run11 dsmAdc vs Adc0 issue

Data are for BHT1 triggered events (dsmAdc > 18).
Distributions from embedding are for electrons from J/psi decay - no weight on input J/psi pT is applied.

 
Below is a comparison between data - left hand side - and embedding - right hand side.
Strange dsmAdc vs Adc0 band seen in the data.

BEMC towers from data:

Black - towers that fired the trigger, dsmAdc > 18, and for which dsmAdc < 0.1*Adc0 (strange dsmAdc vs Adc0 band removed)
Red - towers from the strange band (dsmAdc > 0.1*Adc0)


The issue seen only for some towers.

DsmAdc and Adc0 vs towerId: