Phase2 Tracker Premixed DIGIs are broken

[mmusich]

Two problems have been detected in the Premixing vs Standard Mixing validation for the Phase-2 Trackers:

Issue with Inner Tracker digis

Full details in this presentation from @suchandradutta:

https://indico.cern.ch/event/879033/contributions/3840120/.

In short:

• premixing stage1 uses a single electron to ADC (=135) conversion factor for both IT (inner tracker) / OT (outer tracker)
  

  cmssw/SimTracker/SiPhase2Digitizer/python/phase2TrackerDigitizer_cfi.py

  Lines 230 to 240 in 9abb3c7

  	# NOTE: It is currently assumed that all sub-digitizers have the same ElectronPerAdc.
  	from Configuration.ProcessModifiers.premix_stage1_cff import premix_stage1
  	_premixStage1ModifyDict = dict(
  	premixStage1 = True,
  	PixelDigitizerAlgorithm = dict(
  	AddNoisyPixels = False,
  	AddInefficiency = False,
  	AddThresholdSmearing = False,
  	ElectronPerAdc = phase2TrackerDigitizer.PSPDigitizerAlgorithm.ElectronPerAdc.value(),
  	AdcFullScale = phase2TrackerDigitizer.PSPDigitizerAlgorithm.AdcFullScale.value(),
  	),

• premixing stage2 (first step) uses again the same value for both IT and OT

  cmssw/SimGeneral/PreMixingModule/python/mixOne_premix_on_sim_cfi.py

  Line 227 in 9abb3c7

  premixStage1ElectronPerAdc = cms.double(_phase2TrackerPremixStage1ModifyDict["PixelDigitizerAlgorithm"]["ElectronPerAdc"])

• the final digitization step in stage2 uses instead 600 for IT:

  cmssw/SimTracker/SiPhase2Digitizer/python/phase2TrackerDigitizer_cfi.py

  Line 23 in 9abb3c7

  ElectronPerAdc = cms.double(600.0),

  while 135 is used for OT:

  cmssw/SimTracker/SiPhase2Digitizer/python/phase2TrackerDigitizer_cfi.py

  Line 105 in 9abb3c7

  ElectronPerAdc = cms.double(135.0),

As a result of the double (lossy) conversion performed with different conversion factors in the Inner Tracker there is a noticeable change in the final ADC distributions (red is Premixing, green is Std Mixing)

Few concrete examples for some SimHits are as follows:

Charge [e-]	Charge (ADC) (standard mix) e/ADC = 600	Charge [ADC] (premix step1) e/ADC = 135	Charge [e-] (premix step2) e/ADC = 135	Charge [ADC] (premix final) e/ADC = 600
3639.4	6 (6.07)	26	3510.0	5 (5.85)
8435.7	14 (14.06)	62	8370.0	13 (13.95)
9000.6	15 (15.0)	66	8910.0	14 (14.85)
4234.6	7 (7.06)	31	4185.0	6 (6.97)
13896.1	23 (23.16)	102	10037.7	22 (22.95)

This in principle can be recovered by allowing the collections of PixelDigis produced by the premixing stage1 and the signal digitization in step2 with two different conversion factor for IT and OT and then let the PreMixingPhase2TrackerWorker to undo the conversion separately for IT and OT.

Issue with Outer Tracker digis

In addition it was noticed that the fraction of OverThreshold digis in the premixed case for the PS modules is always 1, while it has a non-trivial distribution for the standard mixing case, see plot below:

As a possible solution we need to configure the digitizer to have analog readout for OT (at least for the PSS layers) at step 1 of pre-mixing, as the actual charge information is essential.

cc:
@suchandradutta @emiglior @skinnari @tsusa

[cmsbuild]

A new Issue was created by @mmusich Marco Musich.

@Dr15Jones, @silviodonato, @dpiparo, @smuzaffar, @makortel can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

[Dr15Jones]

assign simulation

[cmsbuild]

New categories assigned: simulation

@mdhildreth,@civanch you have been requested to review this Pull request/Issue and eventually sign? Thanks

[civanch]

assign upgrade

[cmsbuild]

New categories assigned: upgrade

@kpedro88 you have been requested to review this Pull request/Issue and eventually sign? Thanks

[kpedro88]

@mdhildreth @makortel if you have time, can you take a look? hopefully the first issue is easy to fix, the second one might take some more work

[mmusich]

a possible solution is at #29565

[kpedro88]

@mmusich excellent, thanks!