Work on StSpaceChargeDistMaker since version 1.3:

• Use dE/dx StTpcdEdxCorrection class on real data, excluding dX corrections, to take advantage of wisdom learning in converting hit charge into energy (our proxy for ionic charge in the TPC)
  • Remove anode voltage gain compensation (handled in dE/dx tsspar corrections)
  • Remove anode voltage alive check (handled in dE/dx tsspar corrections, which kills rows if voltage is more than 100 below nominal 1170V / 1390V)
  • Assume drift length = half length, as an average, for dE/dx correction (don't really know z; RMS of error should be reduced from using random z)
  • Use latest dE/dx calibration (current work from Mike Skoby for Run 12 pp200 as pp500 is not yet available; database currently provides a copy of Run 11 AuAu200)
  • Exclude outermost 4 (instead of just 3) pads on each row to comply with dE/dx
  • Exclude rows with TpcSecRowB gainScale<=0 to comply with dE/dx
  • Exclude rows with tsspar gain<=0 to comply with dE/dx
• De-smear real data for wrong distortion corrections (wrong z => smearing in r,phi)
  • Smearing matrix determined from the same Monte Carlo used to find acceptance, then inverted for de-smearing
  • Separate for east and west
  • De-smear in both r,phi, and fall back to try just r if the r,phi matrices can't be inverted

Current results of 1/r^N fits for without de-smearing (blue) and de-smeared (red):
 

Update 2012-10-25: Final calibrations for Run 12 pp200 dE/dx are done, and the difference between the previous (partial calibration) result (red and blue) and final calibration (green and magenta) are very small, but this is still not actually a pp500 dE/dx calibration:

 

__________

What follows are animated gifs of the radial distributions and 1/r^N fits. The animation goes from lowest to highest luminosity of the selected runs, and then pauses on the highest before looping again.

Radial distributions without de-smearing:

Radial distributions with de-smearing:

______________

Comments:
The hope was that de-smearing would reduce luminosity dependence and the outliers, bringing back radial bins [120,130] and [190,200] in particular. Neither of these appear to be the case. A previous look at the possible impacts of using the wrong SpaceCharge correction did not pan out, but perhaps it is time to take another look with more dE/dx and de-smearing wrinkles ironed out now. We should also take another look at this as the available dE/dx calibration progresses.

-Gene