and global system status (sequencer states, PPS, electrical, water and related sensors). The fault analysis system should automatically sort this information to find what is relevant. g. Rapidly changing fields In addition to the above, there are critical devices whose fields (or positions) can change quickly, perhaps during the pulse, or (more likely) between pulses. These devices need 1) special controls protocols, 2) redundancy or 3) external stabilization and verification systems. 1) Depending on the state of the machine, there should be programmed (perhaps at a very low level) ramp rate limits that keep critical components from changing too quickly. For example, a dipole magnet should not be allowed to change its kick by more than a small fraction of the aperture (few percent) between beam pulses during full power operation. This may have an impact on the speed of beam based feedback. Some devices, such as collimators should be effectively frozen in position at the highest beam power level. There may be several different modes, basically defined by beam power, that indicate different ramp rate limits. 2) There are a few critical, high power, high speed devices (damping ring kicker, RF, linac front end RF, bunch compressor RF and dump magnets) which will need some level of redundancy in order to reduce the consequence of failure. In the case of the extraction kicker, this will be done by having a sequence of independent power supplies and stripline magnets that have minimal common mode failure mechanisms. In the case of the front end and bunch compressor RF, there will be more than one klystron / modulator system powering a given cavity through a tee. The LLRF feedback will be used to stabilize the RF in the event that one of sources fails ‘mid-pulse’. There are alternate methods of doing this, for example using a sequence of modestly powered devices controlled completely in parallel, as in the case of the critical damping ring extraction system. 3) There are several serious common mode failures in the timing and phase distribution system that need specially engineered controls. This is necessary so that, for example, the bunch compressor or linac common phase cannot change drastically compared to some previously defined reference, even if commanded to do so by the controls, unless the system is in the benign – beam tune up mode. h. Sequencing system depending on machine state

v. Dec.12, 2005

Instrumentation and Controls 31

The ILC will be divided into segments delineated by beam stoppers and dump lines. There may be several of these in the injector system, two beam dumps in each RTML, and 2 (or 3) in the beam delivery and undulator system. In addition, the ring extraction system effectively operates as a beam stopper assuming the beam can remain stored in the ring for an indefinite period. This part of the MPS assumes that the beam power in each of these segments can be different and reconfigures the protection systems noted above accordingly. i. Protection collimators

The entire ILC will require protection collimators that effectively shadow critical components. In the main linac, for example, there should be a few collimators per betatron wavelength, with an aperture of about ½ of the nominal aperture. These devices must be engineered to withstand innumerable single pulse impacts. There is a collimation system that defines the launch into the main linac.

j. Supporting Documentation

[1] Mattison, T. NanoBeam 2005 presentation. http://atfweb.kek.jp/nanobeam/files/presen//presen-WG2a-12.pdf

[2] Ross, M., http://www.slac.stanford.edu/cgi-wrap/getdoc/slac-pub-8605.pdf. [3] http://www-project.slac.stanford.edu/ilc/MPS_design_rules.htm[4]http://alcpg2005.colorado.edu/alcpg2005/program/accelerator/GG3/peter_tenenbaum20050812215847.ppt[5] http://lhc-mp-review.web.cern.ch/lhc-mp-review/Review-Programme.html

k. Required Research and Development

RD is needed to: 1) determine the abort kicker and spoiler system. Tests indicating the single pulse damage threshold were done in 2000 and need to be extended to study effects of single pulse impacts in niobium, related ILC vacuum chamber construction materials (titanium, stainless steel and aluminum) and on thicker samples (such as the proposed absorber block. RD is also needed to determine how to resurface the material ‘in-situ’ so that an accumulation of many small impact holes does not reduce its integrity. Systems of the highest value (cryo-cavities, collimators and instrumentation) need specific testing in order to determine the threat posed by single bunch impact. Components such as vacuum chambers and associated hardware may be protected by much simpler means and tests are needed for these also. 2) to develop the kicker system. The system is fairly high power (25MW) and must have some degree of redundancy and fail-safe performance. It should be designed to reduce the possibility of partial-kicks. RD is also needed to determine the effect of small holes in the niobium cavities. It may be possible to protect the most vulnerable section of the cavity (iris) with some backing material. 3) An evaluation of possible failure modes must be made in order to determine the risk associated with the lack of abort systems in the linac.

v. Dec.12, 2005

Instrumentation and Controls 32

4) Controls RD is needed to determine the integration strategies for the ramp rate limit and mode distribution system.

3. Alternative lower cost Configuration a. Description Alternate MPS designs, with lower initial investment cost, involve the removal of one or more of the above systems. For example, the pilot bunch may not be needed if the possibility of strong transverse kicks in the linac is shown to be acceptably small, and the BDS / BC systems are shown to be robust enough. Since the pilot bunch is primarily an operational choice, its direct cost impact is small, so this alternate is not very compelling. Other systems, such as redundant upstream RF and abort kicker/ dump systems may also be eliminated as part of an ACD evaluation. These have substantial costs and will need careful evaluation.

v. Dec.12, 2005

Instrumentation and Controls 33

11.6 RF Control System – Low Level RF

11.6.1 Overview Low Level RF encompasses the programming and regulation of the cavity field vector, as well as the control and protection of other parts of the RF system. The cavities in the Main Linacs are operated in pulsed mode at gradients of 35 MV/m with each klystron driving 24 cavities. The pulse repetition rate is 5 Hz and the rf pulse length is 1370 µs from which 420 µs are required for cavity filling while the remaining 950 µs with constant gradient (flat-top) are necessary for the acceleration of the beam. During the beam pulse the fluctuations of the accelerating field - defined as the vector- sum of the fields in 24 cavities - must be kept small. The major sources of field perturbations, which have to be controlled by the low level RF system, are caused by fluctuations of the resonance frequency of the cavities and by fluctuations of the beam current. Fluctuations of the resonance frequency are a result of deformations of the cavity walls induced by mechanical vibrations (microphonics) or the gradient dependent Lorentz force (also referred to as radiation pressure or ponderomotive force). Slow changes in frequency, on the time scale of minutes or longer, are corrected by a frequency tuner, while faster changes are counteracted by fast amplitude and phase modulation of the incident RF power.

11.6.2 Functions LLRF programs and regulates the cavity vector; provide resonant control of the cavity, program control of the modulators and machine protection. The system must diagnose fault conditions and handle fault recovery. It is also the primary data acquisition system for the overall performance of the RF system and provides information such as RF heat load to cryogenics and LCW systems. Each LLRF station will part of the interlocks and beam inhibit and aborts. Fast amplitude and phase control can only be accomplished by modulation of the incident wave that is common to the 24 cavities. The incident wave may be adjusted on a slower time scale with motor driven three stub tuners.

11.6.3 Key parameters The regulation specifications of the bunch compressor sections of 0.08% and 0.03 degrees are the tightest requirement for a single station. The rest of the Linac requirements are somewhat relaxed at least for non-correlated errors. Error in the cavity phase has many sources that sum, including the master oscillator, RF reference distribution, local oscillator, cavity probe, cable, ADC clock, modulator ripple, Klystron etc. These errors are composed of drifts and noise terms and must be understood fully in both time and frequency domains. Fast resonant control is implemented with piezo actuators that are programmed to cancel the Lorentz detuning and to damp microphonics. The required travel and bandwidth of the actuators are a function of the cavity and cromodule design.

11.6.4 Description Fast amplitude and phase control can only be accomplished by modulation of the incident wave which is common to the 36 cavities. The modulator for the incident wave is designed as an I/Q modulator to control the in-phase (I) and quadrature (Q) component of the

v. Dec.12, 2005

Instrumentation and Controls 34

cavity field. This scheme minimizes coupling between the loops and guarantees control in all four quadrants. The overall scheme of the rf control system is shown in Fig.2. ??? The detectors for cavity field, and incident and reflected wave are implemented as digital I/Q detectors. The rf signals are mixed with a local oscillator to down-converted to an IF frequency in the range of 50 MHz and sampled at a rate that harmonically related to the 1300 MHz RF. The exact frequency will be based on the state of the art digitizers and will be optimized for signal quality. Present 14 bit digitizers are operating up to 105 MSPS. The digitized signal is then down-converted to baseband to produce I and Q data streams that describe the cavity field. The IQ analytic signal for each cavity is multiplied by a 2x2 matrix that calibrates the magnitude and phase for each cavity probe. The vector-sum is calculated and corrected for systematic measurement errors. Finally the set point is subtracted and the compensator filter is applied to calculate the new actuator setting (I and Q control inputs to a vector modulator). Feed forward is added from a table in order to minimize the control effort. The feed forward tables are adaptively updated to reflect slowly changing parameters such as average cavity detuning, changes in klystron gain, phase shift in the feed forward path, and general changes in

11.6.5 Automation

The operation of the more than 560 linac rf systems will be highly automated by the implementation of a finite state machine finite state which has access to high level applications including the adjustment of the loop phase, vector-sum calibration, frequency and waveguide tuner control, and exception handling. The area of automation is viewed as a major area of R&D needed for the successful operation of the accelerator complex. The needs of RF system automation will help to define the structure and complexity of the control system.

11.6.6 Present State of the art As in much of the instrumentation electronics, LLRF is benefiting from the wireless telecom industry. The core components of ADC, DACs and FPGAs have seen dramatic advancements in the last 10 years. While there may be a flattening of the rate of advancement in the foreseeable future, one can only make a general guess what will be available in the next several years. While the requirements for stability and regulation are quite stringent and much R&D is needed we believe that the technology is currently available to realize a design. In order to be ready for a final design, a robust continuous design effort must be in place.

11.6.7 Path to Specification 1. More complete modeling of longitudinal plane for the damping rings and main linac will provide a better idea of the full regulation requirements. This model must include both the systematic and stochastic errors and noise of the ground motion, cryomodules and RF system. 2. Identify bunch compressor feedback loop requirements. 3. Identify any additional beam-based control loops required from gun to IP. 4. As new technologies are developed for modulators, such as Marx Generators, new specification for both modulator spectral content and LLRF bandwidth requirements need

v. Dec.12, 2005

Instrumentation and Controls 35

to be spelled out. 5. Error budgets that span many systems are needed. This must cover areas that are sources of disturbance as well as the regulation performance of the LLRF control loops.

11.6.8 Required R&D There are many aspects of R&D for LLRF for a machine on this scale. Many present LLRF systems need careful and time consuming calibration and adjustment by experts. Fault diagnosis is often time consuming and can generate significant down time. Key improvements are needed for installation and commissioning, fault diagnosis and recovery.

Beam based energy and phase global controllers need to be developed and studied on existing machines. “Variations in beam intensity creates systematic errors for both the damping rings and Linac. However, because the individual bunch intensities are known while in the damping ring, feedforward information can be passed up the Linac before beam extraction. Also because most of the cavity errors are within 100kHz of bandwidth, uncorrected errors from individual LLRF controllers may be accumulated and passed on to other station controllers. This approach has the potential to increase overall Linac performance by more than 5% as klystrons could be operated closer to the true required power level with less overhead for feedback.

Platforms are a major concern for instrumentation, controls and LLRF. It would be idea if there is a single platform for all systems. One the other hand standardization must not compromise individual system performance. There are many approaches within the LLRF community to this issue which includes standard instrumentation crates like VXI, Cpci and PXI, to crateless systems with Ethernet interface. As mixed signal IC technology advances and the total electronic circuit footprint is reduced, the value of a crate based system is reduced. Crates may keep there advantage especially if multiple systems are able to share crates and even processing modules.

R&D with the current state of the art components in ADCs, DACs, FPGAs, DSPs and analog processing must be an ongoing process. Keeping up with current computer science advances is equally important. Because of the large effort involved, it is imperative that this work span many institutions in open collaborations

11.6.9 References: 1.

11.6.10 Alternate Configuration No alternates under active consideration at this time.

v. Dec.12, 2005

Instrumentation and Controls 36

11.7 Beam-based Feedback Systems

Overview

Beam-based dynamical feedback control is essential for meeting the high performance and luminosity needs of the ILC accelerator.

Beam based feedback systems will stabilize the electron and positron orbits in the damping rings and trajectories throughout the machine. It will also be used to correct for emittance variations, and for measurement & correction of dispersion in the Main Linac.

Detailed requirements for several of the beam-based feedback systems are included in other sections of the overall BCD, with only a brief synopsis of physics requirements being provided in this section.

Only beam-based feedback systems are discussed in this section, all of which employ instrumentation such as beam position monitors (bpms) and fast kickers. Other (non-beam based) feedback systems, such as cavity temperature control are considered out of scope for this section

A summary of anticipated beam-based feedback loops follows…

Damping Ring: Injection trajectory control Purpose: maintain injection efficiency close to 100% Monitors: injection orbit via bpms Actuators: setpoints for injection kicker and septum. Correction plane: horizontal Correction sampling rate: 5Hz

Damping Ring: Dynamic orbit control Purpose: compensate for drift and low frequency disturbances to keep beam

through center of the multipoles Monitors: closed orbit via NN bpms. Actuators: MM correctors. Correction plane: horizontal and vertical Correction sampling rate: 10-20KHz.

