1. C/S-band cavity BPMs A. Aryshev (KEK), S. T. Boogert (JAI@RHUL), G. Boorman, F. Cullinan, J. Frisch, A. Heo, Y. Honda, J.Y. Huang, S.J. Hwang, N. Joshi, E-S Kim, Y. I. Kim, A. Lyapin, D. McCormick, S. Molloy, J. Nelson, Y.J. Park, S.J. Park,T. Smith,T. Tauchi, N. Terunuma, G. White. SLAC, KNU, PAL, KEK, JAI-RHUL, KEK, ATF https://www.pp.rhul.ac.uk/twiki/bin/view/JAI/BeamP osition

2. Summary of progress C-band Normal resolution with 20 dB 200 nm Best resolution recorded, 20 nm S-band About 1 um (with 15 dB loss in cable) Unstable when DR-RF ramp is on IP Commissioning phase Y resolution ~100 nm (needs good IP steering) Write paper on Hardware and initial results (YI Kim, Lyapin, Boogert)

3. Introduction Earthquake damage : relatively light Main change to network : made old VME controllers inoperative, unable to boot Hardware Exchanged : MVME167 to MVME3100 Integration of SLAC 16-bit digitisers Software upgrades : Over 40 large upgrades Automatic EDM Each polarisation has own reference, displays, processing, DAQ, etc Measurements and checks 714, C-band, S-band LO and CAL Digitiser noise spectra Analysis Calibration scale errors Multi-bunch

4. Hardware changes in detail Exchanged : MVME167 to MVME3100 Integration of SLAC 16-bit digitisers C-band CAL tone broken S-band sources, rationalised and checked S-band cables, reorganised due to new IP region shield penetration Checked out Zygo straightness monitor system

5. SLAC digitisers 16 bit 4 channels 120 MHz internal clock Excellent linearity Problem in impedance Appears to be 1 MOhm

6. SLAC+SIS digitizers Fully integrated 6 SLAC ADC cards 4 channel 16 Bit Bipolar High linearity Low phase noise (internal clock) Impedance issue....

7. SLAC impedance check IPY1 : Inserted

8. SLAC digitiser impedance Diagnostic check list Checked with multimeter (DC measurement) Triggered 25 MHz source Scope test at 50 Ohms Signal looks like it has no reflections Can be avoided in software even for references which have large Q compared with IPBPMs Check TOMORROW

9. C-band RF distribution 714 Input into RF locking box, generates 100 MHz digitiser clock LO LO has some sidebands Should check at down mixer electronics in tunnel Large amplification 5W

10. C-band cal tone Only visible damage due to quake Temperature interlock for C-band CAL tone 5W amplifier Fixed! Need to check CAL tone to MPIP (done!)

11. S-band system digitiser New MVME-3100 2 8-Channel SIS modules, enough for S-band system Clock and trigger still maybe an issue DR-RF on!!!!!! Propose to internally clock as have spare channels for multiple references now (IP BPMs moved to SLAC digitisers) Install TODAY

12. S-band signal cables Cut down cables Length reduced by approximately 20 m Attenuation similarly reduced Need for hybrids/amplifiers probably gone Will install hybrids BPMOld attn (dB)New attn (db) SF1FFx14.9 SF1FFy14.4 QF1FFx14.5 QF1FFy14.9 SD0FFx14.6 SD0FFy 14.7 QD0FFx14.7 QD0FFy14.8 REFS114.8 Measure TODAY

13. S-band sources 8 and 20 GHz Hittite sources LO and CAL tones Source power and frequency stability not such an issue but spectrum is important Need to propagate effect to down- converted signals LO 2kHz span CAL 8GHz 2 MHz span

14. S-band RF distribution At mixdown electronics LO looks reasonable Completely unlocked from 714 MHz of ATF, why is there such a coherent change in BPM performance (IQ rotation) when DR-RF ramp is switched on

15. Zygo straightness system Email from Mike Hildreth Check Laser (ok) VME system (ok-ish) Computers (ok) Laser path needs realignment Preparing for August planned trip by Mike H.

