BEMC Calibration

This page is intended to provide documentation of how I arrived at the numbers and figures relevant to the BEMC calibration quoted in the paper.  Data and macros can be found at

If you want to reproduce these plots, you're probably best off just downloading the whole directory.  Here's a tarball:


Nothing to see here just yet.  I asked Alex about the availability of the data and was told that they used the last dAu combined production.


Run 4 BTOW Calibration

Nevents = 20-57 M.  Run 4 is funny because we've reproduced P05ia since I did my analysis, so the old query I used isn't really valid anymore.  I took these numbers from Jamie's page at

The 20M number is the total number of trigId==15007 events.  These are what we used for the MIP calibration, but we also applied centrality and vertex cuts that aren't reflected in that number.  The number is also complicated by the warning not to use minbias triggers from productionLow || Mid || High before 5042040, which I did not heed (not sure if we knew about the problem at the time).  For the electrons we required trigId!=15203, which if I take all-bTid15203 from that link gives me 57.2M events.

No attempt has been made to figure out exactly what fraction of these data was actually analyzed in my jobs.  In Run 5 that number was 94%, here I've just taken it to be 100%.

Nmips = 1.4 x 10^{6}.  This is the number of entries in the "mH0" histogram in 2004_mips.root

Nelectrons = 1.5 x 10^{6}  makeElectronCounts.C The reason we have Nmips ~= Nelectrons is that the electron sample dropped the refMult and vertex cuts and took some triggered data as well.  Specifically, we used trigId==15007 for the MIPs and trigId!=15203 for the electrons.

Figure 2a:  makeMipFigures.C

Figure 4a:  makeElectronFigure2.C

Figure 5:  makeGainsFigure.C

Note that in each case you'll have to download the ROOT files required by the macro and put them in the same directory.


Run 5 BTOW Calibration

Nevents = 57M.  This one is a bit tricky, as I never saved a tree with all the events I looked at.  Here's how I arrived at the number.  I used a catalog query -cond 'production=P05if,filetype=daq_reco_mudst,storage=HPSS,filename~physics,sanity=1,tpc=1,emc=1' -keys 'sum(sanity),sum(events),sum(size),grp(trgsetupname)'

to get number of files and events for each trgsetupname I used in the calibration (ppProduction || ppProductionMinBias || ppTransProduction).  The query returned a total of 73716 files in HPSS.  I did save my original filelists, and when I cat them together I find that I analyzed about 94% of available files (69297 to be exact).  The event sample from catalog was 60.8 M events, and I just multiplied that number by the percentage of analyzed files.  This doesn't reflect the |vz|<30 cut that was applied to the MIP study.

Nmips = 4.3 x 10^{6}  This is the number of entries in "h" in 2005_mips.root.  All the cuts from Run 5 BTOW Calibration had already been applied.  You might have noticed that this number is much larger than the number of MIPs in Run 4.  The reason for this is that in Run 4 we used only minbias (trigId == 15007) and we also applied a centrality cut of 60-100% (refMult<57).  In addition, it's quite a lot harder to find an isolated track in a AuAu event versus a pp event, and Run 5 had a 3/4 barrel instead of just 1/2.

Nelectrons = 0.4 x 10^{6}  makeElectronCounts.C applies all the cuts and spits out the final counts for 2004 and 2005.

Figure 2b:  makeMipFigures.C

Figure 4b:  makeElectronFigure2.C

Figure 5:  makeGainsFigure.C

Figure 8:  makeElectronFigure3.C