STAR task force for evaluating tracking efficiency uncertainty

[Dmitry Kalinkin <dkalinkin AT bnl.gov>]

Dmitry Kalinkin writes:

> Robert Líčeník via Star-tf-trkeff-l writes:
>
>> Hi everyone,
>>
>> thanks to Petr, who produced MuDsts and picoDsts from daq files in both
>> SL12d_embed and SL13b_embed libraries, we can now conclude that the
>> differences between the official production and the production done by
>> Dmitry (and used for embedding in his study) is very likely due to the
>> difference between the two libraries. See attached slides for the
>> comparison plots. For simplicity, black and green points are from
>> SL12d(_embed) and other points are from SL13b_embed.
>> Would anyone happen to remember what changed between these two libraries
>> that might cause such a large difference in refmult and other track-related
>> quantities?
>>
>> Thanks,
>> Robert
>> _______________________________________________
>> Star-tf-trkeff-l mailing list
>> Star-tf-trkeff-l AT lists.bnl.gov
>> https://lists.bnl.gov/mailman/listinfo/star-tf-trkeff-l
>
> Hi Robert,
>
> I think I know the answer to this question. There is a difference in the
> PPV vertex finder in SL13b that was introduced for run12pp510
> production. I think it was called "W boosting", and the scope of the
> change was just to change the default parameters (like NHits/NHitsPoss
> cut) of PPV to boost vertex the finding efficiency. You can browse the
> relevant differences yourself:
>
> diff -r 
> /afs/rhic.bnl.gov/star/packages/SL{12d,13b}_embed/StRoot/StGenericVertexMaker
>
> There may be other differences between the libraries, after all, experts
> have unrestricted ability to tweak things in the core codes.
>
> In my early steps I was also trying to understand what was the practical
> difference between SL12d(pp200 data)/SL13b(pp510 data) and
> SL13b_embed(what is available). I've looked at the event by event
> differences of multiplicies of global and primary tracks. See slide 4 of
> https://drupal.star.bnl.gov/STAR/system/files/iu_local_meeting_2018_11_08.pdf
> I'm not sure if that contradicts your findings.
>
> Best,
> Dmitry

For the reference, here are PPV parameters when running BFC with SL12d_embed:

StGenericVertexMaker:INFO  - PPV::cuts
 MinFitPfrac=nFit/nPos  =0.7
 MaxTrkDcaRxy/cm=3
 MinTrkPt GeV/c =0.2
 MinMatchTr of prim tracks =2
 MaxZrange (cm)for glob tracks =200
 MaxZradius (cm) for prim tracks &Likelihood  =3
 DeltaY (cm) for BTOF local posision = 1.5
 Min/Max Z position for BTOF hit = -3 3
 MinAdcBemc for MIP =8
 MinAdcEemc for MIP =5
 bool   isMC =0
 bool useCtb =0
 bool useBtof =1
 bool DropPostCrossingTrack =1
 Store # of UnqualifiedVertex =5
 Store=1 oneTrack-vertex if track PT/GeV>10
 dump tracks for beamLine study =0

And with SL13b_embed:

StGenericVertexMaker:INFO  - PPV::cuts 
 MinNumberOfFitPointsOnTrack = unused
 MinFitPfrac=nFit/nPos  = 0
 MaxTrkDcaRxy/cm= 6
 MinTrkPt GeV/c = 0
 MinMatchTr of prim tracks = 5
 MaxZrange (cm)for glob tracks = 200
 MaxZradius (cm) for prim tracks &Likelihood  = 6
 DeltaY (cm) for BTOF local posision = 1.5
 Min/Max Z position for BTOF hit = -3 3
 MinAdcBemc for MIP = 8
 MinAdcEemc for MIP = 5
 bool    isMC = 0
 bool  useCtb = 0
 bool useBtof = 1
 bool nFit/nPoss weighting = 1
 bool DropPostCrossingTrack = 1
 Store # of UnqualifiedVertex = 5
 Store=1 oneTrack-vertex if track PT/GeV>10
 dump tracks for beamLine study = 0