SE_upate

[jingqiangye]

This PR is mainly for https://github.com/darkmatter-ucsd/SanDP/issues/9, https://github.com/darkmatter-ucsd/SanDP/issues/12 and https://github.com/darkmatter-ucsd/SanDP/issues/13 by the following ways:

1. https://github.com/darkmatter-ucsd/SanDP/issues/9, https://github.com/darkmatter-ucsd/SanDP/issues/12. Peaks after S2 split are clustered into S1 or S2 only based on their width above threshold from requirement in config file.
2. For https://github.com/darkmatter-ucsd/SanDP/issues/13, currently I simply switched the order, i.e: a) Merge s1s if they're close b) See if the merged s1s is s1 or s2 based on width above threshold c) Ingore s1s if they're behind main (biggest) s2

Previously it's c->a->b.

[jingqiangye]

Some preliminary results:

1. We can see the difference by the waveforms. For one particular event, event 67 from 20190530T201235, let's see the event waveform before this update.

In the waveform below, a large s2 triggered DAQ, however at later stage another even larger s2 occurred and became the main s2. We have observed a few events like this in SanDiX, the reason is still unknown. But clearly there's no small s2s after the main s2, as mentioned in the related issues. This problem will be more severe for those events where main s2 triggered DAQ.

Before:

After the change, there's a small s2 identified after main s2. That's great! And if we take a even closer look at the s2, it really looks like a single electron pile up event. After: Zoom up:

2. I processed one data and compared the s2 spectrum (all s2s), see below. Clearly after the update more small s2s are identified, and when s2 is larger than 150 PE the two spectrums become similar because s1-merged s2 usually stops there.

[jingqiangye]

As I mentioned in a new issue https://github.com/darkmatter-ucsd/SanDP/issues/15, the current s2 width requirement could be too high, thus I lowered that down to 10 samples (40 ns) now. The result is much better now.

The single electron peak is visible now, see the plot below. The SE gain looks reasonable. (under 5 kV anode)

Also we found the single electrons from cathode, which is around 40 us after the main s2.

[jingqiangye]

Unfortunately 40 ns s1 width seems to small for S1s, we lost s1 efficiency when it's larger than 5 PE, see the plot below. There're many lone s2 in the data now. Need to make requirement a bit looser.

[jingqiangye]

A test with 80 ns width requirement on S2. Looks like most single electron events are NOT identified. Definitely we want to keep the S2 width requirement small, and in the meanwhile use rise/drop time to further differentiate S1 and S2.