Dear Detlef,

Below I summarize our joint visit with Bob Meller (216 Wilson, 5-7418) and
my hallway conversations with Dave Rice and Mike Billing (all of LNS)
regarding the details needed for the G-line bragg switch proposal.  Please
add your comment/changes to the material below and then send revised
version to us all.  Thanks, Don

******************

CESR bunch pattern

CESR has a circumference of 2.56 microseconds and space for 183
buckets.  2560nsec/183=14 nsec at closest spacing.  A period of 1/14 ns
implies a frequency of 71.4 MHz, a frequency derived by tripling a 24 MHz
master clock in Room 101 of Wilson (source: Bob Meller).

Mike Billing and Dave Rice summarized the current 9 by 5 pattern as 5
bunches in a row with 14 ns spacing between bunches.  The next train of
bunches (compared to the lead bunch in the first train is either 20 or 21
of the 14 ns periods away.  The structure of the bunch trains is 3 times
repeated of (20,20,21) spacing (time delays of 280 or 294 nsec each). The
longitudinal bunch length is 65 ps (sigma).   According to Mike Billing,
synchrotron oscillations at 20-30 kHz rate cause a 1 to 5 ps jitter about
the nominal position.
 

Timing Electronics

Bob Meller offered to give us connections to signals in Room 101 that would
bring the quality 71.4 MHz timing signal from which to derive the actual
trigger signal to a fs laser.  The fan out cards are in VME and feature a
CMOS driver compatible with TTL loads and capable of slewing at 1 ns/ 5
volt (about 50 of them in use).  [Bob himself has designed and built the
cards.]  A second enable line with the bunch pattern (generated from a CESR
loadable bit pattern) can be transported on an ordinary coaxial line to
gate the higher quality signal from the 71.4 MHz clock.  We will need to
provide the delay circuits to synchronize the two sets of signals.  Bob
suggested using 1/2" OD semi-rigid coax as has been used on some of the
critically timed signals to the RF system, etc.  I spoke with John Barley
of the CESR electronics shop who said the cable is probably RG214 and with
Margie Carrier who showed me a recent PO for this cable of $579 for a 500
foot roll.  Using G-line drawings, I am estimating the cable run from Rm
101 to G2 to be between 200' and 250'  - but plan to buy the 500' roll to
be sure.  This cable need type N fittings.  A reasonable selection of
fittings are carried in the LNS stockroom.  Bob will need some lead time to
prepare a fanout card when the time comes.
 

Fast detectors

We also need a fast x-ray sensitive photodiode to experimentally measure
the pulse structure at G2, a fast triggerable storage scope, and quality
delay circuits.  I have sent Denny Mills an email request for fast diode
information.

[His response today:  Don: I'm not sure what you mean by
switch (detector?). If you mean detector, 20 psec rise time is tough.
However, if you do not need the rise time but rather just a signal with 20
psec jitter that may be possible with a zero crossing timing SCA and a fast
APD (maybe). Denny Mills].

I also just dug out a copy of the
transparencies of Alfred Baron presented at the Detector Workshop in
Washington on 30 October 2000 where he says with avalenche photodiodes:

"20 ps time resolution demonstrated with visible light in Geiger mode (Cova, et
al., RSI 60 (1989)1104).  Same should be possible with x-rays".

This approach could be tried as it would simplify life if it worked!  We could
also try to call Bob Austin (at Princeton about 15 years ago, now?) who
supplied Denny with GaAs fast diode used to measure CESR bunch length.

As to the earlier timing circuits that Denny formerly used - they are all
disconnected and replaced with higher quality parts according to Meller.

Also, I did learn of some significantly related work.  Robert Holtzapple
(due to leave to SLAC at the end of January to work on PEP II) has been
operating a streak camera in the L5 North Area of CESR.  Bob Meller has
provided him with timing signals as explained above.  Bob has managed to
successfully trigger his streak camera from similar timing signals and is a
lot further away from the oscillator source in Rm 101 (maybe 300 to 400'
away).  I have a very draft copy of a paper "Measurements of the
Longitudinal Coupled Instability at the CESR Storage Ring".  He gives
evidence for beam-beam interactions spoiling the longitudinal arrival time
of bunches.  His measurements shows + or - 5 psec jitter with the
longitudinal feedback on during collisions (normal operations).  The jitter
increases from 5 to 30 ps if the instability threshold is exceeded (180 mA
in one beam for the example given) and the feedback is turned off.  This
paper is being polished up and he will send me a copy of the paper when
completed.  I asked if there were any information available about
transverse bunch motion, but he would have to add a lens to the streak
camera to look for such an effect - so no information is available here.
 

In summary, the prospects of getting an appropriate timing signal are
excellent.
***************
Don Bilderback  Email Address: dhb2@cornell.edu
CHESS (Cornell High Energy Synchrotron Source)
281 Wilson Laboratory           Office Phone: 1-(607)-255-0916
Cornell University              Fax:  1-(607)-255-9001
Ithaca, New York 14853, USA    CHESS home page:www.chess.cornell.edu