Tevatron Beam Diagnostics: Boosting Collider Performance Vladimir Shiltsev, Ron Moore, Andreas...

Tevatron Beam Diagnostics: Boosting Collider Performance

Vladimir Shiltsev,Vladimir Shiltsev, Ron Moore, Andreas JanssonRon Moore, Andreas Jansson

FNAL/Accelerator DivisionFNAL/Accelerator Division

05/01/2006 Tevatron Instrumentation 2

Winston Churchill

“Generals are always prepared to past wars.”

“Soldiers usually win the battles and generals get the credit for them.”

Napoleon Bonaparte

Tevatron Collider Run II Was/Is a Battle


So, Purposes of This Talk Are:

• Give credit to soldiers• Show some remarkable luminosity-related

examples of diagnostics progress:– Lattice /helix issues BPM Upgrade– Reliability/uptime orbit stabilization + HLS +

vibrations– CDF/D0 discrepancy IP vertex diagnostics – Beam-Beam 1.7GHz Schottky + Tune

Stabilization – Losses on ramp IBEAM + FBI + Tune Tracker– Losses at 150 Head-Tail Monitor – DC beam AbortGapMonitor + LPM +SBD– Injection mismatch FWs + IPM + OTR

• Outline lessons learned during the Run II


Tevatron Collider

1.96 TeV Proton-Antiproton Colider1.96 TeV Proton-Antiproton Colider- worlds’ most powerful accelerator- worlds’ most powerful accelerator

Run I (1992-96) t-quarkRun I (1992-96) t-quarkRun II (2001-09) Higgs, SS, >SMRun II (2001-09) Higgs, SS, >SM


Overview of “the Battle”


0.5-1% of loss Poor lifetime 0.5-1 mm orbit error Collimator malfunct Sequencer error

QUENCH !2-4 hrs to recover4-8 hrs for new pbars

or Blow Up

tune few 0.001 coupling ‘s wrong

instability kicker malfunct

1000’s of PSs

“Driver’s Nightmare”

Store #535Jun 15, 2001


“Things going wrong”: 2001

Store #535Jun 15, 2001

So called “comfort plot” (at that time – more of discomfort) as seen by operators in Main Control Room and automatically saved in E-log

Shows intensity over the most critical times of injection, ramp and squeeze:

Blue – protonsRed – antiprotonsGreen – total (DCCT)


Situation improved: 2005

Store #4116Apr 27, 2005

Note :Losses 1/20 of ‘018x proton intensity40x pbar intensity


Importance of Helical Orbits

• Beams share beam pipe be separated – Helical separation ~(10-22)mm at 150 GeV– S ~(3-6) mm at 980 GeV

• Lifetime is strong function of S– 30 sec at 2σ , 50 hrs at 7σ– Aperture limited at 150 GeV “smooth”


Orbit Smoothing and Understanding Optics

• ~0.5 mm rms orbit drift in 1-2 weeks• proton and pbar Qx,y, Q’, lifetimes vary significantly due to such COD needed regular orbit smoothing at 150, ramp, flat-top, squeeze, low-beta. … PLUS• lattice had to be known to <5% (was ~25%)• TBT not reliable• dependent on bunch structure/length• did not see pbars Decided to upgrade BPMs (2003-05)

“orbit – silver orbit”

+-2 mm at collisions

after about 2 weeks in September’02


Beam Position Monitors

• Existing BPMs fitted with new electronics based on Digital Down-Conversion

• Resolution ~5µm (previously LSB was 0.15mm)

• Eliminated difference between coalesced and uncoalesced beams.

• Measures pbar orbit,too

S.W

olb

ers

et

al

A/DDDC FPGA

A/DDDC FPGA

Echotek board

Echotek board53MHz BPF

MVME2400

*Note that “resolution” includes real beam motion, mainly in the horizontal plane (synchrotron osc.)