Damping Ring: Bunch-by-bunch transverse feedback Purpose: reduce coupled-bunch instabilities. Monitors: single wide-bandwidth bpm to provide bunch-by-bunch signals. Actuators: fast deflecting cavity or striplines. Correction plane: horizontal and vertical Correction rate: full bunch rate (500/650MHz)

Damping Ring: Extraction orbit control Purpose: preserve emittance through extraction septum

v. Dec.12, 2005

Instrumentation and Controls 37

Monitors: emittance of extracted beam from RTML Actuators: correctors in damping ring. Correction plane: horizontal and vertical Correction sampling rate: 5Hz

Ring to Main Linac: Pre-Turnaround emittance correction. Purpose: reduce emittance growth Monitors: emittance measurement. Actuators: dipole correctors and skew quads Correction sampling rate: 5Hz for dipole correctors, <1Hz for skew quads

Ring to Main Linac: Turnaround trajectory feed-forward Purpose: correct for extraction kicker jitter. Monitors: beam trajectory measured upstream via bpms. Actuators: 2 fast correctors per plane. Correction plane: horizontal and vertical Correction sampling rate: bunch spacing (~3MHz)

Ring to Main Linac: Post-Turnaround emittance correction Purpose: minimize emittance growth. Monitors: emittance measurement. Actuators: 4 skew quads Correction sampling rate: 5Hz for dipole correctors, <1Hz for skew quads

Ring to Main Linac: Beam energy at bunch compressor (two stages) Purpose: control the final beam energy Monitors: bpms in high-dispersion sections. Actuators: klystron phase shifters

Correction sampling rate: 5Hz Main Linac: Trajectory Feedback (several cascaded loops)

Purpose: compensate for drift and low frequency disturbances to keep beam through center of multipoles and RF cavities.

Monitors: multiple bpms in each large section. Actuators: nominally 4 horizontal and 4 vertical correctors per section. Correction plane: horizontal and vertical. Correction sampling rate: 5Hz.

Main Linac: Dispersion measurement and control Purpose: provide means to measure dispersion; provide means to apply local

dispersion correction. Monitors: dispersion measurement, laser wire. Actuators: use local RF amplitude control to generate local dispersion ‘bumps’

(Dispersion free steering). Correction sampling rate: ??

Main Linac: Beam energy (several cascaded sections) Purpose: control the final beam energy

v. Dec.12, 2005

Instrumentation and Controls 38

Monitors: bpms in high-dispersion sections. Actuators: klystron phase shifters Correction sampling rate: 5Hz

Positron Source: Beam energy at undulator Purpose: control the final beam energy Monitors: bpms in high-dispersion sections. Actuators: klystron phase shifters Correction sampling rate: 5Hz

Beam Delivery System: Trajectory feedback from pulse to pulse Purpose: compensate for drift and low frequency disturbances to keep beams

directed towards the interaction point. Monitors: nominally 9 bpms per plane. Actuators: nominally 9 correctors per plane. Correction plane: horizontal and vertical. Correction sampling rate: 5Hz

Interaction Point: Trajectory feedback from pulse to pulse Purpose: maximize average cross-section of colliding beams Monitors: post-IP measurement of beam trajectory, beam charge Actuators: nominally one corrector per plane. Correction plane: horizontal and vertical Correction sampling rate: 5Hz

Interaction Point: Trajectory feedback within bunch-train Purpose: maximize bunch-to-bunch cross-section of colliding beams. Monitors: bunch-by-bunch bpms. Actuators: 2 fast kickers per plane. Correction plane: horizontal and vertical Correction sampling rate: bunch spacing (~3MHz)

Key parameters that influence beam-based feedback

Performance metrics for the beam-based feedback systems are entirely derived from the physics requirements of the ILC, specifically beam quality and beam stability at the IP. Derivative requirements will come from the physics requirements of each upstream machine (gun, damping ring, compressor, linac, beam delivery system).

Baseline Configuration

The following subsections provide more description of specific feedback loop requirements, and discuss overall architecture and infrastructure for implementing feedback loops.

v. Dec.12, 2005

Instrumentation and Controls 39

Damping Ring: Dynamic orbit control

Real-time dynamical orbit feedback will be used to stabilize the electron orbit both against long-term drift and against dynamic disturbances such as ground motion. Damping ring orbit stability requirements are likely to be similar to those for existing 3rd-generation synchrotron light sources

The orbit will be corrected using the response matrix method that takes orbit measurements from multiple bpms around the ring, and corrects the orbit using multiple correctors, also distributed around the ring. Algorithms and technology are well established in the synchrotron light sources, and usually revolve around solving a response matrix using least-squares or SVD (singular value decomposition). Sufficient phase advance is required between the bpms to properly sample the lattice phase response, similarly the correctors. It is typical to use a similar number of bpms as correctors in the algorithm. A larger number of bpms provides more robust measurement of orbit disturbances, while a increasing the number of bpms and correctors allows orbit corrections to be more localized.

To provide good feedback loop gain for ground motion and other sources of dynamic orbit motion in the 10’s or 100’s of hertz, a feedback loop sample rate of several kilohertz is required. For a 17km damping ring, it would be feasible to implement orbit correction turn-by-turn (ie at 17.6 kHz), but this would be considerably more challenging for a 6km ring.

Local processing will convert raw bpm button signals into x and y position information at the full turn-by-turn rate (or faster). These x and y positions will be pushed onto a fast synchronous network dedicated for real-time orbit correction. A reflective memory network would be the choice for this network today.

Similarly, local processing for the corrector magnet power supplies will receive newly updated setpoints for each corrector at the orbit correction sampling rate.

At one location in the network, a dedicated real-time processing crate will implement the orbit correction algorithm, receiving real-time bpm values from the reflective memory network, and synchronously pushing new corrector setpoints onto the network. The choice of a single local processing crate (over the distributed processing used in many systems today) has the advantage of increased flexibility of algorithms, and increased convenience for servicing, lower overall cost, and easier high availability implementation. Although this potentially creates an I/O bottleneck to the processors, rapid advances in digital processing technology mean it is likely this would be a non-issue by the time the ILC is built.

Trajectory Feedback Control

All transfer lines and the Main Linac will require trajectory feedback control. With the exception of the damping ring trajectory feed-forward system described below, all

v. Dec.12, 2005

Instrumentation and Controls 40

trajectory control systems will comprise the same basic elements use the same algorithm, and will use similar or identical hardware.

Implementation will be similar to that in the damping ring, with the trajectory control algorithm being developed from response matrices. A minimum of two bpms and two correctors are required to correct trajectory position and angle. The two bpms must be separated by sufficient phase advance (preferably 90 degrees), and similarly the correctors. A response matrix couples the corrector responses and bpm measurements. Bpms must be placed downstream of the correctors if closed-loop correction of the trajectory is required.

Unlike the damping ring, however, most of the trajectory correction loops will operate synchronously at the 5Hz ILC pulse rate. Bpm measurements taken at every bunch will be averaged locally, producing an average beam position for each pulse. These will be pushed onto the controls network, taking advantage of the middle-ware layer of the controls infrastructure. The trajectory control algorithm (implemented as middle-ware in the controls architecture) will synchronously calculate corrector magnet settings for the subsequent ILC pulse, distributing the corrector setpoints synchronously using the controls network.

Given the pulsed nature of the ILC beam that results in beam trajectory measurements at only 5Hz, the trajectory control system bandwidth will be restricted to a fraction of a hertz, making it effective only for long-term drift effects rather than for dynamical beam disturbances that will be corrected by the damping ring orbit correction system.

Ring to Main Linac: Turnaround trajectory feed-forward

Individual bunches are extracted from the damping ring by a fast extraction kicker magnet. Shot-to-shot fluctuations in the kicker amplitude translate to bunch-to-bunch trajectory errors of the extracted beam.

A turnaround section has been included in the Ring to Main Linac (RTML) section to allow bunch-by-bunch trajectory measurements immediately after the damping ring to be fed forward over a shorter path length to a Trajectory Correction section comprising two fast correctors/kickers per plane, separated by 90 degree phase advance.

Processing time is critical for this success of the correction system, with a turnaround section length of 170m, giving less than 0.5 microseconds to measure, process, and apply the kick angle correction.

Further details of the machine requirements of this system are described in the RTML section of the BCD.

Main Linac: Dispersion measurement and control

Mitigation of emittance growth in the Main Linac includes a scheme for measuring

v. Dec.12, 2005

Instrumentation and Controls 41

dispersion by locally changing beam energy for low current pilot bunches sent down the linac. This is known as Dispersion Free Steering. A variant of this algorithm is to create small dispersion "bumps" to cancel others.

Alternative methods of correcting unwanted dispersion involve changing corrector settings or moving quadrupoles, so do not involve beam-based feedback. If the dispersion can not be measured sufficiently accurately, we will have to optimize dipole corrector or quad settings based on emittance measurements. Thus, it not only the BPM data that we will need, but also the laser wire data.

Main Linac: Trajectory and Energy control

Several cascaded sections will provide position and energy control in the bunch compressor, main linac, and in the undulator for the positron source. As described above, the position control will use response matrix techniques to adjust the trajectory through each linac section based on bpm measurements. In addition, two bpms in each section will be used to measure beam energy and provide local feedback using klystron phase/amplitude control. As described in the NLC Zeroth Order Report, Chapter 7 [ref 4], cascaded loops for each section limit their correction to disturbances only immediately upstream of the section.

Beam Delivery System: Trajectory feedback from pulse to pulse

We plan a 5Hz orbit feedback system that may be cascaded with the linac 5Hz systems, and/or augmented with feed-forward information from upstream in the machine (i.e. from the linacs and/or the damping rings). In addition a 5Hz interaction-point (IP) feedback system will similarly be implemented. All of these systems will use similar hardware: BPMs, digital feedback processors and kickers. Corrections will be made to both x and y trajectories, as well as to x’ and y’.

Interaction Point: Trajectory feedback within bunch-train

Integrated simulations of the linac and BDS trajectory feedback systems show that, for noisy sites (eg. ground motion models ‘C’ and ‘K’), these systems recover only of order 20% of the nominal design luminosity. Via one-to-one steering, relative motion of the final quadrupoles on opposite sides of the IP leads to relative offsets of the electron and positron beams at the IP and the degradation in instantaneous (and hence integrated) luminosity. The problem is most serious in y where the beam is of order 5 nanometres in size.

For collision optimization, and luminosity stabilization, an intra-train (bunch-to-bunch) feedback system will be implemented in the interaction region. The BPM sensor will be placed several metres downstream of the IP to record the trajectory of the outgoing bunches, and the correcting kicker will be placed several metres upstream of the IP to correct the trajectory of the incoming bunches. Such a system can ‘lock in’ within the first 100 bunch crossings to achieve roughly 80% of luminosity attainable if the beams were in

v. Dec.12, 2005

Instrumentation and Controls 42

perfect collision. Duplicate systems, one for each beam, are planned to allow redundancy of control, and several BPMs at different locations would provide robustness with respect to potentially large electromagnetic backgrounds for the BPM sensors. Additional upstream BPM-kicker sets would be used to provide angle correction.

Additional benefit can be provided by an intra-train position/angle scan(s) based on optimization of a bunch-by-bunch luminosity signal from the ‘beam calorimeter’. After a further 1-2 hundred bunch crossings for the scan(s) it is possible to achieve a total of roughly 90% of the nominal luminosity attainable if the beams were in perfect collision.

Inputs to the feedbacks from additional diagnostics such as beam charge, transverse size, and bunch length monitors would allow adaptive gain control as collision conditions change. Inclusion of feed-forward information from the upstream trajectory feedbacks is desirable.

Architecture for 5Hz feedback systems

The relatively low correction rates and the distributed nature of many of the monitors and actuators make it appealing to consider using the integrated controls infrastructure for the 5Hz feedback systems. The controls network would be used to distribute monitor readbacks and actuator setpoints without requiring dedicated hardware and interfaces. The feedback algorithms themselves would be implemented in the Middleware layer of the control system, using dedicated processor units.

Implementing a feedback infrastructure into the integrated control system offers many advantages, such as:

Simpler implementation, since dedicated interfaces are not required for equipment involved in feedback loops.

Higher equipment reliability, since there are fewer components and interfaces. Greater flexibility, since all equipment would inherently be available for feedback

control, rather than limiting functionality to pre-defined equipment. Offers the ability to develop ad-hoc or un-anticipated feedback loops with the same

inherent functionality and tools. This could significantly enhance the commissioning process and operation of the ILC.

However, for this to be possible, the controls infrastructure must provide synchronous network activities and time-slicing of the network traffic, which in turn would require that all network attached devices comply with the time-slicing rules. The fact that the entire control system must in some way be synchronized to the 5Hz ILC pulse repetition rate means that to some degree, such requirements will be implicit in the control system implementation.

Details of the integrated control system architecture are provided in the control system section of the BCD.

An alternative, but less desirable solution, is to implement dedicated networks for all

v. Dec.12, 2005

Instrumentation and Controls 43

equipment associated with the feedback systems, perhaps using the same architecture and hierarchy as implemented for the integrated control system.

Additional functionality and versatility will be provided by integrating high level applications such as Matlab into the feedback system infrastructure, simplifying the task of developing and protyping feedback control algorithms that are then implemented on control system service-tier applications.

Architecture for intra-bunch feedback systems

Dedicated local systems will be required for intra-bunch feedback systems that must operate at the bunch rate of ~3MHz, such as the RTML turnaround trajectory feed-forward control, and intra-bunch trajectory control at the IP.

In addition, a fast synchronous infrastructure will allow implementation of delayed bunch-to-bunch feedback/feed-forward along the length of the linac.

Local input/output processors will acquire beam position, cavity fields, beam current, and other local beam parameters at the full 3MHz bunch rate, and distribute that information to a fast synchronous network that runs the length of the linac. Local interconnections with the low-level RF systems provide opportunities for local feedback loops at the full 3MHz bunch rate. Distributing the bunch-by-bunch information on a dedicated network allows dedicated external processors to perform dynamical feedback control within the 1mS ILC pulse.

