Looking at runs 19061001 - 19061019, running on the .dat files, and using the epd stream trigger (hit in bbc e +tof or bbc w + tof) which had prepost -1,0,1 turned on I had the following results:

Small Size: drupal.star.bnl.gov/STAR/system/files/Adc_small_190610pp.pdf

Large Size: drupal.star.bnl.gov/STAR/system/files/Adc_big_190610pp.pdf

Example:

Figure 1: EPD ADC spectra for cosmic runs with bbc+tof triggers with prepost = -1,0,1

Using a threshold of 10 (which may be a little small) I found:
Number of hits in  prepost -1 = 446
Number of hits in  prepost  0 = 1035
Number of hits in  prepost  1 = 1489

No evidence on my side of Mike's hot tile ... (I even extended the range of the pdf to check: drupal.star.bnl.gov/STAR/system/files/Adc_small_190610pp_zoomout.pdf )

This actually looks a little like we are in the wrong prepost...  The prepost settings were also turned on for 19063026 - 19063058, which I will look at next.  This analysis has 13,739 bbc triggered collisions in it.

These looked rather similar, see:

Small Size: drupal.star.bnl.gov/STAR/system/files/Adc_small_19063pp.pdf

Large Size: drupal.star.bnl.gov/STAR/system/files/Adc_big_19063pp.pdf

Number of hits in  prepost -1 = 685
Number of hits in  prepost  0 = 1589
Number of hits in  prepost  1 = 2152

I wondered whether there was any way for the pedestal to be "fatter" in a different prepost (yeah, I don't know how this would happen), so I increased the integration threshold from 10 to 15:

Number of hits in  prepost -1 = 542
Number of hits in  prepost  0 = 1287
Number of hits in  prepost  1 = 1843

With pedestal subtraction and the fact that the pedestal width is fairly uniform, I summed over all channels connected to a tile, the results (in various formats, but the histograms are the same):

Figure 2: Sum of all the channels in a given prepost from day 63 in various plotting distributions.