To examine the BSMD relative gains for the 500 GeV run I used VPDMB triggered events from fill 10471.

I start with a 2d histogram of zero-suppressed ADC values, which have been both pedestal-subtracted and had offline pedestal corrections subtracted from them.  Each strip is then checked to make sure that it has good status: if not, it is skipped.  Then the fitting range is determined by starting at the end of the ADC distribution and integrating up until the integral from the end equals .0000475*the total number of events for eta strips and .000045*the total number of events for phi strips.  The point at which this occurs is the upper end of the fitting range, xhigh.  The lower end of the fitting range, xlow, is found by integrating until the integral from the end is .00031*the total number of events, unless xlow is less than 5*pedestal RMS, in which case it is set to 5*pedestal RMS (in order to avoid having xlow be on the pedestal part of the ADC distribution).

Figure 1: The ADC vs softId distribution for module 119.  The black lines at ADC=50 indicate bad channels.

Figure 2: Fits for four eta strips from module 119.  The green lines show three and five times the pedestal RMS.

Figure 3: Slope vs. strip id (from 0-150) for eta and phi strips in this module.

Attached are two pdfs.  The first has several summary plots: the ADC vs softId and slope vs. strip id plots above, as well as slope vs eta and phi.  The second has individual strip plots for every strip in module 119, as well as several summary plots for each plane, showing the distributions of xlow, xhigh, xdiff (xhigh-xlow), the number of channels with zero entries between xlow and xhigh, the pedestal rms, the number of strips that have xlow set to 5*RMS, the slopes, and the relative errors.