Dedicated processing crates will provide both dedicated real-time bunch-to-bunch control, such as RF cavity fields, dispersion free steering, etc, while additional uncommitted crates will provide feedback systems to be implemented as required by operations and for physics studies. High level applications such as Matlab and Simulink will simplify the process of developing and prototyping such algorithms that are then deployed in the real-time processing crates.

Hardware Implementation

Most of the feedback processing requirements described in this section can be met using commercial hardware, including the 5Hz feedback loops and dynamic orbit control in the damping ring.

Custom hardware solutions will be required in cases where low latency or unique capabilities are required, such as for the RTML turnaround trajectory feed-forward and the IP intra-bunch trajectory feedback.

High Availability solutions will be implemented as appropriate, using the same standards and approach as for other instrumentation and control system equipment.

v. Dec.12, 2005

Instrumentation and Controls 44

Present State of the art

Present third-generation synchrotron light sources have refined orbit correction systems to the level likely required for the ILC damping ring. Ongoing advances in digital processor performance and fast high performance analog to digital conversion chips has allowed the conversion from the analog to digital domains to be performed much earlier in the signal chain. Most challenging are systematic effects in beam position monitoring when required resolutions are at or below the few micron level.

Fast intra-bunch trajectory control systems for the IP presently being developed under the FONT and FEATHER collaborations, with the latest implementation (“FONT-4”) aiming to demonstrate feedback with 100nS latency in the electronics and stabilization at um level.

Path to Specification

1. Identify stability and bandwidth issues for damping ring orbit correction.2. Identify beam position monitoring resolution and stability requirements.3. Identify trajectory stability requirements for the linac pre-IP, and for the beam

delivery system.4. Identify bunch compressor feedback loop requirements.5. Identify any additional beam-based control loops required from gun to IP.

Required R&D

1. The most difficult challenges surround the bunch-by-bunch trajectory feedback system in the beam delivery system. High precision measurement at 3MHz bunch rate, fast (low latency) processing, and fast kicker systems all require development.

2. Development of fast highly stable kicker magnets is required for the intra-bunch feedback in the beam delivery system and turnaround section of the RTML.

3. Depending on beam stability and measurement bandwidth requirements, substantial development may be required for beam position monitors and associated electronics. , so their experience can be drawn upon. unique requirements will likely drive specific development needs.

References:

1. P. Burrows, Beam Based Feedback Systems, Snowmass 2005.2. G. Decker, Beam Stability in Synchrotron Light Sources, DIPAC 20053. Beam-based Feedback: Theory and Implementation, NLC Zeroth Order Design

Report, Appendix D.4. Main Linacs: Design & Dynamics, NLC Zeroth Order Design Report, Chapter 7.

v. Dec.12, 2005

Instrumentation and Controls 45

Cost Model

A bottom-up system cost model will be produced following cost estimating rules of the GDE.

Alternate Configuration

No alternates under active consideration at this time.

v. Dec.12, 2005

Instrumentation and Controls 46

11.8 Introduction and Scope of System Integration

Definitions:

Beam Instrumentation: All Instruments in the ILC accelerator complex, which measure beam related parameters, i. e. intensity, position, emittance, etc. The core data and the control of the instrument have to be handled by the control system. Some instruments may also be part of a feedback system.

Control System: Remotely handles all systems and data in the ILC accelerator complex, including the beam instruments. We may define the accelerator-wide distributed control data bus, e. g. a standard Ethernet-based multi-gigabit serial network - as the distributed I/O - interface of the control system, which connects to all instrument modules.

In this subtopic we limit the scope of the integration into controls to beam instrumentation devices, i. e. beam monitors. At a later state one may/should extend the scope to general instrumentation and related topics, like LLRF and instrumentation of cryogenics, vacuum, magnet PS, etc. While the new developments and discussions on advanced ILC beam monitors mostly focuses on beam pickup physics and signal detection techniques, this paragraph tries to cover data acquisition (DAQ) and control aspects in a common way, applicable for most of the ILC beam instrumentation.

Concerning data acquisition aspects we can divide the ILC sub-accelerators into:

Linac: Injectors, Main Linacs and Beam Delivery System (BDS), including bunch compressors, spin rotators, etc., i.e. all straight sub-accelerators. Here the beam is present at the 200 ms repetition rate and has its nominal bunch-to-bunch spacing (~ 300 ns). Nearly all Linac beam instruments acquire their data on a bunch-by-bunch basis, identifying each bunch in the bunch-train.

Damping Rings (DR): Here bunch spacing of the beam is compressed to 6...20 ns. For most beam instruments (some exceptions) a turn-by-turn, or even slower data acquisition is sufficient, which relaxes DAQ and integration requirements. On the other hand a BPM-based, fast global orbit feedback system has to be established to minimize the beam emittance.

Major, high quantity ILC beam instrumentation systems are:

Beam Position Monitors (BPM): Data and control of about 2000 Linac and 800 DR BPM's has to be handled. The BPM's are distributed everywhere along the ILC accelerator complex. Also a fraction of the 20000 HOM-coupler signals might be equipped with read-out hardware. Some BPM's are dedicated for the machine protection system (MPS), others are part of feedback-loops.

Beam Intensity Monitors (Toroids, WCM): 50-100 toroids and wall current monitors are used to measure the bunch charges along the ILC accelerator complex.

Beam Loss Monitors (BLM): 100-500 ionization chamber BLM's are distributed to check and minimize beam losses.

Other beam instruments (beam phase, transverse and longitudinal beam/bunch profile) are of more complex and dedicated character. The number of these systems is smaller; some are combined within a special diagnostic section.

v. Dec.12, 2005

Instrumentation and Controls 47

The beam instrumentation integration efforts have to focus on the major components, i. e. BPM systems, without excluding the integration of new "exotic" instruments. This is particular important, as ILC beam monitors will be developed at different institutes, universities and industries all over the planet. The system integration has to be of open architecture and expandable to include other ILC subsystems, e. g. LLRF.

2. Example of the Linac BPM System Integration

Fig. 1 gives an example of the required hardware, which integrates a single Linac BPM into the control system. The 2-tunnel solution prefers minimum or no (radiation tolerant) hardware in the accelerator tunnel, so only the BPM pickup (RF BPM) is located here. Coaxial cables connect through links every 40 m to the electronics hardware in the support tunnel. This read-out and control system splits into two parts:

Dedicated Analog Hardware: An instrument specific signal processing interface.

Common Digital Hardware: A common, versatile digital DAQ and control platform.

Figure 1: Example of the Linac BPM Hardware

To acquire bunch-by-bunch data a timing module is required, which references to the

v. Dec.12, 2005

Instrumentation and Controls 48

phase stable 1.3 GHz RF signal. Signals derived by this timing board include marker signals (bunch # 1 -based on the time, fiducial of the timing distribution system -and ticks at every bunch), as well as ADC clock signals. The analog bunch signals are acquired with a high sample rate, e. g. rf synchronous at 1.3 GHz. Further digital signal processing provides an entry of physical data, including all required calibrations and normalizations for each bunch. Beside this core data processing, the integration platform has to handle instrument status data and control information. A smart I/O processor feeds the information on the control data bus for shipment to the control room, as well as to other systems if required. Within the 200 ms repetition rate all the acquired data of up to 8000 bunches has to be send through the control data bus. Some beam monitors may require a dedicated, faster feedback data bus, which has to send data within a fraction of the 1 ms macropulse duration.

The electro-mechanics (crates, racks, PS...) and interfaces have to be specified. Beside the upcoming high availability ATCA standard, also traditional standards (VME, VXI, PXI, etc.) and custom solutions have to be considered. While there is a high availability demand for systems like modulators, magnet power supplies, RF, etc., redundancy is a build-in feature into many beam instrumentation systems, like BPM's, which relaxes redundancy/hot-swap needs on the hardware side. However, it is highly desirable to have a uniform hardware-software platform and interface for instrumentation wherever possible throughout the ILC.

3. Present State-of-the-Art

3.1 Key Components

Key hardware components in the common digital part of Fig. 1 are fast analog-to-digital converters (ADC), field-programmable gate-arrays (FPGA) and digital signal processors (DSP). Table 1 lists current (End 2005) state-of-the-art ADC technology. Top-of-the-line FPGA's and DSP's have similar characteristics, or can parallel process ADC data. While the characteristics of these components will improve over the next years and prices will decrease - not every kind of beam instrumentation DAQ needs the maximum speed or band - width performance. Compatible ADC mezzanine boards within the DAQ section will handle the different demands.

Company Type Resolution (Bits)

Sampling Rate

Analog Bandwidth

National Semiconductor

ADC08D1500 2 x 8 3 GSPS 1.7 GHz

Atmel TS83102G0BMGS 10 2 GSPS 3.3 GHz

Anlaog Devices AD12500 12 500 MSPS ?

Texas Instruments ADS5500 14 125 MSPS 750 MHz

Linear Technology LTC2208 16 130 MSPS 700 MHz

v. Dec.12, 2005

Instrumentation and Controls 49

Table 1: State-of-the-Art ADC's, 2005

3.2 System Integration

Most of the current integration developments, regarding linear collider instrumentation, are done at the TESLA Test Facility (TTF) at DESY [1] (more information at http://tesla.desy.de/doocs/doocs.html). Their data acquisition system is based on an in-house developed VME 8-channel ADC board (14-bit, 10MSPS), capable of acquiring synchronous bunch-by-bunch data of the whole accelerator complex for 1 µs spaced bunches (only 1 sample per bunch). At the Spallation Neutron Source (SNS) multi-lab integration experience has been gained, not limited to beam instrumentation [2]. Here a commercial PC/PCI platform defines the hardware integration standard.

4. Path to Specification and Future R&D Requirements

At present the ILC system integration requirements are not fully specified, except we should try to be able to integrate all different kind of (beam) instrumentation:

1. Standard beam monitors (BPM, BLM, toroids, beam phase monitors)

2. Complex beam monitors systems (Laser wire scanners, longitudinal bunch profile monitors)

3. "Exotic" and experimental beam monitors from universities and collaborators

Many requirements, like bandwidth, hardware and software standards, etc. still have to be established. Technical aspects of the integration activities may change, dictated by new technology achievements in the communications industry.

Present and upcoming ILC Test Facilities (TTF, ATF, SMTF...) are the preferred environment to develop and test future R&D integration efforts. These test accelerators have requirements and demands similar to the ILC, but obviously on a smaller scale.

5. References

[1] A. Agababyan, et. al; Integrating a Fast Data Acquisition System into the DOOCS Control System; ICALEPCS2005, Geneva, Switzerland.

[2] D. P. Gurd, et. al; First Experience with Handover and Commissioning of the SNS Control System; ICALEPCS2003, Gyeongju, Korea.

v. Dec.12, 2005

Instrumentation and Controls 50

11.9 Beam Position Monitors

Main LinacBeam position monitors will be placed at each of the approximately 800 quadrupoles in the two main linacs. BPMs will be included in the croymodule. They shall report the position of each bunch in every bunch train.

Table 1. Main Linac Beam Position Monitor RequirementsParameter Requirement CommentsQuantity ~800 Every quadrupoleEnvironment Cold In cryomoduleAperture 60 mm – 70 mm BPMs not to be limiting apertureResolution 0.5 micronStability <10 microns Over cryomodule thermal cyclingTemporal resolution bunch-by-bunch

BPM pickups must be compatible with superconducting RF cryomodules. In particular during fabrication/assembly they must be cleanable with standard techniques to prevent contamination of the cryomodule.

Beam Position Sensor TechnologiesThere are two candidate solutions for the beam position pickup sensor. Both are based on resonant cavities. The “conventional cavity” BPM relies on dipole modes of a resonant cavity which is essentially a pillbox. The “re-entrant cavity” BPM relies on the dipole resonant mode of a coaxial resonator where the beam duct is the center conductor. Both candidate technologies require further R&D to be established as a proven solution to all of the system requirements.

Cavity BPMBy “cavity BPM” we mean essentially a pillbox-type cavity BPM with common-mode free coupling to external electronics. That is, the signal produced at the position output coupler is essentially that of the dipole mode and is therefore proportional to the product of beam charge and beam position. Cavity BPMs show extremely good resolution and stability. Examples are those designed at BINP and studied extensively at KEK’s ATF and those built by KEK’s Shintake and evaluated at SLAC’s FFTB. These feature dipole-mode couplers that reject the cavity monopole modes. This reduces the dynamic range required to achieve sub-micron resolution and is thought to yield excellent accuracy and stability. At C-band resolution in the 20 nm range or better have been reported for the Shintake and BINP BPMs. Centering stability better than ±50 nm over 2 hours has been observed. However we are not aware of a common-mode-free cavity BPM that has been qualified for use in a cryogenic clean environment.

v. Dec.12, 2005

Instrumentation and Controls 51

Re-entrant Cavity BPMThese are RF BPMs using coaxial resonant modes in a shorted coaxial structure. They have proven cryogenic and cleanroom compatibility as demonstrated in the TTF cryomodule. The signal at the output couplers have considerable common-mode signal. Careful monopole mode cancellation is required here; much of it is accomplished externally by RF hybrids in the processor electronics. A recent design should achieve 1 micron resolution. More R&D is required to demonstrate required resolution and stability.

