As I have mentioned before, we now have calibration for the Au+Au 54 GeV from 2017.  Thanks to Xinyue Ju for helping extract the calibrations and to Prashanth for getting us set up on the STAR database.  I have implemented the calibrations into StEpd (manual here) so you get the calibrated number of MIPs passing through a tile with the following method:

StEpdTile::nMIP();

The picoDST infrastructure has been updated to include these calibrations, and thanks to Yang Wu and Isaac Upsal for making the calibrated picoDsts.  Yang has put 1000 of them (about 800k events) on rcf here: /star/data06/EPD/ywu27/pico/picoDSTs_new/   You can quickly download them to your laptop and analyze them there with StEpd.  The manual has examples, and there is a nice StEpdExample.C in the Macros area of StEpdMaker.

I took just a quick look, and at first glance, with nearly no real effort at all, things look surprisingly good.




Figure 1 - Histograms from 800k Au+Au events.  From left to right, top to bottom:  ADC spectra for PP04TT01 and PP05TT05, followed by nMIP distributions for the same.  Vz and RefMult distributions.  RefMult-vs-EPDmult for collisions with |Vz|<50 cm and <10 cm.  eta-phi distribution from the EPD; note the non-sharp features.  Eta-versus-refmult for EPD, followed by two plots of eta-versus-RefMult for TPC and EPD for collisions with |Vz|<50 cm and <10 cm.




Figure 2.  Left panels show EPD multiplicity and RefMult on the y- and x-axes.  Right panels are the eta distribution from TPC and EPD as a function of RefMult.  Top panels are for primary vertex |Vz|<50 cm; bottom are |Vz|<10 cm.  Note much sharper acceptance cutoffs for the smaller range in Vz.



Figure 3.  Pseudorapidity distributions from TPC and EPD for collisions of various centrality, as determined by RefMult.  On the left (right), |Vz|<50 cm (10 cm).


It looks great except something is clearly wrong.  The scale on the eta distributions is at least too high by a factor of 4.  (In fact, for the EPD, it is too high by at least a factor of 16, since I should multiply the above values by four, since we use a quarter wheel.)  While it is not complete, it is worth discussing.  The plots above were made with a macro you can have here.  It is really REALLY quickly done and normally I would clean it up before letting anyone see it, but hey, we're friends, right?  At least it shows how things work.

In particular, notice that this is how you fill an eta distribution.  It may seem strange and convoluted, but it's the right way.

    TVector3 PrimVert(pico->Event_mPrimaryVertex_mX1[0], pico->Event_mPrimaryVertex_mX2[0], pico->Event_mPrimaryVertex_mX3[0]);
    TVector3 TrackEnd = Epd->GetTile(-1,PP,TT)->RandomPointOnTile();
    TVector3 StraightLine = TrackEnd-PrimVert;
    Double_t eta = StraightLine.Eta();
    Double_t phi = StraightLine.Phi();
    h_EpdEtaVsPhi->Fill(phi,eta,Epd->GetTile(-1,PP,TT)->nMIP());