In an effort to tag "bad" APVs and electronic channels I have analysed pedestal and pedestal rms data from March 6th through April 11th.  For each channel, flagged with a GOOD status, on each day I filled:

I) a histogram of pedestals

II) a histogram of pedestal rms

If the status wasn't good then zero was filled for these channels. From these two histograms I extract four quantities:

A) pedestal mean for each channel (mean from histogram I)

B) the rms of the pedestal distribution for each channel  (rms from histogram I)

C) the mean pedestal rms for each channel (mean from histogram II)

D) the rms of the pedestal rms distribution for each channel (rms from histogram II)

Figure 1:  A versus electronic ID with C as errors (red) and B as errors (black).  I only show the plot for channels 0-1279 but the remaining channels can be found at: http://www.star.bnl.gov/protected/spin/rfatemi/fgt/PedMean_RmsMean_PedRms_#.gif.  Naively, if no systematic effects are present,  I think we should expect the rms of the pedestal distribution (B) to be comparable to the the mean rms of the pedestal distribution (c).  If B is larger than C it indicates that the pedestal move more on avearge during the runs than is statistically meaningful as determined by the mean pedestal rms. In this case the 5th APV chip looks bad.  Looking at additional plots you will see this position (#5) in the assembly is systematically different.

 Figure 2:  To see this more clearly I plot the ratio C/B versus electronic  ID.  Just another way of highlighting the same problem channels.  More plots at: http://www.star.bnl.gov/protected/spin/rfatemi/fgt/PedRMS_RMSMean_#.gif.  Everything > 10 is clearly wackadoo.  But even APV # 5 shows the same deviations.

 Figure 3: Deviations in the RMS of the pedestal RMS distribution also tags the same bad APVs:http://www.star.bnl.gov/protected/spin/rfatemi/fgt/RmsMean_RmsRms_0.gif.  APV five shows a large RMS distribution.

For Pedestal and RMS a channel is defined as bad for the entire run if:

• Pedestal Mean > 800
• Pedestal RMS/ Mean RMS > 2
• RMS of Pedstals RMS distribution > 2

APV is flagged as BAD if 33% of channels are flagged as bad for the run.  These apv are documented in BadAPV_33.txt.

For Status a channel is defined as bad for the entire run if:

• A channel has a bad status for > 50% of the run
• A channel flips status > 6 times

APV is flagged as BAD if 20% of channels have BAD status for the run.  These apv are documented in BadStatus_20.txt