New BPMs De-coupling and Lattice measurement

• Calculating all optics functions and coupling correction using TBT data from many BPMs

E.G

ian

felice-

Wen

dt

# corrections

min

tu

ne

split

• Beta-functions measured to better than 5% accuracy on both helices, errors cor-rected, lattice modified so that β*=36cm29cm, giving ~10% gain in Luminosity


-functions : before & after

Upgraded Tevatron BPMs !

Oct’05

Mar’05Vertical beta-function

A/V

alish

ev,

Y.A

lexah

inJ.

An

nala

, V

.Leb

ed

ev

150 m Horizontal beta-function

B0 D0

Change of beta @ IPβ*=36cm29cm gave ~10% gain in Luminosity


Orbit stabilization

• Slow (~1/min) automatic continuous correction of orbit variations using several dipole correctors close to the IPs

• Standard at most Light Sources – only recently commissioned in Tevatron with new BPMs (old were too sensitive to bunch structure quench fear)

• Now need fast FB

orbit stabilization

ON

vertical

horizontal

0.4mm

1 day

V.R

an

jbar


Sensitivity to quakes

“Fire-truck-quake” ~200 um orbit jitter

New diagnostics: tiltmeters, Water Levels on every LB quad

M8.7 Sumatra earthquake 03/28/05


CDF detector sinking ~0.5 mm/year Interaction Point Moves inefficient b-tagging

Intentional move upon CDF

request

Reported by CDFSilicon Vertex Detector


CDF and D0 IP Waist Diagnostics

J.Estr

ad

a/A

.Ch

an

dra

01/04

01/05 01/06

Vertical beta-function measured at D0

Jan’2004 to Mar’2006(“beam diagnostics for free”)

• Vertices of p-pbar collisions analyzed and processed

(on-line @ CDF , off-line with ~day delay @D0) IP

position, luminous region waist size vs z


Tevatron: Life in the “Tune Box”

7th order resonances:

Q=4/7=0.571 -

HIGH LOSSES


Q=7/12=0.583 -

Bad lifetime


Q=3/5=0.600 –

EMITTANCE BLOWUP

proton

pbarevery bunch has every bunch has

its own tune!its own tune!


Tune Diagnostics: 21.4 MHz Schottky

• Workhorse for shot setup. Operators determine tune from spectrum peaks

• Often needs excitation (VTICK)

• But:But: a) does not see pbars anywhere, b) very complicated by coherent tune lines,

• c) does not see bunch-by-bunch tunes

B.

Fellen

z


1.7 GHz Schottky Detector

Slotted Waveguide Pickup

1.7 GHz 109 x 75 mm aperture

Bunch by Bunch Gate Generator

Analog Multiplexer

OscilloscopeVector Signal Analyzer

Phase Locked Multiplier Ethernet

BPF

53 MHz Tevatron RF

LPFBPFH or V Pickup

Protons

Antiprotons

Same as above

Injection Calibration Signal

Closed Circuit TV

1.7 GHz

38 NS

Tevatron Schottky Signal Processing

100 MHz @ 1.7 GHz 5 MHz @ 1.7 GHz 10 MHz

Tunnel Service Building

Revolution Marker

Gating Down Conversion R.P

asq

uin

elli

• Vertical and horizontal units• Proton and pbar ports• 100 MHz bandwidth


1.7GHz Schottky Spectra

#322602/11/04

• Q and 1-Q lines are seen

• Fit gives:– Betatron frequency (accuracy ~0.001)– dP/P sum of two

widths– C_vh difference

of two widths– Emittance area

under the peaks

• For each bunch!

… non-invasive!

A.J

an

sson

/P.L

eb

run


Tune stabilization

Starting tune correction

Pbar horizontal tune

Pbar vertical tune

• Operators use 1.7 GHz Schottky data to keep pbar tunes within a predefined range as the beam-beam tune shift changes

4 weeks

R.M

oore


1.7GHz Schottky Bunch Tunes

Bunch-by-bunch tune variation ~0.005 – an indication of parasitic beam-beam interactions


Plotting Bunch-by-Bunch Datafor each of 36 p + 36 pbar bunches

Live dataInstability @ 150 GeV resulted in

interesting intensity and emittance patterns

Logged dataProton bunch centroid motion during

longitudinal instabilitity

Observing differences in bunch-by-bunch behavior is very useful for understanding beam dynamics in the Tevatron