Table 2. Main Linac BPM Pickup Options: Pros and ConsType Pros Cons

Reentrant Cavity

Proven cryo-compatibilityGood bunch-bunch resolution

Unproven sub-micron resolutionUnproven stability; depends on tuning in electronics module

Cavity

Proven resolutionExpect excellent stabilityGood bunch-bunch resolution

Unproven cryo-compatibility

Beam Delivery SystemBeam jitter to be kept less than 50% of the beam size in most of the BDS. To verify that the jitter requirement is met and to understand the sources of jitter if the requirement is not met, BDS BPMs must have resolution significantly better than half the beam size. We adopt a BPM single pulse resolution requirement of one-quarter of the beam size. Cavity BPMs are favored for most of the BPMs here for resolution, accuracy, and stability. The intra-train IP feedback BPMs are exceptions; these are likely to be stripline BPMs for ease of low-propagation delay processing, and for the 2 milliradian crossing angle scheme, the possibility of directional beam pickups.

Table 3. Beam Delivery System Position Monitor RequirementsParameter Requirement CommentsQuantity ~400Aperture Various sizesResolution /4 ~250 nm

Stability <10 microns long term< 1 micron / hour Energy Spectrometer only

Temporal resolution bunch-by-bunch many places, assume all

A potential problem arises from the beam halo generated in the collimators. The center of gravity of this halo does not necessarily coincidence with the position of the beam core. The BPMs will measure the position of the sum of beam core and halo position weighted by their relative intensities, while for optics tuning and feedbacks the core position is the relative quantity. Dedicated studies of these effects are required to estimate the magnitude and potential risks of this effect.

v. Dec.12, 2005

Instrumentation and Controls 52

Injector Systems Beam Position MonitorsThe beam position monitors in the electron and positron injectors have in comparison with the downstream systems relatively relaxed requirements. They are summarized in Table 4. Transverse beam dimensions are typically three orders of magnitude larger than in the downstream systems; these resolution requirements can easily be met with present BPM technology. A possible approach could be the use of stripline pickups for beamlines at ambient temperature and re-entrant or cavity pickups in the parts of the cold linac upstream of the damping rings (see main linac BPM description above). For the BPM’s in the positron capture system special care has to be taken due to large particle losses and integrated radiation doses. Appropriate solutions exist. Table 4. Injector BPM RequirementsParameter Requirements CommentsQuantity 600Environment Both ambient

temperature and cryogenic

Aperture diameter

40 mm – 100 mm

Resolution < 100 mPrecision < 100 mTime resolution Better than 300 ns Single bunch position has to be resolved for

a bunch spacing of 300 ns

Damping RingsEach of the three damping rings (one electron DR and two positron DRs) will be equipped with a number of BPMs corresponding to roughly four times the horizontal tune value. Most of the BPMs are needed for slow orbit measurement and control. The requirements for these BPMs are summarized in Table 5. Similar performances have been achieved in existing storage rings with button or stripline type pickups.

Table 5. Damping Ring Orbit BPM RequirementsParameter Requirements CommentsQuantity 900 3 rings with Qx70Environment Ambient

temperatureAperture diameter

16/40/98 mm Different apertures correspond to locations in wigglers, arcs and straight. Numbers taken from www-library.lbl.gov/docs/LBNL/570/45/PDF/LBNL-57045.pdf

Resolution 0.5 mPrecision < 100 mRoll error < 20 mradBandwidth > 100 kHz

A small number of BPMs with fast signal processing will be needed longitudinal and transverse feedback systems. Their requirements are summarized in Table 6. The BPM pickup can be of the same type as the slow BPMs, i.e. button or stripline, but a fast readout

v. Dec.12, 2005

Instrumentation and Controls 53

integral to feedback electronics is required. Similar performances have already been achieved in existing storage rings.

Table 6. Damping Ring Feedback BPM RequirementsParameter Requirements CommentsQuantity 9 3 BPMs for each of the 3 ringsEnvironment Ambient

temperatureAperture radius 20 mm We assume that at least one of the feedback

BPMs has to be located in the arcs for longitudinal feedback

Resolution 0.5 mPrecision < 100 mTime resolution Better than 6ns Single bunch position has to be resolved for a

bunch spacing of 6 ns

R&D Required Develop Cryo-compatible cavity BPM, prove cleanability Prove stability and resolution capability of re-entrant cavity BPM Develop and prototype cavities, evaluate in cryomodules Develop and prototype electronics Study effects of beam halos/beam tails on interpretation of beam position data, set

requirements on halo/tail generation.

v. Dec.12, 2005

Instrumentation and Controls 54

11.8 Beam Profile Monitor System

1. Overview The ILC beam profile monitor system is the collection of devices to be used to measure the transverse size of the beam throughout the complex. In the damped beam section (damping rings, bunch compressor, main linac and beam delivery) these are based on ‘laserwires’. A laserwire uses a finely focused, very high power laser to sample the particle beam density in the same way that a wire scanner does. However, the operational issues surrounding a laserwire are more complex because of laser control, laser focusing, laser profile, pulsed laser timing and etc. These have been studied and are described in [1]. The laserwire has the added advantage that there is no added material inside the vacuum chamber. This reduces the risk of contaminating the nearby cryo-cavities. The un-damped beam section of ILC, (source, injector linac, damping ring injection), will use wire scanners and related conventional profile monitors.

2. Baseline Configuration

a. Description In the baseline configuration, there are laserwires in 1) the damping rings, 2) the ring to BC transport, 3) between the two BC stages, 4) the BC to main linac transport, 5) within the main linac (3 sets; at the 10%, 25% and 50% energy gain locations), 6) at the entrance to the beam delivery and 7) within the beam delivery, downstream of the collimation systems. The total number of laserwire systems per side is thus 9. With the exception of the damping ring laserwire, each laserwire system has 3 to 5 interaction chambers distributed along a fraction of a betatron cycle, extending perhaps 40 meters, depending on the beam optics. Each interaction chamber has a focal system for x and another for y and may have a ‘u’ scan direction also, for monitoring x y coupling. The damping ring only has two interaction chambers, one for a dispersion - free and one for a non-zero dispersion region. The main linac laserwires use a modified inter-module insert that allows ‘warm access’ to the vacuum chamber. [2] The total number of interaction chambers, for both sides of the ILC is 70. It is important to note that a single laser (a cost driver in the system) may feed many IP’s through the use of an extended laser transport system. The number of lasers for the ILC laserwire system should be 12 or less (total for both sides). The performance requirements for the laserwire profile monitors are summarized in section 10.0, Table 1. Typically, the system must provide a 10% measurement of emittance (5% beam size). The system in the damping ring and the system at the entrance to the beam delivery should be 2x better than that. A laserwire consists of four main subsystems: 1) the laser and its control and timing system, 2) a laser transport system that carries the light from outside the accelerator enclosure to 3) the Compton interaction chamber, including its strong focusing system and 4) the scattered radiation detector. Typical subsystem parameters are listed in the references.

b. Supporting Documentation [1] Preliminary analysis of laserwire error (systematic and statistical) budget:

v. Dec.12, 2005

Instrumentation and Controls 55

Blair note of November 1, 2005. [2] Answer to Snowmass Question number 29; ‘Linac Diagnostic Sections’, Marc Ross, September 2005. http://www.linearcollider.org/files/WGGG/DECISION%2029.doc Design and performance of existing laserwires; an outline of laserwire RD: Snowmass reports from Grahame Blair and Yosuke Honda. http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/GG2/grahame_blair20050818043011.pptYosuke Honda, et.al., Nucl.Instrum.Meth.A538:100-115,2005 Design of NLC laserwires: J. Frisch, 2002. http://www-project.slac.stanford.edu/lc/local/systems/special_projects/Nanobeams2002/laserwire.pdf

c. Required Research and Development RD is required to show: 1) high resolution performance with high power lasers. In order to make precise measurements of sub-micron beams, very high power (>100MW peak) lasers with ultra-violet light must be used. 2) effective use of the 01 mode, which effectively halves the minimum beam size that can be measured 3) implementation of precision scattered particle detectors in complex accelerator systems, such as the BDS collimation region and the main linac. 4) component durability in this environment. 5) effective use of fringe-laser based monitor: ‘Shintake’ monitor.

3. Alternative lower cost Configuration

a. Description

v. Dec.12, 2005

Instrumentation and Controls 56

11.10 Longitudinal Diagnostic System

1. Overview The purpose of the longitudinal diagnostics is 1) qualify the beam at the extraction of the damping ring, 2) determine the performance of the bunch compressor system and 3) measure correlations between directions in phase space (x z, y z, E z). The required resolution is set by the feature size in the bunch longitudinal distribution, which is typically not Gaussian. The longitudinal beam size monitors will also be used to measure the energy spread. The resolution should be a fraction (20%) of the nominal energy spread. This will typically be done with wire scanners or laserwires at regions of high dispersion. For simplicity these are known as energy spread wires. Energy spread wire locations

1) Source capture beam 2) Damping ring entrance 3) Damping ring exit 4) Linac entrance 5) Linac exit (and undulator entrance if different)

2. Baseline Configuration

a. Description There will be 4 longitudinal diagnostic systems in the damped beam regions (per side),

1) at the exit of the damping ring 2) at the exit of the bunch compressor 3) at the exit of the main linac 4) near the IP. This can be integrated with the crab system and should be focused on

measurement of the y z correlation ‘banana effect’.

Typical performance requirements are to resolve structure which is a fraction of the bunch length, about 20% or 30 microns. In the injector system, there is a similar set of requirements to monitor the phase space of the beam to be injected in the damping ring, the damping process and the capture process.

1) at the exit of the sub-harmonic buncher 2) at the entrance to the damping ring 3) in the damping rings themselves 4) after the capture RF.

Typical performance requirements is to have resolution of about 150 microns. Excellent demonstrations of the use of two dimensional deflecting structures have recently been completed. These facilitate buncher and related low-velocity beam tuning. The above systems, as with the transverse systems, work best if properly integrated into the surrounding lattice. This may have cost impact.

v. Dec.12, 2005

Instrumentation and Controls 57

There are several longitudinal diagnostic systems under development for use at FEL’s.

1) high power RF deflecting structures (LOLA, or crab structures) [1] 2) electro-optical sampling [2] 3) coherent radiation (synchrotron, diffraction, Smith-Purcell) [3] 4) ultra-fast laser wire [4]

The BCD should have 1), high power RF deflecting structures, in each of the above listed locations. The RF deflecting structure provides a ‘z-dependent’ kick that allows the longitudinal structure of the beam to be imaged. The direction perpendicular to the kick shows the correlation. This device is the only one of the four techniques listed which has demonstrated excellent correlation measurements. Since the correlation information is 2 dimensional, they work best with video imaging, and not with sampling (laserwire) like systems. RD is needed to make sure that reliable, durable video imaging can work with low emittance, high energy ILC beams. In the BCD, the deflecting structure is a warm copper structure capable only of deflecting a single bunch, but with very high gradient. The BCD structure is either 2600 or 3900 MHz.

b. Supporting Documentation Paper on LOLA systems at SLAC and TTF: http://www.slac.stanford.edu/cgi-wrap/getdoc/slac-pub-9241.pdfhttp://docdb.fnal.gov/ILC/DocDB/0001/000108/001/lolafel2005_version05081701.pdf Paper on use of a two dimensional LOLA structure by Jake Haimson of Haimson Research Corp. ‘Longitudinal Phase Space Measurements of Short Electron Bunches Using a 17 GHz Circularly Polarized Beam Deflector’, Advanced Accelerator Concepts Workshop, AIP Conference Proceedings 737, 2004. Very high resolution OTR studies done at ATF, J. Frisch et al., http://atfweb.kek.jp/atf/Reports/ATF-01-08.pdf

c. Required Research and Development RD is required to show: 1) performance of very high resolution two dimensional imagers for use with micron-sized damped beams. Present state of the art demonstrations have shown two micron minimum feature resolution. 2) operation of warm structure LOLA systems in the low emittance transport system 3) structures with the capability of switching between x and y deflections, or structure pairs with both kicks – 90 degrees out of phase 4) integrated performance of fast non-invasive bunch length monitors – calibrated using the LOLA type devices.

3. Alternative lower cost Configuration

a. Description

v. Dec.12, 2005

GDE White Papers 1

12. GDE White Papers

Five white papers were written by GDE members on critical BCD issues .

12.1 Energy Upgrade Scenario 1. Task Group Charge This task group is charged by the GDE Executive Committee to review the recommendations from the Snowmass working groups on the energy upgrade path. Specifically, the Task Group is instructed to -

Comment on the pros and cons of the baseline choice made, and on the options not selected. Identify and describe any other possible options that should be considered. Provide a recommendation to the GDE on whether the baseline option selected at Snowmass should be chosen for the ILC BCD, and, if so, why. If the option selected at Snowmass should not be chosen, provide a recommendation for the ILC BCD, together with a justification.

2. Options

At the Snowmass workshop the WG5 reviewed [1,2] the following three options, and recommended to adopt the Option 1 for BCD:

• Option 1: In Phase-1 (500GeV), 22km of cryomodules operating at 31.5MV/m will be installed in the first part of the 41+km tunnels. The upgrade to Phase-2 (1TeV) will require 19km of additional cryomodules, operating at 36MV/m, additional RF and refrigeration. This is the BCD recommendation from Snowmass 2005.

• Option 2: In Phase-1, 24.4km of cryomodules will be installed in the first part of the 41+km tunnels. The RF power sources and the cryogenic systems to install in Phase-1 will have the same total power capacity as Phase-1 of Option 1. Thus, the Phase-1 Option 2 linacs can accelerate full current beams for ECM=500GeV but at a reduced initial accelerating gradient of 28MV/m. Because of the increased length of installed linac cavities compared to Option 1, the Phase-1 of Option 2 can reach ECM=560GeV with the gradient of 31.5MV/m at the cost of a reduced current and reduced luminosity. The upgrade to Phase-2 will require 16.6km of additional cryomodules, operating at 36MV/m, additional RF and refrigeration.

