From the analysis of charged-particle detection efficiency, the results were 96% when 2cm from the both edges of each counter was ignored and the two center counters were neglected. This analysis was from a pair production run. When the full width of the counters was included and the two center paddles were still dropped, the efficiency decreases to 78%. Below is a plot of the HYCAL x vs y distribtuions for the events that had good clusters but no coincidences with a Veto counter. No cuts were applied with the timing. The blue vertical lines are the limits in x which I applied for each Veto counter. These ranges were used to match a HYCAL cluster's x-position and the Veto counter's x-position.
The missed clusters appear to start at the counter's edge closest to the beamline. Furthermore, the range in x of the missed hits gets wider when the Veto counter is farther from the beamline.
My idea is that the particles are obviously traveling at an angle to the beamline and not parallel to it. The overlap of the Veto and HYCAL which I plotted above is if all particles were normal to the Veto counters. In reality, the particles can hit the edge of a Veto counter and hit a HYCAL element directly behind the neighboring Veto counter.
Here is my rough estimate of the distances from the target
Next, I applied the ratio of 730/690 to the limits of each Veto counter and get the coverage of HYCAl plotted below. The efficiency calculated with the new limits and not fiducial cuts on the counters is 96%.
When I look for coincidences bewteen HYCAL x-position and two veto counters(counter and its neighbor next over from the beamline), the efficiency increases to 102%.
That 102% efficiency includes the overcounting of out of time veto hits. If I remove the overcounting, the efficiency becomes 98.5%. The clusters which do not have any veto coincidences are plotted below. I believe these clusters leak through the gaps between veto counters.
For the runs where the low-intensity photon beam is incident on the counters, the beam is normal to the counters and the original limits are fine.