[tubagundem]

The common mode is created by the charge on the pads of a GEM stack. Through capacitive coupling it is induced, with the same amplitude, on all pads of that stack. On each pad it appears as a negative baseline shift. The baseline cannot go below 0 ADC, so the shift that pad j can show is limited by its own pedestal:

CM_induced   = sum_i(q_i * c * k_i) / N_pads      (unchanged)
CM_seen(j)   = min(CM_induced, pedestal_j)        (new)
CM_output    = mean_j CM_seen(j)                  (new)

• CM_seen(j) is passed to the digitization of pad j instead of CM_induced.
• The CM output container stores the stack average of CM_seen. This is what
  the data CMV measure. The sum uses double precision because it adds up about
  3000 pads per stack.
• The flag TPCEleParam.doCommonModePerPadSaturation (default true) switches
  it on. The pedestals come from CDBInterface::getPedestals() and are stored
  in DigitContainer::updateCalibrations() as the 4th per-pad parameter.