From: Pawel Staszel (staszel@alf.nbi.dk)
Date: Thu Oct 17 2002 - 12:06:42 EDT
Hi Kris,
I found a first difference between your and my calculation which in probably
on the
local tracking level. I run on 3 sequences of raw data for run 6249 (which I
found in data02/data/pp)
and on 3 seq. of reduced data from data02/data/pp/R6249/seq. I use the same
software and the only one
difference is that for rung on raw data I have to include modules for the
local tracking in T1 up to T5
(the geometry is the same in both cases (from db)).
The main difference is in TPCs. The two attached plots show dy distribution
for t1-t2 matched tracks.
FFSDy-1.ps - is for raw data (my local tracking) and FFSDy-3.ps - for reduced
data.
As you see the sigma is almost the same but in case of my tracking the offset
is -0.03cm and in case of
your tracking offset is 0.22. Also in the case of my local tracking the
distribution is more gaussian-like.
Probably you have a different setup for TPC tracking modules (and rather for
TCP calibration module).
My script is
/direct/brahms+u/ufstasze/brahms/brahms_app/ps_app/fs/receff/fsRecEff.C.
We also differ slightly in tracking in T5 (in T3 and T4 seem to be identical).
Check whether you have
additional setup for T5 like:
if(rundbOption->GetValue()>5580)t5->SetViewCombinerParams(1,1,4,0,0);
else t5->SetViewCombinerParams(1,1,2,0,0);
It is because after run # 5580 we run without one module of T5 - so we have to
be less strict.
But as I remember I put t5->SetViewCombinerParams(1,1,4,0,0) as a default.
Please, sent d me the name of your script, so we can find out the problem.
Regards Pawel.
Kris Hagel wrote:
> OK,
> So it looks like I need to reconstruct again, this time running
> BrDcRdoModule. I will start that process as soon as HPSS is back up.
>
> For the matching, I used defaults. (ie not the matching files) I looked
> at the matching parameters and accepted vs total and saw in the runs I
> looked at that I was getting almost everything. My conclusion was that
> pp being cleaner that there were less problems. I did, however, worry
> about that and had plans for the next round of matching to do that in a
> formal way. It was clear to me how to do that for the MRS and FFS, but
> not BFS. So I will make look into that in more detail and do it. I
> have a hard time imagining that would account for the decrease in
> efficiency you see, but since I didn't pay attention, it just might.
>
> For the matching, I used BrFsTrackingModule which has BrModuleMatchTrack.
>
> Kris
>
> Pawel Staszel wrote:
>
> >
> > Kris Hagel wrote:
> >
> >> Pawel,
> >> You are right, I do not have BrDcRdoModule. I did not forget; I knew of
> >> its existence, but it was not used in the example script I got (I don't
> >> know where) when I started pp analysis. I polked around somewhat and
> >> never found where it would be used, so I didn't include it.
> >>
> >> So, what is it used for? Will I have to reconstruct all the runs again?
> >> Am I missing important information by not having it in there?
> >>
> > BrRdoDc are used in the later stage of analysis, when doing the global
> > tracking, to
> > retrieve some local T3 and T5 tracks which were missed in "ordinary"
> > tracking during the
> > first level of reduction. The module that does it is called
> > BrDcEnhancementModule - it is not yet in brat but in
> > rk_app/bfs/dcenh. Concerning pp, this is specially important for T5.
> > We run with T5-II turn off,
> > and the "ordinary" (or rather "regular") tracking is around 80%
> > efficient. BrDcEnhancementModule
> > can improve final tracking efficiency in T3 and T5, by utilizing the
> > global tracks found in the other detectors.
> >
> > Any way, looking at the data you have reduced I seen 91% efficiency in
> > T3 and 97% in T4.
> > From my calculation on raw data (run6418) I saw 99% for T3 and 99% for
> > T4. So first
> > I have to find out where this discrepancy came from and have a question.
> > How you do the global tracking - do you use the
> > BrMatchingOffsetModule. If so, you need
> > files that contain sigmas and offsets required by MrModuleMatchTrack.
> > Do you have such files for pp?
> > (for AuAu we called them offset5300_6000.FFS, offset5300_6000.T2_T4,
> > offset5300_6000.T3_T4
> > and offset5300_6000.T4_T5).
> > If so, I would like to use the files. I not - how you do the matching?
> >
> > Pawel.
> >
> >>
> >>
> >> Reconstructing the raw data again will not be a big deal once HPSS comes
> >> back (probably next week) if that is necessary. This is also a good
> >> time because I have been looking at a relatively small number of pp runs
> >> and am now ready to build up to the full statistics.
> >>
> >> Regards
> >>
> >> Kris
> >>
> >> Pawel Staszel wrote:
> >>
> >> > Dear Kris,
> >> > I found out that there are no BrDcRdo tables for T3 and T5 in
> >> > pp reduced files.
> >> > You probably forgot to add BrDcRdoModule to your reduction script.
> >> > The example of how to include DcRdo is in file
> >> > /brahms/u/bramreco/reduce/ProductionReduction.C.
> >> > Check also whether your BrDcTrackingModules are set in the same way.
> >> >
> >> > Regards Pawel.
> >> >
> >> >
> >> >--
> >> > ----------------------------------------------------------------------
> >> >| Pawel
> >> Staszel |
> >> >| Niels Bohr Institute Tb 8 email:
> >> staszel@nbi.dk |
> >> >| Blegdamsvej 17 phone: (+45) 35 32 53
> >> 51 |
> >> >| København Ø FAX: (+45) 35 32 50
> >> 16 |
> >> >|
> >> Danmark |
> >> > ----------------------------------------------------------------------
> >> >
> >> >
> >>
> >--
> > ----------------------------------------------------------------------
> >| Pawel Staszel |
> >| Niels Bohr Institute Tb 8 email: staszel@nbi.dk |
> >| Blegdamsvej 17 phone: (+45) 35 32 53 51 |
> >| København Ø FAX: (+45) 35 32 50 16 |
> >| Danmark |
> > ----------------------------------------------------------------------
> >
> >
> >
> >
> >
--
----------------------------------------------------------------------
| Pawel Staszel |
| Niels Bohr Institute Tb 8 email: staszel@nbi.dk |
| Blegdamsvej 17 phone: (+45) 35 32 53 51 |
| København Ø FAX: (+45) 35 32 50 16 |
| Danmark |
----------------------------------------------------------------------
This archive was generated by hypermail 2.1.5 : Thu Oct 17 2002 - 12:08:34 EDT