±4deg


Longitudinal Instability Sampled Bunch Display and Phase Monitors

bunchlength

RF phase of bunch

bunch #

Bunch tomography

Weird bunch shapes during instability burst(snapshop taken by SBD)

No instability, continuously “dancing” bunches (RWM)

Fast (200Hz) longitudinal phase

monitor is under development

+-4

deg


Sampled Bunch Display (SBD)

• Measure bunch intensities, lengths– Few – 350 × 109, 1-4 ns– Updates ≈ 1 Hz

• LeCroy WaveRunner 6200 captures waveforms over many turns in memory

– 2 GHz bandwidth, 10 G-samples/sec

• Macintosh G5 does signal processing– 200 tap (0.5 ns/tap) FIR filter removes

effect of dispersion in the long cable

• Resolution of intensity ~0.05% (5e9)

• Resolution of centroid position and RMS length ≈ 0.02 ns (RMS is ~2ns)

Figure 2: a proton bunch signal (raw) and after application of the FIR filter. The feature at the far right is a 3/4 % reflection. from one channel through the splitter to the other. Full height of the main bunch is ~5 amps.


Longitudinal Oscillations Lead to Generation of DC beam in Abort Gaps

• The Tevatron operates with 36 bunches in 3 groups called trains

• Between each train there is an abort gap that is 139 RF buckets long– RF bucket is 18.8 ns Abort gap is 2.6 s

• Protons leak out of main bunches to the gaps. Tevatron is sensitive to few x 109 particles in the abort gaps (total beam ~ 1013) as they lead to quench on beam abort (kicker sprays them)

139 buckets

21 buckets 1113 RF buckets totalTrain

Bunch

Abort Gap


Abort Gap Intensity Monitor

Photocathode

Dynode 1

Photocathode

HV below photocathode

Nominal PMT Behavior (Gated On)

Gated Off PMT Behavior

Dynode 2

Dynode 3

Dynode 4

Dynode 1

Dynode 2

Dynode 3

Dynode 4

HV below dynode 3

Hamamatsu gated MCP style PMT on loan from LBNL

• 5ns minimum gating time w/no noticeable settling time

• Very large extinction ratio

• Somewhat expensive (~$20K /tube)

• 2-stage Micro Channel Plate PMT – Gain of <= 106

• No sensitivity to pre-gate light

R.T

hu

rman

-Keu

p


DC Beam in Abort Gap:TEL On/Off/On

AGM is calibrated wrt to DCCT - the most precise Tev instrument.


Intensity Measurements: DCCT and FBI

• DCCT (DC Current Transformer)– Typical intensities 109 → 1013

– Noise ~0.5 e9 or ~0.005% max

– 24-bit ADC samples @ 6.9 MHz

– Output 128-sample average @ 54 kHz

– Calibrate via external pulser

– N_p+N_pbar together

• FBI (Fast Bunch Integrator)– Bunch-by-bunch intensities via RWM

– Narrow (1) & wide gates (5 buckets)• Main and satellite bunches

– Updates @ up to few hundred Hz

– Sensitivity to temperature improved

– Calibrate via DCCT• Few % correction for satellites

DCCT

Receiver

TBEAM Front-End * ~18 bits

T:IBEAMB - E13

T:BEAM - E12

T:IBEAM - E12

T:IBEAMS - E10

MADC (single sample, 12 bits)

DCCT

Amp.

DCCT

Receiver

TBEAM Front-End * ~18 bits

T:IBEAMB - E13

T:BEAM - E12

T:IBEAM - E12

T:IBEAMS - E10

MADC (single sample, 12 bits)

DCCTDCCT

Amp.

B.

Fellen

zT.M

eyers

Resistive Wall Monitor: Ceramic break with 80 120Ω resistors.

Signals sampled at four locations are summed.


