To examine whether or not the appearance of band structure in the SMD response was correlated to the response of the towers overlapping with a given SMD strip, I filled a histogram of BSMDE strip id vs. ADC count only if one of the two towers it overlaps with had a response greater than 2.5 sigma above pedestal.  Then, as I did before (see this old blog entry) for each group of 150 channels, I plotted the ADC distribution, fit the region between 400 and 950 to a falling exponential, and plotted the chi2/dof of the fit.  Each bin i on the x axis represents the ith group of 150 channels, so the first bin is channel 1-150, etc.

The chi2/dof values are smaller than they were for the previous plots: here the average (neglecting the zero and -1 values) is ~.5, while in the previous non-correlated plots the average was ~2.5.  Note that a number of the chi2/dof values are zero.  In the case of bin 97, that's deliberate, as the actual value is ~80.  The others are zero because (for whatever reason) there are no ADC values above 200 or 300, or no entries at all in the case of bins 7 and 61.  Bin 10 has chi2/dof of -1 because there is only one entry between 400 and 950.  Note also that bins 91-105 have elevated chi2/dof values because they are taken from the module that had the old configuration under which the ADC valules maxed out at ~800.

The attached file smdAdcHists.pdf contains the ADC distribution for each set of 150 channels.

The data was taken from the following runs: 8340056,8347077,8354043,8361021,9003020,9009040,9020048,9042010,9042024,9043034,9044022,9045003,9046007,9047017,9047034,9047074, 9048006,9048041,9049012,9049053,9050022,9050088,9050183,9051011,9052001,9052063,9053030,9053036,9053113,9054005,9054039,9054088, 9055035,9056042,9056101,9057050,9057085,9059012,9059065,906002,9060091,9061005,9061060,9061080,9062031,9062122,9063032,9063140, 9064017,9064065,9065007,9066010,9066025,9066084,9067038,9068004,9068053,9068121,9069047,9070003