CDC peak lengths
To find the peak (signal) length I have to decide where the peak ends and noise takes over.
Histogram of difference between adc sample value and previous sample value for first 190 samples in many events (trigger is approx sample 200 so samples 0-190 should be almost all noise).
Pedestal s.d is 11 (adc value) for these straws in this run. For other straws s.d. ranges from 9 to 13.
Peak length would be too long if the required decrease is too small and is caused by noise near the end of the signal window. Peak length=0 if it cannot be found (each decrease in adc value is smaller). These correspond to small signals (<200 adc value when requiring a decrease >15.) Required decrease>15 gives a good balance.
Max drift time (according to Garfield) is 660ns, 83 sample periods. Min peak length seems to be 5 samples so a max signal length of 88 samples is reasonable.