16. Software changes in detail Main change use to IP region BPMs Previously code used one reference cavity per dipole PAIR Totally automatic system configuration Generates all files (displays included from one master control file) More stable database, rarely have to change main DB configuration even if hardware changes drastically Complete system configuration in database Determine how the BPM system is operating, what is connected, how it is being processed

17. Automation Completely automatic configuration. Edit signal file and all settings down to hardware, analysis and display changed

18. Digitiser noise Two types of digitiser Need to understand the effect of digitiser noise on BPM processing Two types of investigation Calibration tone (fit to sine wave) Data taken but need to process Beam off Done, results on right BPMDigitiserNoise [ADU] QD10Xx raw SIS (14 bit) 2.53 QD10Xx pro SIS (14 bit) 0.12 IPAx raw SLAC (16 bit) 7.44 (/4=1.86) IPAx pro SLAC (16 bit) 0.72 (/4=0.18)

19. Frequency shifts In normal operation often notice temperature shifts Appear as linear phase in down- mixed signal Temperature changes cavity size, hence frequency Need to temperature monitor entire BPM system? Possible? Resolution? Basic : -112 kHz/K Beam pipe-50-60 kHz/K

20. t0 resolution Beam arrival time (t0) Substantial jitter and then longer term drifts Phase correction algorithms depend on this number Measured in REFC1 diode detector Linear fit to rising edge of signal Better method? Charge [10^10]t0 RMS [ps] 0.3602 0.5160 0.833 1.032 0.3 x 10^10 e-

21. Calibration Errors due to Slow Orbit Drift and Random Beam Jitter Record ~1000 pulses of position measurements Smooth using moving average with flat window of length equal to calibration step length in pulses Fit points separated by window length against positions in calibration scan to determine fractional error on position scale Subtract smoothed signal to leave random jitter Calculate random errors analytically Subtract jitter for BPMs on movers then repeat analysis Frankie Cullinan

22. Results S/m Frankie Cullinan

23. 23 IQ : Without bunch subtraction Bunch No  (rad) | Δ  | (rad) 1 0.98778 2 0.328980.65883 3-0.597770.92671 Rotation Angle: o BPM is moved along Y axis at 200, 100, 0, 0  m positions respectively. o I & Q are calculated in a similar way as single bunch calibration, without any bunch subtraction. o With change in mover position, I&Q from 1 st bunch moves along a straight line in IQ plane passing through (0,0) o Steps along straight lines in IQ plane from 2 nd and 3 rd bunches shows the behaviour expected from a cavity BPM. o IQ signal from a bunch is polluted by the decayed signal from previous bunches. Nirav Joshi

24. 24 IQ : After bunch subtraction Bunch No  (rad) | Δ  | (rad) 1 0.98778 2-0.041821.029609 3-0.967940.926120 Rotation Angle: Bunch Subtraction: o i & q signals from BPM and reference cavities are subtracted separately before normalization. o Decayed i & q from previous bunch is subtracted from current bunch. o Subtracted signal is then normalized. o Phase difference between bunches becomes even. Nirav Joshi

25. Future work Cavity temperature monitoring Monitor temperature by frequency drift Ok for a phase detector but not dipole system Consider installing temperature monitoring on all BPMs Resolution of current system? High charge operation (question from Terunuma- san) Dipole saturation is always a problem Reference saturation is not an option (should not happen) Ordering 5 remote controlled C-band attenuators Understand entire BPM system error budget, statisical and systematic errors etc.

26. Summary System is working well, compared to pre-earthquake situation Many cases improved Significant software changes Expect some bugs, tried to remove most of them Improved maintainability Continue with pre-quake program Attempt to get high resolution operation for >1 week without calibration