Beam Lifetime Depends on Chromaticity

• Two methods for fast Q’ measurements: – Head-Tail Monitor– Fast and Accurate TuneTracker

Loss rates (LOSTP) versus chromaticity

0400800

120016002000

0 2 4 6 8

Horizontal Chromaticity

Loss r

ate

s (

Hz)

(CDF Detector Proton Loss Counters)


Fast Q’ Head-Tail Monitor

• Particles with different dP/P have different tunes head-tail phase difference ~Q’

• Just few pi d kick• Accuracy ~0.5

unit • Very fast method

– Ops like it!

• Currently used for monitoring– No difficulty to

measure Q’ with octupoles

V.R

an

jbar

0 200 400 6002

1

0

1

2Head minus Tail position

turn number

mm DXi

i

.


Fast and Accurate Tune Tracker

• Beam is lightly excited over a frequency range around f_betatron

• Zero phase response frequency declared =Q

• Accuracy in Q ~0.0001• Very fast method

(3Hz)– Works on every Tev

ramp and in LB squeeze

• Change dP/P and determine Q’– Stat accuracy ~0.2– Syst error ~0.5 unit

Determine and track Phase=0 frequency

C.Y

.Tan


0 10 20 30 40channel #

0

5

10

15

20

25

30

35

nrut#Quadrupole Oscillations due to

Lattice Mismatch: Ionization Profile Monitor

• Single bunch turn-by-turn beam size measurement.

• See separate talk.

1 cm

A.J

an

sson


OTR Monitor

• 5 µm aluminized mylar foils• Rad hard camera, 130x170µm

size pixels• See poster session!

V.

Scarp

ine


Both Instruments under development, but the first results are interesting:

IPM shows significant (~30%) quadrupole oscillationsOTR shows no size change over 1 turn (#2 vs #1, note ampl.)

OTR, vert

Turn #1Turn #2

… just one example of importance of having several instruments for cross-checks

0 5 10 15 20 25

1

1.1

1.2

1.3

IPM, vertmm


Cross- Checks and -Calibration

• Intensity: DCCT and FBI and SBD– DCCT is most precise, but limited function– FBI and SBD within 1%, multi-functional

• Phase oscillations: SBD and LPM (slow and fast)

• Tunes: Schottky 21MHz, 1.7GHz, TuneTracker– All three in operations for different tasks– Absolute differences dQ~0.005, relative changes <0.001– TT fastest and precise, 1.7GHz most versatile

• Emittances: Flying Wires, SyncLite, Schottky – Tons of efforts to bring the three to within ~10%– FWs are ~main tool, used for 1.7 GHz Schottky calibr-n

• Luminosity: CDF and D0 different by ~20% (!)


Lessons Learned - I

Due to peculiarities of SC synchrotron operation, non-invasive beam diagnostic instruments should be preferred, effects of intrinsically invasive ones minimized (FW 337um)

Having two or more instruments for same beam parameter measurements (makes life more complicated to bring them together but) makes the data more trustworthy

Fast data collection rate (at least 1 sec for all channels) is a must – at all stages, for all bunches, all the time – and saved for years (for postmortem)

Detectors have tons of beam diagnostics, so good communication channels with them are important


Lessons Learned - II

Accept help/ideas from other groups/labs – many of them have invaluable expertise:- CD: BPM upgr; PPD: BLM upgr/SL/beta* monitors; LBL: Abort Gap Monitor MCP-PMT; ANL: e-cloud detectors, etc.

An instrument development is fast, compared to time needed to make it “fully operational” and satisfactory for operators and physicists – a lot of

efforts went into that

Team up diagnostics developers and users from the very beginning till commissioning of the instruments (and even beyond that – in operation)


Teaming Up for Instrument Development

Instruments Developers Commissioners, Users Beam Line Tuner D.McGinnis/V.Scrapine J.Annala dEmm@ Inj, ”last sec” V.Scarpine A.Xiao/J.Annala BPM upgrade S.Wolbers/R.Webber M.Martens/J.Steimel 21.4MHz Shottky B.Fellentz P.Lebrun/D.Still 1.7GHz Shottky R.Pasquinelli A.Jansson/V.Shiltsev Tune Tracker C.Y.Tan C.Y.Tan/J.Annala SyncLite/Abort Gap Monitor R.Thurman-Keup A.Valishev/V.Shiltsev Flying Wires J.Zagel/S.Pordes A.Jansson/E.McCrory IBEAM+SBD+FBI R.Flora/S.Pordes A.Tollestrup/J.Annala Head-Tail Monitor V.Ranjbar V.Ranjbar Baseband Schottky A.Semenov/C.Y.Tan

