New Booster Board

Created May 1, 2004

Last updated May 1, 2004

The Booster board, built by the UCSD Physics Electronics Shop, is a huge improvement over the first version, incorporating lots of bells and whistles. Unfortunately, it whistles. The functions of the current board are:

• Multiplies 10 MHz sine wave from GPS XL-DC clock into 50 MHz sine
• Generates START and STOP signals according to TTL request based on 50 MHz source
• Generates alternate START that can be fed to lower header of TDC without explicitly crossing grounds
• Converts NIM signal from 9327 (deriving from photodiode) into an ECL START signal for TDC (intended to go to lower header on TDC)
• Generates random START/STOP pair with separation of about 12 to 112 ns to fill all TDC channels uniformly
• Random START alternately comes out on channel that can be fed to lower TDC header without crossing grounds
• Internal noise generator for random STARTs can be replaced with external pulse input.

This version of the booster board had various problems with noise coupling and ground bounce. For instance, the calibration START/STOP only behaves (i.e., 10 ps external jitter) when the 10 MHz feeding the random part of the board is disconnected. Also, the digital logic associated with the random START/STOPs produced much structure in the resulting histogram that completely swamped the TDC variation.

The next version of the booster board should have the following features to make it more immune to these problems:

1. 50 MHz sine should be promptly turned into a square clock and distributed to the various comparators via a low-jitter clock-distribution chip.
2. There should be a front-panel switch to select between primary STARTs (headed to upper header and TEST input on TDC) and alternate STARTs intended for lower header and individual channel inputs.
3. There should be a way to disable the random generator circuitry via front-panel switch and associated power relays. We want all of this to be dead as a doornail during normal operation.
4. It would be nice to have a second 50 MHz square wave output to be terminated in 50-ohms so we could do jitter measurements on the 1 GHz scope.
5. A re-design of the random START/STOP circuitry should be carried out to make it as completely immune to digital noise/ring as possible. Perhaps it should take a fresh approach altogether, ignoring my initial design scheme.

I've debated somewhat about separating the calibration and random functions onto two separate boards. But if we can alternately power one or the other, then we might stick with the economy gained by making a single board.

APOLLO To Do Task List.