• Option 3: In Phase-1, 22km of cryomodules operating at 31.5MV will be installed in the 22km (short) tunnels. The upgrade to Phase-2 will require construction of the additional 19km-long tunnels, additional cryomodules operating at 36MV/m, additional RF and refrigeration.

In all these three cases, the TESLA shape cavities for 31.5MV/m operation or Low-Loss/Reentrant shape cavities for 36MV/m operation are assumed with the Q value of 1010. The “packing factor” of the active volume along the linacs is 72.5%. The WG5 in the 2nd week of Snowmass also discussed a more optimistic (somewhat more aggressive) gradient scenario in which approximately 10% of less-performing cavities/cryomodules

v. Dec.12, 2005

GDE White Papers 2

during the initial construction will be replaced with better performing units by/during the Phase-2 upgrade. This may allow us to reduce the total tunnel length to 38.5km from 41+km. While the written WG5 report does not explicitly review, another Option was also mentioned during the Snowmass workshop [3]:

• Option 4: In Phase-1 (500GeV), ~25km of cryomodules operating at 31.5MV/m will be installed, but in a “sparcified” fashion distributed in, for instance, four groups over the 41+km tunnels. The spaces, each extending up to 2km, between the neighboring cryo-groups will be utilized as diagnostic sections during Phase-1. The upgrade to Phase-2 (1TeV) will require additional cryomodules, operating at 36MV/m, to replace these diagnostic beam lines, additional RF and refrigeration.

While other upgrade options are also conceivable, all of them are considered a hybrid of these four options with some parameter tweaking. 3. Pros and Cons

We fix the energy goal to be 500GeV for Phase-1 and 1TeV for Phase-2, as per the ICFA Linear Collider Parameter Subcommittee report [4]. The emittance control issues and the design requirements derived from them will apply equally to all the four options above, and their considerations are unlikely to favor/disfavor one Option over others. Thus, the Task Group gave pros-vs-cons considerations in the light of the following four technical merits:

• Ease of construction and operability during Phases 1 and 2, • Energy reach of the Phase-I machine using lower luminosity, • Cost • Prospects for high accelerating gradients.

Construction and Operability during Phases 1 and 2 The relevant issues are summarized in Table 1. In the entries, the favorable features are printed in blue, and problematic ones in red. Some comments follow-

• WG5 described Option 2 as being conservative in the sense that it allocates 20% energy overhead compared to 10% of Option 1. We note that the importance of this “conservatism” depends on the prospects for the gradient improvement of 28MV/m 31.5MV/m (or higher). The assessment is expected to vary as the time develops.

• In all options, the linac configuration will end up consisting of a hybrid of cryomodules operated at 31.5MV/m and 36MV/m. However, since the Phase-1 installation of RF source units will be designed to be capable of driving 35MV/m (http://www.linearcollider.org/wiki/doku.php?id=bcd:main_linac:configuration) systems anyways, this is a non-issue. The varying gradients are unlikely to affect the linac operability in fundamental ways, either. Note that the present Linac BCD does not specify RF sources capable of powering cavities beyond 36 MV/m at nominal beam currents.

v. Dec.12, 2005

GDE White Papers 3

• If the new cavities added for Phase-2 are to be operated at accelerating gradients substantially higher than 36MV/m, a new, distinct set of RF systems need to be introduced with matching designs. All four Options can accommodate these. However, Options 1, 2 and 3 offer a cleaner installation pattern for maintenance services and commissioning since the Phase-2 hardware will be more distinctly populated (One might counter-argue that 2+km-long clusters of Phase-2 hardware are sufficiently distinct populations).

• Reports from WG1 point out the need for beam diagnostic sections before the bunch compressor (BC), after the BC, and at the 250GeV point along the main linacs. In addition, a 50GeV beam extraction point may be introduced for the positron system. All four options can accommodate these. A generous set of “diagnostic sections” provided by Option 4 naturally satisfy these.

• A significant discriminator from the construction and operability standpoint is the civil engineering and associated installation work that is required in the Phase-2 upgrade for Option 3, where extension tunnels have to be newly excavated.

Option

1Option 2 Option

3 Option 4

Can the RF sources and cryogenics for Phase-2 be installed during Phase-1 operation?

Yes Yes No Yes

During Phase-1, is significant flexibility still reserved for the details of the Phase-2 RF and cryosystems?

Yes Yes Yes Yes, but mildly limited, because of the fixed geometry.

Setup of machine in initial phase (machine protection, quench handling etc)

- Easier than Options 1, 3, 4

- -

CMS Energy with 20% margin on accelerating gradient (i.e. 28 MV/m)

450 GeV

500 GeV 450 GeV

450 GeV

Do impacts of Phase-2 tunnel excavation during Phase-1 operation need to be evaluated?

No No Yes No

Is recommissioning needed for e- sources (and likely DRs) in Phase-2?

No No Yes No

v. Dec.12, 2005

GDE White Papers 4

Can the benefits of diagnostic section be accommodated?

Yes Yes Yes Yes, also built-in rooms during Phase-1.

Any issues with 31-36MV/m hybrid operation?

OK OK OK OK

Any issues with ~40MV/m hybrid operation?

OK OK OK OK

Table 1: Pro/Con Table from the Operability Standpoint Energy reach of the Phase-I machine using lower luminosity For all the scenarios it seems reasonable to assume that the experience after initial running will allow operation with reduced margin and increased operating gradients in the cavities towards 35MV/m (cavity performance test acceptance criteria). A higher energy can be reached for all options. The lower gradient due to the larger operations margin of Option 2 allows a final energy limit of about 560 GeV at 31.5 MV/m average gradient. This potential can be achieved without additional installations. Its cost implications are discussed in the next section. Cost The WG5 at Snowmass examined the construction cost (not including operation) of the main linac systems for the Phase-1 and the Phase-2 upgrade. The required expenses are computed for construction and installation of: the cavities, cryostats, refrigerators, modulators, RF sources, distribution systems, instrumentation, control and tunnel civil engineering. Then, only the cost ratios normalized to that of the Phase-1 construction in case of Option 3 were quoted. The cost model used there was derived from the TESLA TDR studies with some revisions, such that the conventional facilities are assumed to be based on two parallel tunnels. No cost variations are considered for the cavities, RF systems and cryogenic systems during Phases 1 and 2. It should be also noted that this exercise does not consider the cost for the injectors, damping rings, beam delivery and experiment halls. The first half of Table 2 reproduces this WG5 Snowmass summary. To see the stability of cost ratios, we varied the costs for “cavities and cryostats” (i.e. not including RF sources) and for “tunnel civil construction”, each independently, by up to 40%. The resultant total linac cost is increased by up to ~24%. However, the cost ratios, normalized to that of the Phase-1 of Option 3 in each case, remained very stable. They are summarized in the second half of Table 2.

Option 1 Option 2 Option 3

Option 4

v. Dec.12, 2005

GDE White Papers 5

Cost Ratio (reported at Snowmass 2005)

- Phase-1 Construction - Phase-2 Upgrade - Phase-1 & 2 Total Construction

1.15 0.77 1.93

1.21 0.73 1.94

1.00 0.93 1.93

- - -

Cost Ratio (when linac costs for linac cavity systems, including the cryostats, and tunnels are independently varied by up to 40%) - Phase-1 Construction - Phase-2 Upgrade - Phase-1 & 2 Total Construction

1.13~1.200.72~0.79~1.92

1.19~1.250.68~0.73~1.93

1.0 ~0.92 ~1.92

- - -

Table 2: Summary of the linac construction cost ratios. Some comments follow:

• The total cost for the linac components required in Phase-1 & 2 is likely to be very similar.

o However, it should be noted that the Phase-1 Phase-2 upgrade in Option 3 involves relocation of the electron source. The damping rings may also have to be relocated.

o The Phase-1 Phase-2 upgrade in Option 2 will involve rearrangement of the waveguides, klystrons, and possible modulators.

o None of these costs are accounted for in Table 2. • If the project is cut at the end of Phase-1, Option 3 is the least expensive by

approximately 15%. If the Phases-1 and 2 are put together, Option 3 is the most expensive by due to the additional injector relocation which is not accounted for in this exercise.

• The above calculation neglects the operational cost. The setup of the machine for Phase-1 with a larger operations margin as in Option 2 will ease commissioning thus saving the commissioning / machine improvement time and the cost associated with such efforts.

• Phase-1 of Option 2 assumes installation of 12% more cavities compared to Option 1, yet it assumes the same amount of RF sources as Phase-1 of Option 1. This allows the operation at ECM=500GeV with full current at 28MV/m, or ECM=560GeV with reduced current at 31.5MV/m, depending on the cavity operation margin to allocate. However, in order to reach 1TeV in Phase-2, the cavity sections installed in Phase-1 need to operate with full current at 31.5MV/m. This will require rearrangement of the waveguide connections between the RF power sources and the cavities, unless the klystrons (and possibly matching DC power supplies) are upgraded (see above).

• In Phase-1 Option 1, we may choose to install 560GeV-worth of linac systems also, in an attempt of attaining additional ~10% operational margin. Naturally, this

v. Dec.12, 2005

GDE White Papers 6

comes at an expense of increased construction cost for Phase-1. • Overall, whichever specific strategies to choose, in terms of installed population of

hardware components, operational margins in Phase-1 can be attained for corresponding construction cost increases.

Prospects for High Accelerating Gradients We have note that the assumed accelerating gradient of 31.5MV/m for Phase-1 and 36MV/m for Phase-2 are a major challenge in today’s (2005) technical standard. This is true for any of the options discussed here. The cavity performance for all the options is specified as:

• Phase-I: Average accelerating gradient in low-power acceptance test 37MV/m. The width of the gradient distribution is specified to be 5%. All cavities which perform less than 35MV/m with Q0=8×109 are rejected and need re-treatment.

• Phase-II: Average accelerating gradient in low-power acceptance test 42MV/m. The width of the gradient distribution is specified to be 5%. All cavities which perform less than 40MV/m with Q0=8×109 are rejected and need re-treatment.

Vigorous, coherent and organized international collaboration programs are mandatory to establish especially the quality control measures which reduce the performance scatter currently observed. A detailed list of R&D topics has been established in the WG5 Snowmass reports, which needs to be addressed [5,6].

4. Recommendations

We recommend Option 1, as put forward in the WG5 Snowmass report, as the BCD on the basis of the following:

• Option 1 offers good operability during Phase-1, adequate provision for beam diagnostic capabilities and ability to accommodate upgraded SRF hardware components relatively seamlessly.

• Option 1, being conceptually the simplest among the schemes considered here, helps GDE develop the solid understanding of the practical fundamentals of the engineering designs and the cost analyses the most rapidly.

• Much of the understanding on the engineering and the cost, to be gained from the exercise with Option 1, can be readily applied to examine the technical and cost implications of other Options soon thereafter, if deemed adequate.

We also note the following:

• Option 3 offers the lowest cost for Phase-1 yet it requires the highest cost for the whole Phases 1&2, because of the staged civil construction and relocation of the injector systems associated with it. The Phase-2 upgrade for Option 3 is likely to take the longest time period, because of, again, the staged civil construction and the fact that the installation of RF source components cannot start till the extension tunnels are complete. The relative merit of Option 3 will have to be looked at in the context of the project acceptance from the political or long-term financial standpoint. Such analysis can be done after the complete Option 1 study is done.

• Option 4 offers the operability and upgradeability similar to those of Option 1, plus

v. Dec.12, 2005

GDE White Papers 7

substantially more diagnostic sections during Phase-1. At this point, however, the task group does not see convincing technical justification for this option..

• At the risk of being repetitive, the assumed accelerating gradient of 31.5MV/m for Phase-1 and 36MV/m for Phase-2 are a major challenge A detailed list of R&D topics to be addressed are given in [5,6].

References [1] H.Padamsee, et al; “Update on Q2 – Main linac starting gradient, upgrade gradient and upgrade path”, http://lcdev.kek.jp/GDE/EUpgradeTF/ForPlenary1c.ppt[2] H.Padamsee, et al; “Main Linac Starting Gradient, Upgrade Gradient and Upgrade Patch”, http://lcdev.kek.jp/GDE/EUpgradeTF/Question3-1.doc[3] G.Dugan, private communication, http://lcdev.kek.jp/GDE/EUpgradeTF/redis.pdf [4] S.Komamiya, et al; “Parameters for the Linear Collider”, http://www.fnal.gov/directorate/icfa/LC_parameters.pdf , Sep. 30, 2003. [5] Snowmass WG5 Summaries; http://www.linearcollider.org/cms/?pid=1000088 [6] BCD Cavity Package; http://www.linearcollider.org/wiki/doku.php?id=bcd:main_linac:cavity_package

v. Dec.12, 2005

GDE White Papers 8

12.2 The IP Configuration to be Chosen for the ILC Baseline

Configuration Document Wilhelm Bialowons (DESY), Thomas Markiewicz (SLAC), Hitoshi Yamamoto (Tohoku U.) 16 November 2005

Working Group 4 Recommendations:

• BCD: Two BDS (2 mrad & 20mrad), Two IR halls, separated in z, Two detectors • ACD1: Two BDS (2 mrad & 20mrad), One IR hall, z=0, Two detectors • ACD2: One BDS (X-angle unspecified), One IR hall, z=0, Two detectors, push-pull

capability Other Configurations

• Minimal: One BDS, One IR Hall, One detector designed so as to permit construction of a second BDS, IR Hall and detector at a later date.

Recommendation of this Committee • We recommend that the GDE adopt the WG4 Snowmass BCD configuration of two

beamlines, IPs & detectors in two IR Halls at different longitudinal positions. The baseline crossing angles should be 2 mrad and 20 mrad.

