Scientific potential is reduced if 10MeV range doesn't provide good linearity in energy.

DNL of digitizers screw linearity, causing residue of >3keV for calibration function.

5/18/12: Before the electronics were turned off at Berkeley, we completed the required pulser measurements to extract absolute non-linearity measurements for all 1120 ADC channels. At this point, conversion from raw trace data to a useful non-linearity measurement is complete. David Radford's correction code, is being implemented the correction in offline Mode3 data-sorting code. Once successfully implemented, will discuss with Mario/Carl re: implementing in the pre-processor.

5/29/12: David's correction code, which takes into account the 'local baseline', the resting baseline (equivalently the slope the pulse is sitting on) and the flat-top, has been implemented in offline sorting, and shows an improvement in the residuals from rms = 0.7keV to rms = 0.24keV (inlcorrection-1109.pdf), for old (CR-5) 152Eu calibration data. The correction on the 10MeV scale channels appears to be working, at least at low rate, where the local and resting baselines are very close. The high rate correction still needs to be tested for the 10MeV scale, which will require taking the appropriate data. If possible, once I (Heather) arrive at NSCL, we can take the appropriate data and test the correction. The next step for correcting 10MeV channels would then be to incorporate into pre-processing code (needs Mario's help).

Corrections to the other gain CC channels has not yet been successful. The measured INL corrections do not show agreement with the residuals as was the case for the 10MeV full scale data, which may indicate a problem with the extraction of the INL – Heather and David will continue to look into this, and try to confirm the source of the problem. It is possible that correction to the 30MeV CC may not be possible without more INL measurements.

6/4/12: Offline correction (in Heather's unpacker) has been demonstrated for all 10 MeV CC in the MSU set-up, and appears to work for all channels except for 1 that is messed up. The correction code has been sent to Mario for incorporation into the pre-processor.

6/5/12: Offline correction now works all 28 crystals, and has been confirmed in 226Ra data from 500Hz/CC to 6kHz/CC.