Sphenix-calibration-l mailing list

[Ross Corliss <ross.corliss AT stonybrook.edu>]

Dear All, 

The attached minutes from Tuesday's meeting are now posted on the indico page.  Feel free to add or amend.

-Ross

Central Membrane (Sara):
Showed 0th-order CM matching.  All stripes have a stripeID, positions inside 
a stripe return that stripeID, positions outside do not return an ID.  
The CM stripes straddle the radii of TPC padrows, but this is not ~relevant 
because of the expected radial distortion due to the magnetic field.  Ross 
suggests we could modify this so they straddle after mag field distortion.  
Tom points out if we do this we should continue to instrument the full radial 
extent, since there is no penalty for pads that are not reconstructed.
Q:  How far along we we to simulating the CM hits?
A:  PHG4TpcElectronDrift (Henry's version) releases one electron from each 
end of the stripe, but this is not yet realistic.  Realistic would need to be 
a poisson distribution number of electrons randomly distributed along the 
stripe.
Q:  Would this need a dedicated/different cluster finder?
A:  Probably not.  Diffusion blur the stripes similarly to tracks near the 
TPC (add 400um diffusion + 288um square-wave width in quadrature), and the CM 
flash is attenuated to get ~2x the MIP most probable value for a track 
segment of that length.  Can add more light to system if needed.

Overview and CM distortion measurement (Ross):
Showed overview, reminder that we are now using the z-asymmetric 3d magnetic 
field.  This produces an asymmetry in the north and south halves of the 
detector that is not smooth, because at z=0 it jumps from drifting through 
one integrated field to the other.  Showed a slide explaining why the 
helicity changes from half to half, but the sign of the phi distortion does 
not.
Showed the lack of correlation between distortion at CM and charge integrated 
in a column from the CM point to the readout.  There is also no correlation 
between the integral of the fluctuation charge (integral of (single event - 
average event charge) ) and the distortion fluctuation.  There is, however, 
correlation between fluctuation distortion at the CM and at z=50.  This needs 
to be studied.
Q:  We need to create a flowchart of how to combine this information.
A:  Agreed.  Should assume the Outer Tracker in that flowchart.  Expect, 
roughly:  Measure distortion at CM, interpolate between that and mid-z 
measurement due to outer tracker, then interpolate between that and 
distortion=0 at z=readout.  (though this still neglects measurement 
timescales)
Q:  What is the cpu power we need for the distortion mapping?
A:  It is significant.  the largest step is the field calculation from the 
charge map.  We can calculate this, but expect many optimizations not yet 
applied.  Not sure what resolution we will need from this.
Q:  Granularity studies still need to be done.
A:  Yes.  

Comment on distortion map binning: 
Ultimately it would be nice to have the radial boundaries of the distortion 
maps match the padrows (or a subset thereof), because
- this is the natural segmentation for the anode currents (to which one must 
apply the charge-to-distortion model)
- this is the natural segmentation for the reconstructed distortions, and 
thus would make comparison easier.


Pull request (Henry, Ensemble):
We should create a DistortionMap module that is controlled via flags in 
fun4all - one to turn on, one to select source file (assume one set of maps 
in that file for now), flags to activate particular r, phi, or z distortions 
only
electron drift code should draw distortions from that module, no handling of 
histograms directly.
histograms can be pull-req'd to a calibrations repository.
Can generate the DistortionMap request once it is ready, and expect to 
iterate with Jin once the request is made.
Q:  Geant makes assumptions about what padrow a hit is on.  This is probably 
wrong already, and will get worse.  We need to look into that at some point.
A:  Yes, but let's wait until we confirm this code works as before, and we 
see it break.
Q:  We should apply the full distortions with this code and see the tracking 
break.

==========

Dr. Ross Corliss

If it's work hours, and I'm not in a meeting, feel free to look for me in my virtual office:
(my zoom office is out of order, alas.  Reach out, though, and I can generate a link.)

On Oct 5, 2020, at 2:11 PM, Ross Corliss <ross.corliss AT stonybrook.edu> wrote:

Dear All,

Per the doodle poll, it seems we can pick up more of the folks working specifically on the TPC distortion/calibration software if we shift to >>> Tuesday, 9am EDT <<<.  We will meet tomorrow at that time.

This shifted time doesn't affect the usual timing of the outer tracker discussion on alternate weeks.

https://indico.bnl.gov/event/9369/

https://stonybrook.zoom.us/j/94081538941?pwd=Y256N3ptUC9zV2M4cjNtUGgyWnJYdz09

-Ross

==========

Dr. Ross Corliss

If it's work hours, and I'm not in a meeting, feel free to look for me in my virtual office:

https://stonybrook.zoom.us/j/9710804637?pwd=ZmZ2SlRIMVFvUUJwbUkyOVNVTmE5QT09