Justification of Recommendation • There is not yet an international consensus on the optimal parameters of an IP

configuration based on a single IR hall. Partisans for the different IP options have passionate opinions which can only be swayed by the results from detailed studies. A down-select now would require the authority of an internationally agreed upon ITRP process and cannot be made by a GDE task force.

• In order to make a fair and unbiased comparison, we must at this point maintain a

configuration where both crossing angle options continue to be developed and studied. Work is needed to study machine and physics performance issues equally for both designs, demonstrate hardware feasibility through R&D, and develop accurate and mutually agreed upon cost estimates. Choosing the WG4 BCD allows the concerns of all parts of the physics community to be satisfied while these tasks are completed.

• The main argument for overturning the Snowmass BCD decision is lower cost. At this

time, the cost estimates from different regions vary widely and are not agreed upon. The time to make the down-select to one IR is when the potential savings and their impact are better understood.

v. Dec.12, 2005

GDE White Papers 9

Compatibility with Gamma-Gamma Option The BCD covers small (0-2mrad) and large (14-20mrad) cases; the 20mrad and possibly 14mrad numbers are sufficiently large to be said to be consistent at this stage of R&D with the needs of a γγ upgrade of the IP. A crossing angle below 14 mrad is incompatible with γγ.

Costs If cost sensitivity requires an initial BCD of one IR, that IR hall should be sized to house one detector. We believe it is imperative not to make the crossing angle choice at this time for the reasons given above, but rather to carry forward the studies, R&D and cost estimate in preparation for a later decision. Details on the staging strategy to either a 2nd IR hall or a Push-Pull configuration in an expanded IR hall do not need to be made at this time. However, longitudinal space should be left for a low emittance big bend to a possible future second IR hall and the civil engineering plan should incorporate schemes to facilitate this, for example tunnel stubs or alcoves. Cost information is available from the TESLA TDR (TESLA), the US Linear Collider Technical Options Study (USLCTOS) and the GLC Project Report (GLC200302), whose total project cost estimates vary significantly. We do not attempt to resolve discrepancies, nor do we include the cost of detectors. These cost estimates were in 2000 €, 2003 $, and 2003 ¥, respectively, and need to be updated and reconciled by the GDE cost engineers in 2006. While the total estimates of the absolute cost of the 2nd IP differ by only ~10%, the individual components can differ by factors of two or three. The TESLA estimate was 250M Euros for a 2nd IR including beam lines, tunnels, IR halls and dumps. Assuming these savings were used to purchase additional main linac cryomodules and rf, this would be sufficient to increase the energy of the machine by 19% (with an existing tunnel) and a few % less if additional tunnel needed to be constructed. This assumes that the unit costs are those of the last units produced for the initial 500 GeV machine (80% of their average cost). This also corresponds to the TDR estimate of two years of operational costs. The corresponding estimate from the USLCTOS was 229M$ for a 2nd IR. This included a parallel support tunnel as well as the items listed above. Calculated as above, using USLCTOS cost estimates for cryomodules and rf, this would be sufficient to increase the energy by 14% (with an existing tunnel) and a few % less with tunnel. The corresponding estimate from the GDC200302 was 303 Okuyen1 for a 2nd IR. This is quite consistent with the TESLA and USLCTOS estimates. Calculated as above, and using its own cost estimates for cryomodules and rf, this corresponds to a 17% increase in energy with an existing tunnel and a few % less with tunnel. 1 1 Okuyen = 108 ¥ ; exchange rates on 2005.11.16 are 1 € = 139.5 ¥ = 1.169 $ or 1 $ = 119.3 ¥

v. Dec.12, 2005

GDE White Papers 10

Pros And Cons of BCD vs ACD1 or ACD2 • In our present understanding, the costs of BCD and ACD1 are similar. Cost estimates

were made based on linear feet of tunnel required and volume of the IR halls to be excavated. The difference between BCD and ACD1 is only in the Z offset of the detectors. This affects the staging strategy in the event that the initial project has a single IR, and can be decided when the decision to build a second IR is made.

• For ACD2, the cost increment, relative to a one IR sized to hold one detector, of a hall large enough to house two detectors is estimated at 30M Euros (TESLA), 58M$ (USLCTOS), or 78 Okuyen (GLC200302). This is relevant for ACD2 with a Push-pull configuration where the larger hall is built as part of the initial project. The primary cost impact is the 2nd detector whose costs have not been included.

Discussion of Down-Select Options • The current BCD prepared by WG4 lists the pros and cons of the different crossing

angles, as they are currently understood, and a list of the R&D required to pursue each configuration. The relevant Webpages are:

o Ranking of BDS configurations o R&D Specific to Baseline and Alternatives

• If a down-select to one IR is made, the choice of linac angle (not in the charge to this committee) might need to be reconsidered by the GDE.

v. Dec.12, 2005

GDE White Papers 11

12.3 Number of Tunnels

Dedicated task group: J.P.Delahaye, H.Hayano, N.Phinney.

1. Introduction

A number of options concerning the number of tunnels and their respective configuration were identified at Snowmass in the context of Global Group 1 on Parameters. They are summarized on the figure 1 below.

Two options, each one with two alternatives, are envisageable:

1. One single tunnel that contains the accelerator as well as all the electronics and power sources with two alternatives:

a. Modulators in widely spaced support buildings on surface as specified in the TESLA project (case 2 of fig 1).

b. Modulators inside the same tunnel (case 3 of fig 1)

2. Two tunnels with the linac in one tunnel and the electronics and power sources in a second support tunnel with two alternatives.

a. Modulators and Klystrons in a gallery on or close to the surface (case 1 of fig1). b. Modulators in the second nearby tunnel and Klystrons in one or the other tunnel

(case 4 of fig 1).

v. Dec.12, 2005

GDE White Papers 12

2. Present recommendations

The Global Group 1 in charge of “Parameters” at Snowmass and answering the decision 6 about the number of tunnels finally recommended the two tunnels configuration with the option 2a in case of a shallow site and option 2b in case of a deep site.

The present draft chapter of the BCD about the tunnel layout is in line with the Snowmass recommendation. It reads: The baseline choice is for the rf sources to be located outside of the beam tunnel so they would not be subject to radiation and could be accessed for repairs while the machine is running. To minimize rf power losses and cable runs, the sources are to be distributed along a second tunnel (or surface gallery) that runs parallel and nearby to the beamline tunnel. The rf power is transported into the beamline tunnel through three WR650 waveguide runs from each rf unit (one waveguide per cryomodule).

3. Pros and cons of the Snowmass recommendation

The pros and cons of the recommendation in favor of a double tunnel with respect to a single tunnel are summarized in the appendix. The single tunnel corresponds to the option 1a selected by the TESLA study with a diameter of 5.2 m and with access shafts and surface halls at regular intervals (every 5km) for cryogenics and modulators housing. The double tunnel corresponds to the option 2a or 2b (site dependent) made by one support tunnel of 4m diameter and a beam tunnel of 3.2 m diameter with the minimum of access shafts required for cryogenics (every 5km) and cross-over connection from the two nearby tunnels at regular (~600m to 1 km) intervals. The major criteria are related to maintenance/reliability, safety and cost issues: Cost favors a single tunnel. The cost estimation of the civil engineering varies substantially from one site & region to another. The exact diameter of the support tunnel will have to be adjusted to the size of the selected modulator with a slight influence on the extra cost of the two tunnel configuration. All other considerations favor two tunnels. In particular, the greater reliability demanded of many components and/or the extra tunnel length to provide additional rf overhead and the possible extra access shafts to comply with safety issues offset much of the cost advantage. The total extra cost is finally of the order of a few % of the TPC.

4. Site specificity

The issues can be site-specific because of different safety and access/egress regulations.

v. Dec.12, 2005

GDE White Papers 13

DESY has stated that access would not bepermitted to the support tunnel while high power components were operating but there do not seem to be similar concerns in Japan,the US, or CERN.A shallow tunnel would reduce the tunnel construction costs and hence the cost difference between one and two tunnels. However,none of the regional sample sites proposed in the BCD really qualify as a shallow tunnel and none of them could tolerate the surfacepresence of a SLAC-like klystron gallery. The only exception is the Hanford reference site where a surface gallery would be possible.The DESY site is less deep than FNAL, CERN or the Japan sites but cannot really be thought of as shallow.

5. Other possible options to be considered

All possible options are summarized on figure 1. The only option not being considered here is the case 1b with one single tunnel housing all components including the modulators. This option is worth consideration only if a compact high-reliability redundant modulator like the Marx design were adopted. With the TESLA style modulator, this option is excluded because of reliability and accessibility issues. Since the TESLA style modulator is the baseline design and the Marx modulator is only an alternate, we agree with this analysis. This option could be reconsidered only if the baseline modulator choice changes.

v. Dec.12, 2005

GDE White Papers 14

6. Our recommendation and justifications

We concur with the recommendation of Global Group 1 at Snowmass and recommend the two tunnel option namely:

← • the option 2a in case of a shallow site ← • the option 2b in case of a deep site

with the justifications below: ← • The additional cost is marginal when considering the necessary overhead

and equipment improvements to comply with reliability and safety issues, ← • A better availability with higher risk of success on the necessary MTBF

improvements of the critical components ← • Simpler installation and likely shorter schedule ← • Easier maintenance and consolidation of equipments ← • Smaller exposure of equipments to radiation and corresponding damage ← • Easy access to key electronics for fine tuning during commissioning ← • Easier energy upgrade

Pros and cons of a double tunnel with respect to a single tunnel

(assuming modulators in surface halls with single tunnel as in the TESLA Design Report)

Criterium Pros Cons Favored tunnels

Availability & Risks

Due to the possible access of personnel in the support tunnel even during high power and beam operation, the down-time is estimated to be half as much with 2 tunnels, (30%) with today’s quality of components. The down-time could be reduced to 17% by improving the MTBF of the critical components as compared to 30% in a single tunnel with the same improvements or 22% with robotic repair (which would be extremely expensive).

2

Commissionin Easy access during operation 2

v. Dec.12, 2005

GDE White Papers 15

g

to subtle electronics problems that require hands on with scope and beam to understand.

Radiation

Reduced exposure of electronics to radiation from accelerator. No radiation hard design required.

2

Equipment consolidation

Possible improvements to electronics modules such as BPMs and LLRF can be done gradually a few modules at a time during the runs

2

Upgrade

Installation in support tunnel can go on while commissioning/running occurs in accelerator tunnel.

2

Safety

Possible escape from one tunnel to another in case of emergency due to fire or other hazard using connecting tunnels equipped with isolation firewall as cross-over egress at regular intervals (every 600 m to 1 km). The same path can then also be used for

Personnel underground in case of deep tunnel. Possible danger during operation to personnel working in equipment gallery Following DESY rules, a support tunnel could possibly not be accessible during machine

2

emergency personnel and equipment to gain access

operation due to risk of high powered

to the hazard. In this way the distance between

equipment in the tunnel. This concern does not

vertical shafts to the surface could be maximized for

seem to exist in other regions.

operational and equipment purposes and not based solely on life safety and exiting requirements.

Installation of equipment

Simpler installation procedures because the equipments in the two tunnels can be installed in parallel

Larger effort required for civil construction

2

v. Dec.12, 2005

GDE White Papers 16

Maintenance

Exchange or repair in parallel to machine operation. Reduces need for maintenance periods

Work on modulators in surface halls easier than in an underground tunnel (deep site)

2

Schedule

Possibly shorter total installation schedule because of less in-tunnel conflicts and greater flexibility in

Possibly longer if schedule dominated by civil engineering and installation of conventional 2

interleaving installation and early commissioning

facilities

Minimum extra cost for greater reliability of the

The TESLA TDR estimated a cost difference of

components and RF overhead (US estimate 64M$ for an additional 3% RF overhead when including

350 MEuros (410 M$) between a 5.2m diameter single tunnel and two tunnels. This included the

the linac, tunnel and RF extension). A similar amount is certainly necessary for greater reliability

cost of high power cables and transformers but no additional modulator/klystron stations or

Cost

of the critical components in a Single Tunnel configuration. Minimum number of vertical shafts (in case of deep tunnel) as required by cryogenics (about every 5

reliability improvements. The US Options study estimated a 5% increase of the Total Project Cost for two tunnels offset by 3% for reliability improvements. Recent studies from the US,

1

km). No additional shafts (US estimate 8.4M$ per 120 m shaft) necessary for safety or reliability

Japan and CERN, estimate the cost difference between a 5.2m single tunnel and two tunnels (a

issues. No extra cost for long high power cables and

4m support tunnel and a 3.2m beam tunnel) at respectively about 270, 240 and 350 M$. These

transformers between modulator and klystron (US estimate about $150M$).

estimates vary substantially between regions in total civil engineering cost.

v. Dec.12, 2005

GDE White Papers 17

12.4 Positron Source White Paper

Tom Himel, Karou Yokoya, Nick Walker

November 14, 2005

Executive Summary Summary

The keep-alive source should have at least 10% of the nominal positron intensity. The undulator should go at about the 150 GeV point in the linac

v. Dec.12, 2005

GDE White Papers 18

Executive Summary

We were charged to make two recommendations: the requirements for the keep-alive source and the proper location of the undulator.

