Log entry time 23:37:34 on April 24, 2011

Entry number 229987

Followups:

• 230050 04/25/11 08:10 gaskell Compton: may want to move Happex timing board

Right before run 21739 I changed the gate to a gate generated by the HAPPEx timing board (HAPTB). The gate is as close as it gets to what we were using with difference specified below. The differences are small enough that I don't expect any change in the asymmetry, so if there is a chance something is wrong. There will be a small change (see below) in the signal because of a small change in the length of the integration window itself. The length can be studied now.

The HAPTB is on vmec8 (the Moller VME crate). It uses the HAPTB library in ~/compton/libraries/vme/haptb, which is loaded on reboot. The default startup settings are what we want (for now). You can use VxWorks command line functions to set this to other values (see haptb source files).

• Run 21739 has the HAPTB gate going into the Compton DAQ but still uses the Compton DAQ gate generator for the integration window. This should be sensitive only to shifts in the timing of the integration window with respect to the actual helicity transition.
• Run 21740 has the HAPTB gate going into the Compton DAQ and uses the HAPTB integration window so the gate generator is bypassed.
• If both of the above look fine in first100k, I will keep using the gate from the HAPTB.

The goal of using the HAPTB gate is to be able to reproducibly change the gate around with software controls. We were using a gate generator with knob before, and that's not very reproducible (or reliable). Another goal is to have exactly the same gate for the Moller and Compton DAQ.

Differences are:

1. The HAPTB specifies the integration window and is triggered by the accelerator MPS signal. The time between the accelerator MPS signal and the beginning of the integration window used to be set by the gate generator knob, but is now specified as an integer multiple of 2.5 us intervals. This adds quantization, and shifts everything by something like 1.5 us. This could theoretically affect the latching, but we weren't latching this close to the edge of the integration window (that would have been horribly unstable).
2. Similarly, the length of the integration window is specified as an integer multiple of 2.5 us. Again, quantization will change the length of the integration window (but... see number 3).
3. Because of the way the HAPTB works, we need some arming time before the accelerator MPS arrives (approximately 30 us if I recall correctly, not in the manual, I mean, "manual"). We used to have an integration window up until the start of the accelerator MPS signal, so now we are stopping 30 us earlier, and our integration window is correspondingly shorter (30 us/900 us = 3% change in statistics).