V.Lebedev/J.P.Carneiro Vacuum Diagnostics/RGA B.Hanna V.Shiltsev HLS/Vibrations J.Volk/T.Johnson V.Shiltsev/J.Annala Longitudinal Phase Monitor A.Ibrahim J.P.Carneiro/V.Shiltsev Luminosity+IP diagnostics CDF/D0 V.Papadimitriou/V.Shiltsev IPMs/OTRs A.Jansson/V.Scarpine A.Jansson Beam Loss Monitor upgrade J.Lewis/S.Pordes J.Annala/D.Still Software (DLPlotter, OAC, SDA) T.Bolshakov/Cntrls

R.Moore/J.Slaughter


So, Have We Won The Battle?

Yes, but only one one of many…Yes, but only one one of many…

… …the campaign is not overthe campaign is not over


“Waiting for Mr.Higgs…”

4 more years > quadruple luminosity integral may be Mother Nature will smile on

us


And at the end…

On behalf of the Run II team, we’d like to thank all who took part in beam diagnostics development

and

Thank you for your attention!


backUp Slides


Fast Q’ Head-Tail Monitor

• Particles with different dP/P have different tunes head-tail phase difference ~Q’

• Just few pi d kick• Accuracy ~0.5

unit • Very fast method

– Ops like it!

• Currently used for monitoring– No difficulty to

measure Q’ with octupoles

V.R

an

jbar

0 200 400 6002

1

0

1

2Head minus Tail position

turn numberm

m DXi

i

.


Special BPMs: Beam Line Tuner

K * (A - B) /(A + B)

Beam

A

B

StriplinePickup Position

PhaseAmplitude

Tuneetc

Signal Processing

Analysis

•Measure turn-by-turn position at injection

•Extract betatron oscillations, do closure for next injection

– Reduce emittance dilution from mis-steering

– Calculate expected emittance dilution

•Also determine tune, coupling, energy & phase mismatches at injection

•Also for post-mortem of lost stores– Stop continuous data acquisition on

abort– Look for signs of instabilities in “last

second” buffer

• Tevatron Stripline parameters– 1 meter long (< 1/4 )– ~30 dB directionality– Pickup gap = 83 mm– 0.65 dB/mm Sensitivity– Located near F0 for

maximum proton and pbar separation (in time)


Digital Receiver BLT

• Of course, it measures only DIPOLE oscillations


Instruments under developments, but first results are interesting:

IPM shows significant (>30%) quadrupole oscillationswhile OTR claims no size change over 1 turn whatsoever

17.5 20 22.5 25 27.5 30 32.5 35turn #

1.1

1.2

1.3

1.4

1.5

1.6

smrezismm

OTR, vert

Turn #1

IPM, vert

Turn #2

… just one example of importance of having several instruments for cross-checks


LIST OF INSTRUMENTS

BLT+dEmm+”last sec” RM BPM+related AJ Shottky I + II+ III+TuneTrk AJ SL+AGM+FWs VS IBEAM+SBD+FBI RM Head Tail Monitor VS Vacuum+RGA RM HLS VS RWM+LPM AJ Luminosity+IP diagnostics VS Software (C49, D44, SDA, ArrView, DLPlotter, OAC)

RM


Longitudinal Phase Monitor

sin

cos

FPGAADC

Q-10bits

I-10bits

x

x

ADC Clock

Gate Timing

Gate Timing

Beam Pick-up

(Stripline)

(SIMPLIFIED)

T

T

RF

T

T

RF

RF

RF

dtttf

dtttf

dtttf

dtttf

)sin()(

)cos()(

)sin()(

)cos()(

arctanˆ

A.I

bra

him

Linearity:

Simulated phase:

Measured phase*:

*Discrepancies explained by difference in RF voltage and shape asymmetry during acceleration


BPM Turn-by-Turn

1113 / 5 Artifacts* Betatron Lines

R.K

utc

hke e

t al

*Sampling frequency is 7/5 RF, so clock phase varies slightly with a period of 5 turns. Resolution: 0.3 mm/sqrt(Hz)


1.7 GHz Schottky measurements

Schottky Horizontal Sync lite HorizontalSchottky VerticalSync lite Vertical

#4058Note: Arbitrary scaling

Damper problems→ longitudinal blow up

#3989

NO fudge factors!!!

#4000RMS fluctuation around

trend line: 5 10-4

0.00

2 Pbar bunch #29 Intensity: 19 109

Emittance: ~17 mm mrad

Momentum spread

Emittance

Single bunch tune

SBDSchottky Horizontal Schottky Vertical

*Note that the 1.7GHz Schottky can not resolve the two normal modes of oscillation by frequency, hence it is insensitive to tune changes due to coupling.


Schottky development II

• In-house development

• Similar to 3D-BBQ

• Looks at both positive and negative peaks

• Feed-back to eliminate LF and increase dynamic range

Digital Beam Tune Monitor based on 16 bit 100MHz Digitizer

RF CLOCK,

RevMarker,TCLK,

FPGA100MHz

14 bit DAC

Cable from Tunnel

Ethernet, OAC

INPUTs: BPM Plates

TRACK/ HOLD

10Mbit ETHERNET

(A-B)- LPF(A-B)

TRACK/ HOLD

A+B

100MHz 16 bit ADC

64MSRAM

DSPSHARC

Input from Plate

~40ns Hold

A,B

A–B

100MHz 16 bit ADC

LPF(A-B)

+

- Gain100÷150

Reset

A.S

em

en

ov


Schottky Development

• CERN has provided Direct Diode Detection Base Band Tune (3D-BBQ) module

• By gating with RF switches, were able to separate proton and pbar signals

• Issue with 60Hz lines• Expert/Study tool for

the moment.

C-Y

. Tan


Resistive wall monitor

• Ceramic break with 80 120Ω resistors. Signals sampled at four locations are summed. Calibration signal can be injected.

• Two monitors: one for FBI/SBD, and one for general use

• During early 36x36 operation, ferrite inside vacuum heated up and started outgassing. Fixed with a new type of ferrite

• Recently intermittent problems with some resistors. All swapped out.

old ferritenew

ferrite

B.

Fellen

z


Longitudinal Phase Monitor II

• New development base on the same board as baseband Schottky.

• Yet to be tested.

2 Channels Longitudinal Phase Meter Based on 16bit 100MHz Digitizer

8

4

4 i

i

i

i

SiSi

8i

i

Si

FPGACYCLONE

Cable from Tunnel

Ethernet, OAC

INPUTs: BPM Plates 5MHz

GaussFilter

10Mbit ETHERNET

5MHzGaussFilter

100MHz 16 bit ADC

64MSRAM

DSPSHARC

RF CLOCK,

RevMarker,

TCLK,

100MHz 16 bit ADC

RF Refere

nce

~20ns

Phase

~100ns

ADC Sampl

es 18.8ns

A.S

em

en

ov


Special Equipment on Low-Beta Quads

Each low beta quadrupole is now equipped with :• Tiltmeters with ~1 urad resolution (above)• Water level sensors with ~0.2 micron resolution (right )

J.V

olk

T.J

oh

nso

n

Hydrostatic Level Sensors on Low-Beta quads and Alignment data report that CDF detector is

continuously sinking by ~0.5-1 mm/yr, distorting

vertical position of ~dozen low-beta quadrupoles

Date post:	14-Jan-2016
Category:	Documents
Upload:	kendal-mullinnix
View:	216 times
Download:	0 times

Tevatron Beam Diagnostics: Boosting Collider Performance Vladimir Shiltsev, Ron Moore, Andreas...

Documents