The primary requirement for the keep-alive source which came out of the availability studies is that it be strong enough that diagnostics (primarily BPMs) work as well with the keep-alive source as they do with full intensity beams. There must be no gain, offset, or resolution changes that prevent machine development and beam based alignment results from being as useful as those done with the undulator source. We asked a few diagnostics people what intensity this would take and they thought they could do it with 1% of design intensity but admitted they were uncertain as systematic errors are the problem and there is no design yet. We recommend a minimum intensity requirement of 10% of nominal intensity to reduce the chance of such systematic errors making the keep-alive source nearly useless, and because there are inexpensive ideas on how to make a ≥10% source. This source would have all bunches filled to 10% of nominal intensity. Note that for many purposes higher single-bunch intensity is better even at the expense of populating a smaller fraction of the bunch train.

After considering a large number of pros and cons between placing the undulator at the end (END) or at the 150 GeV (MID) point of the linac, we concluded that all were minor compared to their differing yields as a function of energy. They are both currently designed to have a yield of 1.5 at a beam-energy of 150 GeV. Note that the desired actual yield is 1.0 and the design value of 1.5 was chosen to ensure that 1.0 can be easily reached without a lot of tuning and to provide some insurance in case the real accelerator doesn’t perform to the design. The MID design has a yield that is a constant 1.5 over the full energy range. The END yield varies with beam energy. It has dropped by a factor of 4 by 100 GeV, by a factor of 300 at the Z (meaning the 10% strength keep-alive source would be used at the Z for detector calibration) and has increased a factor of 2.5 by 250 GeV.

The decision basically came down to the advantage of END being that it ameliorates the risk of a low e+ yield at energies above 150 GeV. Its disadvantage is a guaranteed lower yield (and hence luminosity) at beam energies below 150 GeV. With this as the major factor and without additional clarification on the physics requirements (luminosity) at lower centre-of-mass energies, we recommend the MID location for the undulator.

v. Dec.12, 2005

GDE White Papers 19

Required Intensity of the keep-alive source

Considerations (what we gain as intensity goes up)

Consider 3 possible levels for the keep-alive positron intensity

← • LOW: defined to be enough for diagnostics to work, but not enough to do serious MD and beam based alignment work

← • MEDIUM: defined to be enough for diagnostics to work well enough to do serious MD and beam based alignment work, but not enough to work on collective effects or thermal problems in the DRs.

← • HIGH: defined to be enough to work on collective effects and thermal problems in the DRs

Availability simulations showed that the gain from LOW was minimal while MEDIUM allowed the ILC with an undulator source to be up almost as much as one with a conventional source. While HIGH didn’t increase the availability much more, that certainly depends on assumptions as to how much trouble will be caused by collective effects and heating in the DR. If they are more troublesome than assumed for the simulation, then HIGH intensity could be more important.

Note that ability to go to a higher intensity is always better, so a decision on the requirement must also include information on how hard it is to achieve the requirement.

v. Dec.12, 2005

GDE White Papers 20

Description of possible sources

We considered two possible forms of the keep-alive source that have been worked out by the sources group.

1. The first is a 10% intensity source which uses a ~500 MeV linac to direct a beam at the same target used for the undulator source. Details can be found at

http://www.eurotev.org/e158/e1365/e1378/e1520/EUROTEV-Report-2005-019-1.pdf

2. The second time-shares the 5 GeV positron booster linac to produce roughly full intensity bunches at half the nominal bunch rate. It requires a 250 MeV linac, a high power positron target and capture section, and some transport lines. A drawing can be found at https://ilcsupport.desy.de/cdsagenda/askArchive.php?base=agenda&categ=a0533&id=a0533s1t9/moreinfo.

The second source is clearly more difficult and expensive than the first although significantly more powerful.

v. Dec.12, 2005

GDE White Papers 21

Recommendation

The availability studies and source costs make it clear that MEDIUM is the preferred option. Next it is necessary to determine what intensity that implies. The primary requirement for the MEDIUM keep-alive source which came out of the availability studies is that it be strong enough that diagnostics (primarily BPMs) work as well with the keep-alive source as they do with full intensity beams. There must be no gain or offset or resolution changes that prevent machine development and beam based alignment results from being as useful as those done with full beam intensity. We asked a few diagnostics people what intensity this would take and they thought they could do it with 1% of design intensity but admitted they were uncertain as systematic errors are the problem and there is no design yet. We recommend a minimum intensity requirement of 10% of nominal intensity to reduce the chance of such systematic errors making the keep-alive source nearly useless and because there are inexpensive ideas on how to make a 10% source. This source would have all bunches filled to 10% of nominal intensity. Note that higher single-bunch intensity is better even at the expense of populating a smaller fraction of the bunch train.

Location of the Undulator

v. Dec.12, 2005

GDE White Papers 22

Description of the two locations considered

Two locations have been considered for the location of the undulator.

1. 1. At the end of the linac. This would be just downstream of the MPS collimators, energy measurement chicane and fast extraction system. Downstream of it are BDS corrections, diagnostics and the big bends to split the beams to go to two IRs. This will be referred to as END in the remainder of this document.

2. 2. At roughly the 150 GeV point of the linac. This energy is chosen so that one can run from the Z energy up to 250 GeV without changing the electron beam energy that goes through the undulator. The electron beam is decelerated in the rest of the linac after the undulator when the beam energy for collision is below 150 GeV. This will be referred to as MID in the remainder of this document.

v. Dec.12, 2005

GDE White Papers 23

Temporary design choices for the undulator sources

For both END and MID there are further design choices that need to be made. For clarity of arguments in the paper, we have made these decisions in a way that we think makes each option as good as possible. Once the location is decided, these smaller decisions should certainly be considered more carefully and final decisions made. In some of the pros and cons below we will mention the effect it would have if one of these decisions was taken differently.

Provisions for low energy running

END has a luminosity a factor of four less than MID at 100 GeV as will be discussed below. This could be mitigated to only a factor of two with the addition of a bypass line from the 100 GeV point to the end of the linac, or by increasing the length of the undulator, should the physics case require it. For the purposes of this comparison, we assume this is not done to keep the cost down.

v. Dec.12, 2005

GDE White Papers 24

Other design choices

Other design choices such as the allowed emittance growth in the bends, shape of the beam line to separate the gammas from the electrons, and separation distance of the positron target from the electron beam line mainly effect cost. As the cost difference turned out to be smaller than the errors in the cost it isn’t necessary to enumerate the design choices here

v. Dec.12, 2005

GDE White Papers 25

Pros and cons

Positron yield for beam energies between 100 and 150 GeV and at the Z: favors MID. In this energy range

the beam energy in the END undulator varies between 150 and 100 GeV. This decreases the e+ production rate and makes the luminosity roughly drop (in addition to the scaling from adiabatic damping) so that at 100 GeV it is one fourth that of the MID solution. See Figure 1 for the simulation results of Wei Gai which are consistent with those of Klaus Floettmann. Note that the physics requirements are not met in the 100 to 150 GeV beam energy range for END. This is mitigated somewhat by energy run plans (http://www.slac.stanford.edu/econf/C010630/papers/E3006.PDF) that have only 10% of the integrated luminosity in the effected energy range. Having the runs at those energies occur at half the luminosity hence makes the average design luminosity of END 10% less than that of MID. Note that the design yield of 1.5 at 150 GeV is considered to be a necessary margin to make sure a yield of 1.0 is actually achievable without constant tuning. It is wrong to use the factor of 1.5 to say the yields at lower energies are adequate.

The END yield at 50 GeV (for Z calibration) is very small. The keep-alive source would be used instead of the undulator for this running. As the keep-alive source is specified to be 10% of the nominal intensity, the Z calibrations for END will take 10 times longer than those for MID. There are widely varying numbers in circulation for the amount of Z luminosity needed for the calibration so we are not able to determine the overall impact of this factor of 10 luminosity difference.

Figure 1: Positron yield as a function of energy as calculated by Wei Gai at ANL.

v. Dec.12, 2005

GDE White Papers 26

Positron yield at high energies: favors END. At high energies, the e+ yield for the END option will be >> 1 making e+ intensity tuning trivial. This is really a risk mitigation effect. If the design yield of 1.5 is achieved, it is large enough and the extra yield doesn’t help. However, if we miscalculate the yield or the DR acceptance is significantly smaller than planned, then the extra yield at high energies will be very welcome. A numerical example is that if the DR acceptance is 0.04 instead of the design of 0.09, then the calculated yield at 150 GeV drops from 1.5 to 1.13. For MID it would be 1.13 at all beam energies. For END it would increase to 2.8 at 250 GeV.

The above 2 pros and cons regarding yield are considered by far to be the most important. Cost would be important except that the total cost of the e+ system is fairly small and the cost differences we have evaluated have fairly large uncertainties that depend on engineering which has not been done.

Cost: favors neither. We did a crude cost estimate of the two options. They came out equal within errors. Different assumptions could change the relative cost by around 20% which is not a large enough difference to have a significant impact on the decision. For example adding a bypass line to improve the END low energy luminosity had this effect. Things we took into account in the cost estimate were that END made use of the existing BDS protection collimators and fast extraction dump and required slightly longer arcs to limit emittance growth. There are no designs of either option with enough detail to evaluate differences in terms of number of access shafts or costs to avoid interference of the gamma line with the electron line.

The remaining pros and cons are all much lesser weight and are listed primarily to let people know they were considered and to help guide future reviews and reexaminations of the decision. They are listed in essentially random order.

Energy jitter for beam energies less than 150 GeV: favors END. At low energies when the undulator is at the 150 GeV point, the beam must be decelerated after the undulator. Both the acceleration and deceleration add to the energy jitter resulting in a higher energy jitter for this case than when the undulator is placed at the end. The worst case is when each section of the linac has an energy jitter (probably due to phase jitter) that is ) independent of the other sections and the desired beam energy is very low (say 50 GeV to run on the Z0. For this case, the MID undulator must accelerate the beam to 150 GeV

and decelerate it by 100 GeV for a total of 250 GeV of acceleration. The end undulator only needs 50 GeV of acceleration for the luminosity beam. Its energy jitter will thus be = 2.2 less than for MID undulator. The WWS requirements state that the energy jitter should be less than 0.1%. A rough calculation based on numbers from the TDR (and the ILC design is different than that) indicate the MID energy jitter would be about 1.4 times this WWS requirement. This would have to be accepted, or mitigated by reducing the random energy jitter of each RF station.

v. Dec.12, 2005

GDE White Papers 27

Note that no extra emittance degradation is expected for low energy running with MID. If anything, there will be less emittance growth than for END as the beam reaches a high energy sooner and hence wakefields have a smaller effect.

Need for e+ tuning when energy is changed: favors MID. With the undulator at 150 GeV, the beam energy only varies downstream of the undulator as energy at the IP is varied. For the end undulator, the beam energy will change in the undulator and some tuning of the e+ production is likely. Note that increasing the energy should be easy as the yield increases, but decreasing the energy to near 150 GeV is more likely to require more tuning.

Flexibility of linac operation: favors END. The MID solution requires the first section of the linac to always run at full gradient while END allows the flexibility to run that way or to run everything at lower gradient when the maximum beam energy is not required. Running below 150 GeV in MID requires part of the linac to run back-phased (to decelerate the beam). While possible in principle, actual experience in a SC linac is lacking. (It is commonly done at SLAC.)

BDS upgrade flexibility: favors MID. This flexibility could be important to allow for some improvements, such as additional collimation stages, or lengthening the diagnostics section, or addition of a second interaction region. If the BDS is attached to the end of a straight linac, (the case for MID), one can simply remove cryomodules and extend the BDS into the linac tunnel. If the undulator is placed at the end of the linac, the bends and the undulator would have to be moved upstream in this upgrade scenario.

Difficulty of Main Linac Energy Upgrade.

← • If full length tunnels are built and the upgrade is done by adding RF to the downstream end: favors MID. For this energy upgrade, MID needs no modifications. END may need to have its undulators replaced with ones better matched to the higher beam energy and perhaps have its bends made more gently to reduce emittance growth. (An option to put the undulator at the end of the phase 1 linac at the 250 GeV point, allowing the phase 2 linac to be constructed downstream was not considered in this paper.)

← • If short tunnel is built and upgrade is done by digging more tunnel in the upstream end: favors neither. The changes needed for END are the same as above. For MID, one has the choice of leaving it in its original location and making changes similar to END, or moving it upstream to the new 150 GeV point.

v. Dec.12, 2005

GDE White Papers 28

v. Dec.12, 2005

GDE White Papers 29

Recommendation

After considering a large number of pros and cons between placing the undulator at the end (END) or at the 150 GeV (MID) point of the linac, we concluded that all were minor compared to their differing yields as a function of energy. They are both currently designed to have a yield of 1.5 at a beam-energy of 150 GeV. Note that the desired actual yield is 1.0 and the design value of 1.5 was chosen to ensure that 1.0 can be easily reached without a lot of tuning and to provide some insurance in case the real accelerator doesn’t perform to the design. The MID design has a yield that is a constant 1.5 over the full energy range. The END yield varies with beam energy. It has dropped by a factor of 4 by 100 GeV, by a factor of 300 at the Z (meaning the 10% strength keep-alive source would be used at the Z) and has increased a factor of 2.5 by 250 GeV.

The decision basically came down to the advantage of END being that it ameliorates the risk of a low e+ yield at energies above 150 GeV. Its disadvantage is a guaranteed lower yield (and hence luminosity) at beam energies below 150 GeV. With this as the major factor, and without additional clarification on the physics requirements (luminosity) at lower centre-of-mass energies, we recommend the MID location for the undulator.

v. Dec.12, 2005

GDE White Papers 30

v. Dec.12, 2005

GDE White Papers 31

12. 5 Comments on the Laser-Straight vs.Curved Tunnel

W. Funk, D. Schulte, T. Shidara

November 15, 2005

1 Introduction

Three main options exist for the tunnel layout. The first is a laser straight tunnel, the second a tunnel that follows the curvature of the earth, and the third a tunnel that consists of short straight sections that are joint by small bends. In case of a laser straight tunnel, the linac ends that are far from the detector would have an inclination of 3-4mradian with respect to the equipotential of gravity, while the central area would have no inclination.

