1

FONT IP Feedback System

Implications of increase of L*

Philip BurrowsJohn Adams Institute

Oxford University

2

Outline

• Reminder of system concept

• General considerations

• TDR design

• Implications of changing L*• Technical issues + summary

IP beam feedback concept

Last line of defence against relative beam misalignment

Measure vertical position of outgoing beam and hence beam-beam kick angle

Use fast amplifier and kicker to correct vertical position of beam incoming to IR

FONT – Feedback On Nanosecond Timescales:Robert Apsimon, Neven Blaskovic Kraljevic, Douglas Bett, Philip Burrows, Glenn Christian, Christine Clarke, Ben Constance, Michael Davis, Tony Hartin, Young Im Kim, Simon Jolly, Steve Molloy, Gavin Neson, Colin Perry, Javier Resta Lopez, Christina Swinson

4

Beam parameters

ILC 500 1000 CLIC 3 TeV

Electrons/bunch 2 2 0.37 10**10

Bunches/train 1312 2450 312

Bunch separation 554 366 0.5ns

Train length 727 897 0.156us

Train repetition rate 5 4 50 Hz

Horizontal IP beam size 474 335 40nm

Vertical IP beam size 6 3 1 nm

Longitudinal IP beam size 300 224 44 um

Luminosity 2 5 6 10**34

5

Beam parameters

ILC 500 1000 CLIC 3 TeV

Electrons/bunch 2 2 0.37 10**10

Bunches/train 1312 2450 312

Bunch separation 554 366 0.5ns

Train length 727 897 0.156us

Train repetition rate 5 4 50 Hz

Horizontal IP beam size 474 335 40nm

Vertical IP beam size 6 3 1 nm

Longitudinal IP beam size 300 224 44 um

Luminosity 2 5 6 10**34

6

General considerations

Time structure of bunch train:

ILC (500 GeV): c. 1300 bunches w. c. 500 ns separation

CLIC (3 TeV): c. 300 bunches w. c. 0.5 ns separation

Feedback latency:

ILC: O(100ns) latency budget allows digital approach

CLIC: O(10ns) latency requires analogue approach

Recall speed of light: c = 30 cm / ns:

FB hardware should be close to IP (especially for CLIC!)

Two systems, one on each side of IP, allow for redundancy

7

IP FB Design Status: ILC

Engineering design documented in ILC TDR (2013):

1. IP beam position feedback:

beam position correction up to +- 300 nm vertical at IP

2. IP beam angle feedback: hardware located few 100 metres upstream

conceptually very similar to position FB, less critical

3. Bunch-by-bunch luminosity signal (from ‘BEAMCAL’)

‘special’ systems requiring dedicated hardware + data links

8

ILC IR: SiD for illustration

Door

SiD

Cavern wall

9

ILC IR: SiD for illustration

Door

SiD

Cavern wall

10

Final Doublet Region (SiD)

11

Final Doublet Region (SiD)

12

QD0 – QF1 Region (SiD)

13

QD0 – QF1 Region (SiD)

14

Final Doublet Region (SiD)

IP Region (SiD)

IP Region (SiD)

Beamcal – QD0 Region (SiD)

IP FB BPM Detail (SiD)

Tom Markiewicz, Marco Oriunno, Steve Smith

19

Kicker BPM 1

Digital feedback

Analogue BPM processor

Driveamplifier

BPM 2

BPM 3

e-

ILC FB prototype: FONT at KEK/ATF

20

Kicker BPM 1

Digital feedback

Analogue BPM processor

Driveamplifier

BPM 2

BPM 3

e-

ILC prototype: FONT4 at KEK/ATF

21

Kicker BPM 1

Digital feedback

Analogue BPM processor

Driveamplifier

BPM 2

BPM 3

e-

ILC prototype: FONT4 at KEK/ATF

BPM resolution ~ 0.3umLatency ~ 130nsDrive power > 300nm

@ ILC

22

FONT4 latency

• Time of flight kicker – BPM: 12ns• Signal return time BPM – kicker: 32ns

Irreducible latency: 44ns

• BPM processor: 10ns• ADC/DAC (4.5 357 MHz cycles) 14ns• Signal processing (8 357 MHz cycles) 22ns• FPGA i/o 3ns• Amplifier 35ns• Kicker fill time 3ns

Electronics latency: 87ns

• Total latency budget: 131ns

23

ILC IP FB performance

24

Implications of increased L*

• Latency = electronics + beam flight time

• Electronics = 87ns (today’s technology)

• Simple model: BPM + kicker at L from IP

beam flight time = 2 L/0.3 ns

• For bunch-by-bunch FB want:

