Routine plots (plots we should make for every hybrid; we should QA other hybrids against these results)

Interesting stuff

Calibration circuit noise

With calibration enabled, some channels show a non-Gaussian distribution of noise - the sampled ADC value splits into "upper" and "lower" peaks.

Some channels are affected more than others - some dependence on cal group, but varies widely within a group. This is the average RMS value of every sample for a given channel:

We cut on the second sample of channel 32 being above or below a threshold. We see that other samples and other channels get cut in the same way (including channels not being calibrated), so the noise is common to all channels.

Without cut:

Events with sample below threshold:

Events with sample above threshold:

With CSEL set to give no calibration pulses, we still see this effect:

The sequence of "upper" and "lower" events appears to be random.

Calibration circuit crosstalk

If one channel group is being calibrated, other channels see crosstalk at ~1/10 amplitude, opposite polarity. This is visible in the plots above.