The beam delivery system experts require that their system be straight; in addition the last part of the linac shouldalso be straight, about 0.5km.Hence, also the curvedtunnel will needa bend at the end of the linac. The inclination at the end of the linac would in this case be of the order of 0.5mradian.

The choice oftunnel design has to consider the impact on three main areas, the beam dynamics, the cryogenic system as well as the conventional facilites, tunnel construction and installation.

In general, we find that the beam dynamics favours a laser straight tunnel while the cryogenic system favours a tunnelthat follows the curvature ofthe earth.The preference in terms ofconventional facilities and installation costs is strongly site dependent.

In the following, mainlythe curvedand laser-straight tunnelwill be discussed.The piece-wise straight tunnel has essentially the same properties in terms of tunneling and cryogenics than the curved tunnel, given that the bends are close enough.The main drawback is the needofadditionaltunnel length; about 800m for the full machine has been suggested, using two bends pers side. However, the number of bends will mainly be dominted by the cryogenic system, if the piece-wise straight tunnel is chosen.

v. Dec.12, 2005

GDE White Papers 32

2 Tunnel and Conventional Facilities

In a level area, a tunnel built using a cut and cover technology may be cheapest. Such a tunnel requires to follow the earthcurvature.In most areas, it will however be requiredto rather bore a tunnel since the land above the tunnel may not be available over the full length or it may be too costly.

If cut and cover cannot not be used, a curved tunnel may still be cheaper. In a level area, potential cost saving may exist due to shorter access shafts and the possibility to built the experimental hall using a cut and cover approach. However, the potential savings are very site dependent and cannot be assessed without detailed knowledge of the geological constraints and other factors. We realise that under certain conditions a laser straight tunnelwithan inclination withrespect to the earthsurface may be the cheapest solution. This is for example the case for a site close to CERN.

It is clear that not only the tunnel cost plays a role but that also the installation cost for the main linac need to be considered.

v. Dec.12, 2005

GDE White Papers 33

3 Beam Dynamics

The vast majority of the beam dynamics studies have been made for a laser straight tunnel. Only a very limited e ort has been made to study the impact of a curved tunnel. ff

In general, it is clear that beam dynamics favours a laser straight tunnel. In a curved tunnel, the dispersion needs to have finite values along the main linac, i.e. the transverse beam trajectorywill depend on the beam energy. In a laser straight machine, the dispersion can be zero. However, the preliminary beam dynamics studies have sofar not shown any severe implication of the non-zero dispersion for the ILC. It is strongly recommended to perform detailed studies of the e ect of the non-zero dispersion on ffbeam-based alignment, tuning and feedback.

The following points have been considered:

• Emission of synchrotron radiation in the quadrupoles. If one assumes that the beam is bent in the quadrupoles using the corrector coils, the emitted radiation power per linac length is P=50mW for E=500GeV. This power scales linearly with the ratio of quadrupole spacing to quadrupole length, a value of 34m/0.66m has been assumed; for larger quadrupole spacings the power will increase accordingly. This power is not negligeable but should be acceptable compared to the overall heat load budget. The critical enery stays below 5MeV, so excessive component activation is avoided.

• The emittance growth due to the non-zero vertical dispersion. By simulating an ideal machine, three diferent studies performed at DESY [2], CERN [3] and KEK [4] showed that the additional emittance growth is very small compared to the inital value.

• The emittance growth due to current ripples ofthe correction coil power supplies. It is assumedthat the corrector coils ofthe normal linac quadrupoles are usedto steer the beam to follow the earthcurvature. Avariation of the corrector coilcurrent will thus apply transverse kicks to the beam.Simulations at CERN indicate thata power supplycurrentvariation of 0.010.02nm,which is quite small. The e ect on the beam fftrajectory is much more significant. Simulations at KEK indicate thata power supplystabilityof 0.003size. However, one should be able to correct this e ect by an ffintra-pulse feedback, provided the involved timescales are well above tens of microseconds.

• Scaling errors of the beam position monitors and errors in the assumed beam energy. Dispersion free steering is a likely option for the beam-based alignment of the ILC main linac. In this method, beams of di erent energies are used in the correction procedure. In ffcase of a straight tunnel, one aims to make these beams follow the same trajectory, a nullingmeasurement.In case of a curved tunnel, due to the non-zero dispersion, the trajectories of these beams must be di erent. One must thus be able to accurately ffdetermine the required di erence of the trajectories and to measure them precisely. This ffrequires that the scale factors of the BPMs be accurately known, as wellas the lattice andthe beam energy.In case ofthe straight tunnel, the error in BPM response as well as wrong assumptions for the beam energy will have a much smaller eject and can be

v. Dec.12, 2005

GDE White Papers 34

largely mitigated by iterating the correction procedure.

It is conceivable, that a piece-wise straight tunnel has an adavantage over the curved tunnel, since it may be possible to more accurately determine the beam energy in the crucial bending sections. For the straight parts the same advantage as for the laser straight tunnel applies.

Simulations at CERN indicate that a knowledge of the scale factor with the precision of 1% is suffcient for a curved tunnel; studies at DESY indicate that a knowledge of the energy difference between the beams with 2% precision should be suffcient. Understanding of the impact of systematic errors is necessary.First indications are positive for simulations at CERN. If one aims to use a test beam with 20% energy difference but uses 15% or 25% instead, the resulting emittance growth still remains below 1nm.

• Implications for a dog-bone damping ring. In case a dog-bone damping ring is chosen, it will share the tunnel with the main linac or be in a tunnel parallel to the main linac. Studies at DESY have indicated that it is possible for the damping ring to follow the earth curvature, but verification is needed.

In summary, the beam dynamics consideration favour a laser straight tunnel. But in the studies carried out sofar, the impact of a tunnel that follows the earth curvature has been quite small. One can therefor expect that more detailed studies will yield a similar result. In this case the preference for the laser straight tunnel will be rather weak. Having said this, it is very important to actual verify this by carrying out the detailed studies.

The piece-wise straight tunnel will di er from the curved tunnel by avoiding potentialff complications in the accelerating parts ofthe main linac proper by putingthem into dedicated sections.While this has andvantages compared to a curved tunnel, the relative benefits are not established. Studies indicate that the bends are acceptable [1].

It should be noted that at least one bend will be needed at the end of the curved linac, in order to allow for a laser staight beam delivery system. This bend can however be relatively weak.

It should also be noted that a multi-TeV upgrade, which is likely normal conducting,

would strongly prefer a laser straight tunnel[5]. While the studies are not complete they indicate that a noticeable luminosity loss is to be expected. Potentially the loss could be quite large.

v. Dec.12, 2005

GDE White Papers 35

4 Cryogenic Systems

The inclination of 3-4mradian at the ends of the laser straight tunnel has consequences for bothcryomodules and the cryogenic distribution system. Liquid helium is delivered to the superconducting cavities as a two-phase fluid in an 8 cm pipe inside the cryomodules, with liquid flowing in one direction, and gas being pumped in the other. For tilts of up to about 0.3mradian one can expect that no modifications to the cryosystem are need, since the di erence in height for a string length of 150m rmains below half the pipe ffdiameter;this however needs to be verified.

The consequences of a tilted linac are:

← • Cryogenic system designers have no choice about liquid flow direction. The ’conventional’(gravi-tational equipotential) collider helium distribution system envisages the refrigerator in the center of a segment of linac, flowing helium in both directions. This will no longer be possible.

← • The vertical distance between the linac andthe refrigerators, assumedto be on the surface, will be 50-60 m more at the IP than it is at the injectors. This di erence in hydrostatic head will influence the design of the refrigerators, and ffmay necessitate locating some or all of the refrigerator at the linac elevation.

← • Modifications will be required for the two-phase helium distribution pipe in the cryomodule, to ensure that the liquid surface is never less than some minimum distance above the bottom of the pipe (to ensure adequate heat transfer from the cavities)andnever less than some minimum distance below the topofthe pipe (to ensure adequate cross-sectionalarea to pumpthe evaporated gas with acceptable pressure drop). The simplest modification proposed is the addition of periodic dams or weirs in the pipe to ensure both a minimum and a maximum liquid level. The minimum separation of weirs in the most inclined portion of the linac is comparable to the length of a module. Deciding to put one (or perhaps two) weirs into each module would not be technically di cult or expensive, andffi would maintain interchangeability of modules.

Potential negative consequences of the weir system include:

← • A large increase in the time delay between the introduction ofadditional liquid from the refrigerator and the response of a liquid level sensor at the other end of the distribution system.

← • ’Dryout’ of the linac between two weirs becomes a real possibility.

v. Dec.12, 2005

GDE White Papers 36

Both of these might be counteracted by the insertion of an additional level sensor into each module. It is clear that significant R&D is needed to validate this design concept. Topics that need to be addressed include:

← • What is the minimum liquid level required and what margin should be applied to that?

← • What is the minimum cross-section neededto ensure acceptable pressure dropin the gaseous flow? How much margin is needed? How does that impact the topology of the connection between the two-phase pipe and the 30 cm diameter Gas Return Pipe?

← • At what operating parameters willentrainmentof liquid drops in the gaseous flow begin to appear? What impact does this have on pressure drop and on gaseous and liquid dynamics?

← • How are the responses to the preceding questions a ected bynon-ffuniformities of heat transfer from cavities? In other words, what if one cavity in a module has significantly greater losses than the others?

← • Can we design a controlsystem thatcan create an acceptable fluid distribution in a significant length of linac, and then maintain that distribution in the face of the variations in heat load described above, as well as transients arising from cavities tripping off?

Answers to these questions, and many more sure to be generated as these ideas are considered further, will require a substantial R&D program. Other potential solutions to the problem are being considered.

The requirement of the laser straight beam delivery sytem requires in any case that the cryogenics must be able to tolerate an inclination of 0.5mradian. The related R&D must be carried out if one cannot remove this constraint.

In the piece-wise straight tunnel, the distance between the bends needs to be less than about 5-6km, in order not to make the reuiqrements for the cryogenics more demanding in the main linac than at its end. The number of bends per side thus needs to be three to five depending on the site length.

In conclusion, there is currently no evidence that the problems related to an inclination of the cryomodules cannot be overcome. In a tunnel following the earth curvature, the cryogenic system will be simpler and likely less costly than in a laser straight one, but the actual difference is not yet known.

R&D on the potentialto use cryomodules with an inclination will prove valuable.It is

v. Dec.12, 2005

GDE White Papers 37

not yetcertain that no problem is found for the beam dynamics. If geological considerations favour an inclined site, this R&D may also help to reduce cost. For a piece-wise straight linac, one needs to verify the maximum inclination that can be tolerated as this determines the maximum distance between the bends; from the beam delivery sytem an angle of 0.5mradian is required.

v. Dec.12, 2005

GDE White Papers 38

5 Conclusion

Based on a review of the available material, we conclude that there is no evidence that any of the three options is not viable. The choice can therefor mainly be based on cost considerations. The actual optimum choice is site dependent but in most cases it is expected that the tunnel that follows the earth curvature is cheapest, while the piece-wise straight tunnel is only somewhat longer.

For a tunnelthat follows the earthcurvature, a detailedstudyofthe impacton beam dynamics has to be carried out. We recommend that all beam dynamics studies include the curvature, in order to ensure that any potential di culty be found at the earliest time.ffi These studies should be pursued with high priority. The modules close to the end of the linac will need to have an inclination of 0.5mradian, this very likely necessitates some modifications of the cryomodule layout or a shortening of the cryostrings. Detailed studies of how the inclination can be tolerated is needed.

The piece-wise straight tunnel requires that the maximum inclination angle be identified, since it will determine the bend distance. For distances between the bends of 5-6km, it will not tighten the requirements compared to what anyway needs to be fullfilled to satisfy the beam delivery system requirements. The potential cost of the bends can however be substantial. Based on the 200m long bend design in [1] and five bends per side one needs to increase the tunnel length by 2km; also the magnets start to cost. Also for this case the beam dynamics needs to be studied in detail.

In case of a site where a laser straight tunnel would be preferable, R&D needs to be carried out in order to ensure that the cryogenics system can achieve the required specifications. Di erent approaches exist and will need to be investigated in detail. Also fffor a laser straight main linac, the beam dynamics needs more detailed studies.

It should be reviewed whether the beam delivery system and the last part of the main linac need to be laser straight. If also the last part of the linac could be built curved, this would significantly ease the task of the kryogenics and saving the bend would somewhat reduce the cost.

v. Dec.12, 2005

GDE White Papers 39

References

[1] P. Tenenbaum, http://alcpg2005.colorado.edu:8080/alcpg2005/program/accelerator/WG1/petertenenbaum20050812163

[2] N. Walker, Emittance Preservation in a possible ILC Main Linac which follows the Earth’s Curvature. http://www.eurotev.org/e158/e1365/e1378/e1489/eurotev-report-2005-17-1.pdf

[3] A. Latina, D. Schulte, http://alatina.home.cern.ch/alatina/curved linac.pdf

[4] K. Kubo http://lcdev.kek.jp/ILCAsiaNotes/2005/ILCAsia2005-23.pdf

[5] H. Braun, J.-P. Delahaye, D. Schulte, ILC Compatibility with a Multi-TeV Upgrade. CLIC-Note 644 (2005)

v. Dec.12, 2005