ATF 2 NanoBPM Q BPM Electronics. Status Feb, 2006

1
ATF 2 NanoBPM (Q BPM) Electronics. Status Feb,
2006

• Mark Slater Cambridge
• Yury Kolomensky, Toyoko Orimoto UCB
• Stewart Boogert, Steve Malton, Alexi Liapine UCL
• Mike Hildreth Notre Dame
• Jeff Gronberg, Sean Walston LLNL
• Josef Frisch, Justin May, Doug McCormick, Marc
  Ross, Steve Smith, Tonee Smith SLAC
• Hitoshi Hayano, Yosuke Honda KEK

2
Requirements

• Primary
• 25 BPMs to be instrumented (50 channels)
• Sub 100 nanometer resolution single bunch.
• Existing SLAC/BINP NanoBPM has demonstrated 20
  nanometers.
• Large dynamic range (400 microns desired)
• Secondary
• Since absolute state-of-the-art performance not
  required, limit costs, simplify calibration
• Use technology consistent with future large scale
  production for ILC.

3
Basic System Operation

• Signals
• Use C-band dipole mode for beam position
• Use reference cavity of same frequency for
  normalization, and phase reference.
• Downmix signals to frequency which is easily
  digitized (25MHz).
• Double downmix for existing system, single
  downmix for ATF2 bpms
• Digitize at 100Ms/s, 14 bit .
• Reasonable speed / resolution / price point for
  present day digitizers
• Note System easier to diagnose with phase locked
  LO(s) and digitizer clock, but no fundamental
  difference in performance.
• Find signal amplitude and phase (complex
  amplitude)
• Digital down conversion and fitting algorithms
  both tried.
• Much discussion. Joe Frisch, Steve Smith strongly
  favor DDC.
• Normalize signal using reference cavity (correct
  phase)
• Find correlation between complex amplitude and
  beam position from known BPMs or correctors

4
Existing SLAC NanoBPM Electronics

• Dual down mix
• Mix to 476MHz, then to 26MHz
• Dual down mix allows use of wider (percentage)
  bandwidth filters. (20MHz at 6.5GHz is
  difficult).
• Filters used to reject out-of-band signals.
• Note existing system does not bandwidth limit
  amplifier noise sacrifice approximately 3dB in
  noise figure.
• Design error
• Could fix by adding filter after each amplifier
  but additional complexity.

5
(No Transcript)
6
(No Transcript)
7
SLAC / BINP NanoBPM Performance

• Use 3 BPMs in alignment frame
• Use outer 2 to predict measurements from middle
  BPM.
• Noise and stability are combination of BPM system
  noise and structure vibration and drift

8
Resolution
20 nanometer RMS noise, center vs. end BPMs (beam
motion 15 microns p-p)
9
Drift of 50 Nanometers over 1 Hour measured
50nm
10
Changes for ATF 2 Electronics relative to ATF /
BINP electronics

• BINP cavity BPMs with SLAC electronics have
  demonstrated 20nm RMS resolution, 50nm stability.
• Electronics bulky and expensive
• For ATF2 can tolerate slightly reduced
  performance.
• Have learned from SLAC / ATF design.
• Minimize use of narrow band filters
• Expensive, bulky
• Use PC board components
• Use Image Reject mixer to reduce noise
• Replaces filter
• Reduce power consumption for mechanical
  stability want to minimize heat dissipation in
  tunnel

11
Image Reject Mixer

• Standard method to reduce noise from image
  frequency that mixes to same IF.

12
Electronics Block Diagram
13
(No Transcript)
14
(No Transcript)
15
Calculated Performance

• Noise Figure 5.7dB
• Compare with approximately 4dB estimated for
  existing NanoBPM electronics.
• Note, existing mixers increase noise by 3dB by
  combining 2 frequency bands.
• If present system is noise limited, expect
  similar 25nm position noise.
• Signal to noise and Nonlinearity lt-67dB (ratio
  of non-linear power to full scale power),
  corresponds to approximately 150 microns peak
  to peak range at 25 nanometer resolution. (or 600
  microns at 100nm resolution)
• Power dissipation 3.5 Watts / BPM
• Calibration signal included to improve stability
  off center need to experiment to evaluate
  performance.

16
IF power Amplifier
IF preamplifier
LO in
C-band Amplifier
IR mixer IF 90 deg. coupler
Input
Anti-alias Filter
Calibration Coupler
IR mixer (C-band section)
Limiter (not installed In this board)
17
Measured Performance (bench test at SLAC)
PRELIMINARY DATA
18
Electronics Status Tests at KEK
Detection limit was -85dBm. Assume bandwidth is
the 40MHz bandwidth limit of the
board Corresponds to 13dB Noise Figure. Cable
losses were 4.6dB. Electronics noise figure was
8.4dB Measured dynamic range was
64dB Differences from SLAC measurements may be
due to definition issues.
(Graph from Y. Honda)
19
Board Performance Status

• Resolution Requirement was 100nm.
• No measured resolution with beam yet.
• Measured noise gives 30nm calculated resolution.
  (KEK)
• Linearity Need to correct 20 intensity jitter
  with 200 micron offset to 100nm. This is a
  linearity of 4001
• Measured linearity gt20001 at full scale (SLAC)
• 16001 dynamic range measured at KEK.
• Stability Not measured

20
Electronics Problems IF preamplifier failure

• Most channels failed over 2 week test!
• Same circuit used at DESY, but still 2 failures
  out of 80 chanels in 3 months.
• Possible problems
• Very tight tolerances on component placement
• New PC board layout
• Possible component problem
• Component from alternate vendor under test
• Also developing completely new design for IF
  pre-amplifier. (based on fast op-amps)
• Expect test by end of Feb 06.
• In principal better performance, better gain
  stability, reduce power dissipation

21
Electronics Problems Calibration Coupler

• Board includes stripline couplers for calibration
  signal an for LO distribution.
• Circuits did not work
• Discrete components may have distorted strip-line
  geometry
• Possible Blunder in design
• Consulting SLAC experts in RF stripline design
• This is in principal a straightforward design
  should be low risk for a good RF engineer.

22
Electronics design issues

• Saturated operation
• Most ATF experiments done with saturated cavity
  BPM signals.
• Electronics should be well behaved for signals
  20dB above saturation.
• Performance looks OK at first glance

Position Sweep Note, signal saturated by
16dB Peak amplitude from fit to decaying
exponental shown (from Mark Slater)
23
Calibration (means different things)

• RF Calibration of cavity, couplers
• Frequency, Q, coupling for each mode
• Should be stable except for thermal expansion
• Calibration of electronics
• Gain, phase shift
• Calibration of dipole phase / amplitude vs beam
  position
• Goal of install it and it works with no in situ
  calibration still difficult. In principal
• Calculate cavity response to beam
• Measure coupling, frequencies on bench
• Measure cable phase lengths, and electronics
  phase shifts
• Otherwise Need ability to make a known move of
  beam relative to the cavity
• Note movement resolution does not need to be
  small, just within dynamic range of BPM, and with
  accuracy better than desired absolute accuracy of
  BPM.

24
(No Transcript)
25
BPM Remaining Issues

• Mechanical stability may be critical. The LLNL
  and KEK BPM support frames are both the result of
  a lot of engineering.
• Significant mechanical engineering effort may be
  required to make use of lt100 nanometer resolution
  / stability.
• Work underway at KEK
• Calibration algorithm requires work. In principal
  can steer beam (or move BPMs) to find phase and
  amplitude corresponding to position but this
  has not yet been automated.
• Need to integrate signals with ATF control system.

26
What would change for ILC

• Very difficult to guess at future technology.
• At the moment, 100Ms/s, 14 bit provides
  reasonable performance and price
• Higher digitizer speeds do not simplify system
  unless we can directly digitize the C-band (or
  possibly L-band)
• This may be practical when ILC is built
• Alternately 16 -20 bit, 100 Ms/s digitizers may
  be available
• Present system pretty good
• 2X theoretical position noise
• Linearity limited by both analog section, and
  digitizer
• Working on tail of exponential decay gives
  extended dynamic range.
• Cost is quite low dominated by assembly, test,
  connector hardware
• Suggest present design as baseline, but as
  technology develops, be prepared to change
  technology.