The ratio used for these files was:
0.78*(cathode-30)/maxpix
Where maxpix is the height (with baseline subtracted) from the pixel row that sees the maximum signal. The pixel baselines can be derived from the data because for any single row there are events that happen elsewhere and result in no charge being collected. The cathode always sees the event so there is no baseline in the data. It is assumed to be 30 which is a typical baseline for the pixel channels.
The 0.78 scaling is based on the data from the alpha source. These events are expected to be at the cathode so they should have a ratio of one. The scale correction may be correcting for differences in relative amplifier gain or may be due to some charge being lost to other pixels etc.
The black lines show the expected spread due to electronics noise. The lines show the effect of a height change of +- 5 on cathode and anode. The noise level on the pixels is approximately +- 5 based on the half max of the baseline peak. Note that the error effect looks asymmetric at low energies. For example, if the pulse height is 25 and the ratio is 1, the error lines are:
(25+5)/(25-5):1.500
(25-5)/(25+5):0.667
The alpha source shows more spread than predicted by electronics noise but the spread decreases with energy as predicted.
These plot shows calculated depth vs. max pixel pulse height for Cs137 using the EV-11 10mm detector. The top plot shows a run with the source on the cathode side. The bottom plot shows a run with the source on the anode side. Both plots have events with negative strip heights removed.
