L-PID

PID issues:

1,I cut on a small spot on PRL2.vs.PRL1 to select electrons, and look at the L.cer.asum_c (blue line). 
  As you can see, there still have many low ADC channels in Cerenkov SUM, which means that from Cer, 
  there are still many Pions. After I cut on L.cer.asum_c at >300, where the single electron peak is, 
  the efficiency of electron already reduces to 97%.

2, I did the reverse way by cutting on high ADC channels of Cer Sum (4000~5900), which indicates that 
  most of them should be electrons. However, when I cut on the electron peaks of PRL1&2, there are only
  87% of electrons. It looks like some pions deposite very big energy in Cerenkov, which does not make 
  sence since their energy must be very huge to produce big signals in cerenkov PMTs.  

3, I look at the L.tr.beta (raw beta before timing calibration). As you can see, even I cut on PRL1.vs.PRL2
  to pick electrons, in the beta spectrum, the peak of pion is still very comparable to electron peaks.

4, After only applying cut on L.cer.asum_c>4000, which basically can remove most of pion, I plot of L.tr.beta,
  and still find that there is a pion peak.

5, Cerrenkov ADC vs PRL ADC