500GeV: 87 + 2 L*/0.3 < 554ns L < 70m

25

Implications of increased L*

• Latency = electronics + beam flight time

• Electronics = 87ns (today’s technology)

• Simple model: BPM + kicker at L from IP

beam flight time = 2 L/0.3 ns

• For bunch-by-bunch FB want:

500GeV: 87 + 2 L*/0.3 < 554ns L < 70m

1 TeV: 87 + 2 L*/0.3 < 366ns L < 42m

26

Implications of increased L*

• Latency = electronics + beam flight time

• Electronics = 87ns (today’s technology)

• Simple model: BPM + kicker at L from IP

beam flight time = 2 L/0.3 ns

• For bunch-by-bunch FB want:

500GeV: 87 + 2 L*/0.3 < 554ns L < 70m

1 TeV: 87 + 2 L*/0.3 < 366ns L < 42m

??: 87 + 2 L*/0.3 < 150ns L < 9m

27

Increasing L* from 3.5m to 4m

SiD Schematic(Door Closed)

2014.09.03 SiD MDI U. Tokyo T. Markiewicz/SLAC

HCAL Door Yoke PACMAN

QD0 Cryostat QF1 Cryostat

QD0 L*=3.5m

QF1 L*=9.5m

QD0 Service Pipe

FB Kicker

FB BPM

BeamCal

PolyCarbonate

LumiCal

W MaskBeampipe

ECAL

Movers

Beampipe Spider

Support

Bellows & Flange

28 of 51

Increasing L* from 3.5m to 4m• No impact if BPM + kicker stay (roughly) in same

places • If push QD0 back by 0.5m could move kicker in

front of QD0 (Glen says this has some attractive aspects)

• Reduces lever arm x2 (no problem)• Would need shorter kicker than TDR (no problem)• ATF2: 30cm kicker, 1inch diameter aperture

100 urad kick on 1 GeV beam

400 nrad kick on 250 GeV beam

1600 nm correction of beam at 4m lever arm

30

Increasing L* from 3.5m to 4m

• If push QD0 back by 0.5m and leave kicker behind QD0

• Could add a BPM on incoming beamline in front of QD0

SiD Schematic(Door Closed)

2014.09.03 SiD MDI U. Tokyo T. Markiewicz/SLAC

HCAL Door Yoke PACMAN

QD0 Cryostat QF1 Cryostat

QD0 L*=3.5m

QF1 L*=9.5m

QD0 Service Pipe

FB Kicker

FB BPM

BeamCal

PolyCarbonate

LumiCal

W MaskBeampipe

ECAL

Movers

Beampipe Spider

Support

Bellows & Flange

31 of 51

32

Practical considerations

Wherever they go, need to design:

•Mechanical integration of BPM + kicker into beamline

•Routing of cables

•Location, support + shielding of electronics

If downstream of QD0, check backgrounds

Larger distance from IP puts potential sources of broadcast RF further from detector

33

Example: ATF2 IP kicker

CLIC Final Doublet Region

34

CLIC Final Doublet Region

35

36

Conclusions

IP FB system can be adjusted to suit a slightly larger L* without performance compromise

Better to have kicker downstream of QD0?

Would be simpler to keep all hardware on same side of push-pull beamline split (presumably downstream side: need to locate electronics on detector)

37

Extra material

Stripline BPM resolution ATF2 stripline BPMs: single-pass beam, bunch Q ~ 1 nC

330 nm rms

39

New IP chamber installed summer 2013

JAI, KEK, KNU, LAL 3 cavity BPMs

Commissioning started November:alignment, BPM signals,beam jitter …

40

IP kicker Cavity IPBPM

FONT digital

FB

KEK IPBPMelectronics

FONTamplifier

e-

ATF2 ‘IPFB’ tests 2014

41

Layout with new IP kicker

42

June 2014 beam test results

Centre beam in BPM using mover optimise resolution

43

Centre beam in BPM using mover optimise resolution

June 2014 beam jitter

resolution ~ 60 nm

44

Incoming jitter ~ 200nm

June 2014 IPFB results

45

Incoming jitter ~ 200nm

June 2014 IPFB results

Corrected jitter ~ 87nm

46

Latest beam results from ATF2

• Beam size ~ 44 nm achieved (Kuroda)

• First attempts at stabilisation of small beam at nm level (ATF2 goal 2)

• Cavity BPMs with resolution ~ 60nm

• Working IPFB system stabilising beam to

~ 87nm (at BPM resolution limit)

• October: improved BPM electronics

(design 2nm?) coming from KNU

• FB studies ongoing …