In order to reduce the cable path lengths for the pion detector, Jean-Francois, Fang, and I did some surgery on some of the "MD event-mode spare" cable paths. The net effect is that the spare cables 45,56,67,78,81 all come off the Ferris wheel near the octant 67 patch panel, which will be convenient for putting the pion detector in octant 6 or octant 7 with short (~15') patch cables.
Attaching a pulser via a many-way tee to md6+ and all the pion detector channels, we observe the following cable-only relative timings upstairs:
md6+ 0 ns wilbur -5.2 ns templeton 3.5 ns fern 16.8 ns avery 16.8 ns zuckerman 14.0 ns
We have matched the timing of the pion detector triggers to the MD self-trigger. During a pulser-triggered run using the existing timing on MD6+, the "md" pulse arrives 36 ns after the start of the gate. (The minimum for these QDCs if 15 ns after the start of the gate; this comes later because it's always the second MD PMT in a coincidence that generates the trigger, and the transit time for light across the bar is long.) The pulser pulses coming up the fern/avery cableway arrived at the QDC 53 ns after the start of the gate. This is what we tried to duplicate.
There was a very low rate of wilbur-templeton coincidences, so to time the "pion coincidence trigger" we used fern-avery coincidences at about 60 Hz. The delay module was too long at its minimum setting, so we made a direct cable connection from the coincidence module with a coil of cable.
Now that we have discussed it, though, we didn't quite do this timing right: we didn't account for the time difference between the w-t and f-a coincidence. Also, an event near the middle of a bar, the MD light propagation will make the MD signal come a little later than the w-t coincidence (rather than at about the same time). I think that both of these mean we want more delay between the wilbur-templeton coincidence and the QDC gate, but I'm too tired to be sure. It's probably best just to wait and check this with beam, and there is a reasonable amount of headroom for under/over-delay.