Hi Steve et all, On Sun, 08 Apr 2001 13:55:58 -0500 "Stephen J. Sanders" <ssanders@ukans.edu> wrote concerning ": question about recent mult changes": > Hi Christian, > I have been trying to obtain some direct comparisons between the NBI > centrality and that obtained by Hiro. I was able produce an ntuple with > both centralities early Friday, but I then decided to update the > BrRdoModuleMult program to use some revised calibrations that Hiro had > developed. When I went back to redo the calculation later on Friday, I > noticed you had modified a number of classes in the mult directory. I > downloaded the changes and did my replay. It's true I updated classes in mult, and I didn't announce it on the list, since the time was getting late and I hoped to announce it on saturday, but then I got tied up - sorry about that. So here's the list of changes: * BrTileRdoModule now uses the per ring energy to multiplicity conversion polynomials, as outlined by Steve and Hiro in BAN26. This change affected classes: BrTileRdoModule BrTileCalibration * BrSiRdoModule is know (more or less) fully implemented, following the algorithms of Steve and Hiro as previously done in BrRdoModuleMult. This change affected classes: BrSiDig BrSiRdo BrSiRdoModule BrSiCalibration BrSiParameters Please note, that the changes to BrSiRdo and BrSiDig means that it's not compatible with old version of those classes. As a consequnce, the class version of these classes has been bumped. Also note, that the class BrSiRdoModule is not fully debugged yet. Steve and Hiro, could you please take a look at the class BrSiRdoModule and see if I missed anything. * Some clean-up to BrRdoModuleMult. Nothing really of consequnce, just to make it easier to read. I'll bump the version number later today. > However, I am no longer getting a reasonable result for the NBI > centrality. Almost all of the counts are showing as 0 or 1, with a couple > of .05. Would your changes have affected the centrality calculation? What > I see could be consistent with your changing to the newer energy calibration > without making a corresponding change to the centrality class. However, > this is pure speculation. That could be the reason. However, looking over the parameter values from the previous calibration (one polynomial for all tiles) to the present one (one polynomial for each ring), the difference is very small, and mostly on the second or higher term, which are small anyway. Anyway, I'll look into it and change the numbers if needed. I'm inclined to do the centrality cuts in the total deposited energy in the array, rather than the multiplicity, since doing that we'll be independent of the energy-to-multiplicity calibrations, which, as far as I can tell are dependent on some event generator (HIJING so far). I believe using the deposited energy (after outlier corrections and so forth) should be ok, since we're after all looking at the integral rather than the actual number. Any thought on that is welcome. > ...steve > p.s. I stopped working on this about 5:30 pm my time on Friday. With AFS > down all weekend, I would not have incorporated any changes you might have > made after that time. I subitted late friday afternoon (DST CET) and have not done anything else since then. Yours, Christian ----------------------------------------------------------- Holm Christensen Phone: (+45) 35 35 96 91 Sankt Hansgade 23, 1. th. Office: (+45) 353 25 305 DK-2200 Copenhagen N Web: www.nbi.dk/~cholm Denmark Email: cholm@nbi.dk
This archive was generated by hypermail 2b29 : Mon Apr 09 2001 - 09:08:43 EDT