In this page, I studied the BBC hits when EPD tiles see nothing. We saw a smaller compression than we expected and by looking at the TAC vs ADC plots, we found the low compression has nothing to do with the out of time signals. Here, I looked at BBC Ring 1 when EPD tiles covering BBC PMT 1, 3, 5 don't fire respectively and BBC Ring 3 when EPD tiles covering BBC PMT 8 don't fire. In this study, I used 600k AuAu events and just like in the previous study, an unfired EPD tile means nMip<0.1 and a BBC hit refers to the hit with ADC>150. We also come up with some ideas which may (or may not) explain the low compressions. Following are the results, the numbers on the plots at the first two columns are the compression factors. They were calculated as (N_hits of the Tile i with smallest hits)/(N_hits of the tile at the opposite side of Tile i).

When I selected the EPD tiles which overlap with a certain BBC tile, I made the unfired region pretty 'safe'. When you look at the plots at the third column, you might think the BBC tile was not totally covered, actually it was very safely covered if you look at a larger version of the plot. However, the EPD and the BBC might be rotated a little bit, so the tile position we have might not be accurate. In order to rule out this effect, I made the unfired region one circle larger. As you can tell from the following results, the compression factor DO NOT increase much.


Another possible explanation is the geometric efficiency of the EPD (see Mike's post)

Besides the low compression, we also saw some strange patterns. For example the stair-shaped pattern on ring one and the drop on tile 2 (shown in the red circle on the above plots). When I looked at the BBC hits without any event selection, those patterns didn't show up so it might be caused by the cuts I put on events, which doesn't make sense to me.