search

LDMX-Software / Ecal

Software module for digitizing, reconstructing, and vetoing within the ECal.
1 stars 0 forks source link

Update Recon for v14 Geometry #43

Closed tomeichlersmith closed 2 years ago

tomeichlersmith commented 2 years ago

The silicon thickness changed! We need to propagate this.

tomeichlersmith commented 2 years ago

Changing that recon parameter did wonders for aligning the two geometries shower features. There are still differences that need to be understood, but v14 is reasonable with the recon updates on this branch.

avgLayerHit_.pdf deepestLayerHit_.pdf maxCellDep_.pdf nReadoutHits_.pdf showerRMS_.pdf stdLayerHit_.pdf summedDet_.pdf summedTightIso_.pdf xStd_.pdf yStd_.pdf

tomeichlersmith commented 2 years ago

The noise RMS is currently in units of mV:

https://github.com/LDMX-Software/Ecal/blob/a4bca7d0743adcd5ac079bb7bd0b976f450afb2f/python/digi.py#L45

Calculating this out and using the gain from the hardcoded conditions

https://github.com/LDMX-Software/Ecal/blob/a4bca7d0743adcd5ac079bb7bd0b976f450afb2f/python/ecal_hardcoded_conditions.py#L49

Gives a optimistic noise of ~0.6 ADC.

As pointed out in the comment on EcalHgcrocEmulator, the noiseRMS is calculated by estimating the number of noise electrons as a function of the readout pad capacitance. At zero capacitance, the noise is estimated to be 700 electrons and then 25 additional electrons per pF of readout capacitance. We could bump up the "average readout pad capacitance" from 20 pF to a higher number to try to account for the more realistic noise, but I think it would be better to fold the noise into the conditions. This will then allow us to have it vary by channel in the future emulations.

tomeichlersmith commented 2 years ago

Changing how the HgcrocEmulator functions means that this issue is promoted to the ldmx-sw space because it also affects Hcal.