I can safely say that my reconstruction code for the HiRes CAEN pipeline TDC's is working, and that the banks should be a good check on systematics.

Here's TAGTL and TAGR number of rows per event.

And, here's the number of TAGT_LR rows per event. TAGT_LR corresponds to TAGT. TAGT_LR is a software co-incidence, and TAGT is the hardware co-incidence

Looks ok, but then I looked at TAGT and TAGTHIT for the same data/events since I noticed far fewer TAGM_LR photons than TAGM...

This actually isn't a problem. Take look below.

TAGT rows/occupancy

Here we see the entire time window over which each TDC can measure. These plots are the reconstructed TChannel bins (TAGTHIT and TAGT_LR_HIT--where overlap channels are made)

The LeCroy is set to 16 microsecond, and the CAEN pipeline TDC have a 4 microsecond range.

This means that the CAEN TDC's do not measure 75% of the out of time events that the LeCroy TDC's do. This leads to a 75% reduction in TAGM_LR statistics, but the in time peak is untouched.

And, This reduction in out of time photons is noticed in TAGM_LR...

I'm confident that my code accurately reconstructs TAGT_LR, since the statistics in and the shape of TAGT_LR match up with TAGTL and TAGTR as one would expect.

After that, TAGT_LR_HIT (analog to TAGTHIT) and TAGM_LR are made with the well understood tagger library.

The last question to be answered is what timing window should we use to form the coincidence between TAGTL and TAGR.

The following data show TAGTLHIT-TAGTRHIT.

Are these humps at +-5 ns significant?

And, if we are happy with a +-4ns software coincidence window, then here's a study of the correspondence between TAGT_LR_HIT and TAGTHIT.

I binned time diffs as integers.