OK, I looked at the events in the run 11186 burst in detail.
* You were right, there was a problem with the ADCZ-TDCZ cut: the
absolute value was missing! I don't know how this got lost: I found
the ABS there in some older test versions of the code. But anyway,
it's fixed now, and thanks for pointing that out.
* My guess that the problem with hit 6 was a crazy timewalk correction
in Calmod was correct. Here are the numbers:
Raw values:
TDC0H 2206
TDC1H 100
TDC0L 2216
TDC1L 3199
ADC0 244
ADC1 104
Timewalk calibration constants from Calmod (run 10866 constants):
TWCORR1 side 0 = 1510
TWCORR2 side 0 = -4716
TWCORR1 side 1 = 2404
TWCORR2 side 1 = -18900
Corrected ADC values:
ADC0-ped 0 = 244 - 117.75 = 126.25
ADC1-ped 1 = 104 - 91.61 = 12.39
Magnitude of timewalk correction:
TWcorr0 = 1510/sqrt(adc0) + -4716/adc = 97.03
TWcorr1 = 2404/sqrt(adc1) - 18900/adc = -842.45 <-- (clearly wrong!!)
Reconstructed times:
Time = slope*(tdc+twcorr)-off
Time 0 = .1674*(2216+97.03)-24.714 = 362.4881
Time 1 = .1634*(3199-842.45)-14.2537 = 370.8116
Position = -.5*(362.49-370.81)*29.98/1.61 = 77.5
A quickie fix is to put some kind of limit on the magnitude of any
given timewalk correction. I'll put this into the next monitor
upgrade if I don't come up with something better by then.
Meanwhile I'll update the VXMACB constants to Alice's latest set.
Kate.