ESA Bunch Length Measurement Hardware

1
ESA Bunch Length Measurement Hardware

• Josef Frisch, Douglas McCormick, Sean Walston

2
Basic Operation

• High frequency (10-gt100GHz) radiation is
  generated when the beam passes a ceramic gap in
  the beamline.
• A set of detectors is used to measure the power
  in different frequency bands
• April Run
• 16GHz, 23GHz (cutoff frequency) diodes
• Signal transmitted through gt30M of WR-90
  waveguide
• Expect dispersion to broaden signal 100nm.
• 2X 100GHz (75-110GHz) diodes
• Signal transmitted through 0.5M of WR-10
  waveguide
• Expect dispersion broadened signal 1ns
• Future Runs
• Will add a pair 220-330GHz diodes.
• Will add broadband pyroelectric detectors
• Lower sensitivity
• Can arrange for LF cutoff with circular waveguide.

3
Processing Electronics April Run

• Signals large
• Needed 20dB attenuation on 16 and 23 GHz signals
  to bring diodes into linear range
• Removed horns and moved waveguide 10cm from
  ceramic gap for 100GHz detectors.
• Note signals should be adequate for Pyro
  detectors will try in July.

4
Raw Signals (5Gs/s Scope)
16GHz
100GHz
80ns/div 5mV
20ns/div 10mV
23GHz
100GHz
80ns/div 5mV
20ns/div 20mV
5
Raw waveform comments

• 16 / 23GHz Long pulse lengths (50ns) not
  unreasonable considering long WR90 waveguide to
  diodes
• 100GHz, 7ns pulses too long to be due to
  waveguide dispersion.
• Suspect tail is due to mm-wave radiation
  generated downstream propagating backwards.
  (forward radiation would be prompt).

Note clear what part of the signal is ringing,
but pulse is clearly wider than the expected 1ns
from waveguide dispersion
6
Measurement Electronics
Gated integrator installed to allow 2ns gates
for expected signals from diodes Actual signals
20ns, so could use standard GADC with little
signal to noise loss DC output from gated
itegrators read by SLC control system SAM
7
Data taken over a shift with varying beam
conditions 100 GHz detectors do not track 16GHz
detector Measuring something (hope it is
bunch length)
Two 100 GHz detectors track with RMS error of
1.5 Still dont know we are measuring bunch
length, but at least the 2 detectors give
consistent results Non-linear behavior at low
signal not understood maybe due to unusual beam
conditions Mike Woods will show plots vs.
accelerator parameters
8
With fixed beam conditions (39.779 deg phase ramp
38.614 V) See good correlation, 100GHz detector
track to 0.6 difference RMS
9
Electronics for July Run

• Hope to add two 220-330 GHz Detectors.
• Power sensitivity similar to existing 100GHz
  detectors
• Reduced waveguide size WR3.4 vs. WR10 will
  decrease input power X10
• Horns should gain the power back if needed
• Adding diode preamplifiers (better match) to
  improve signal to noise 5X. (all detectors)
• Also planning on Pyroelectric detector.
• Noise 20X worse than existing diodes
• Should be OK for ESA, marginal for LCLS
• Frequency response from aperture cutoff (100
  GHz) to light
• Since pulses are relatively long, can use
  conventional. GADC for data acquisition.
• Ideally use multi-gate GADC if available (for
  ESA), otherwise use cable delays.
• For LCLS, could add a mixer for additional gating
  if signal to noise is marginal.
• GADC will give pulse to pulse correlation.

10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
Issues

• Component delivery Not clear what will be here
  by July
• Pyro-detector sensitivity With good
  pre-amplifier, sensitivity 100X worse than diode
  with matched amplifier.
• Electronics low noise charge amplifier for
  diode may not work on first try