You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 61 Next »

September 1, 2009

Phone bridge number {} 617 324 7374 {} , Time 2:00 pm to 3:30 pm, up to 12 connections
Agenda

  1. result of analysis of SL09b production
    1. Jan (Drupal blog )
      1. presentation of algo ver 4.0 - not linked
      2. Run 9 data
        1. histogram-based dump of selected events
        2. Larger collections of events: di-jets , W-candidates , Z-candidates
        3. Matt's 3 interactive events in skechup or just one view
        4. sequential histograms from pass through 1M events (LT=~8/pb)
      3. M-C simulation analysis histograms for Pythia W-events (LT=320/pb) , filtered QCD-events LT=6.3/pb, Z-events LT=4200/pb
    2. Joe
  2. discussion of Run 9 data publication strategy
    1. Questions/comments from former prominent STAR collaborator:
      • Could the Z-candidates be di-jets? cosmic rays? Do you try to do any charge sign discrimination on the electrons (i.e., opposite sign on the 2 sides for "Z-candidates")? Phi distribution does not appear indicative of cosmics.
      • Most importantly, is "Jacobian peak" absent when you look at 200 GeV data with the same algo and cuts?
      • Lots of questions, but this does look really promising, seems to give you NN W candidates NN Z candidates in the barrel region. It's a bit surprising that such a crude isolation cut, with no away-side ET upper threshold or SMD criterion, would suppress jet background so effectively. We always thought we would have to work much harder than this to unveil a signal above background.
  3. Ross - vernier scan analysis
  4. Joe - BTOW relative gains
  5. AOB

August 27, 2009

Minutes (not finished editing)

  1. Evaluation of SL09b production
  2. W analysis: who is working on it? What is your strategy? Do you need help?
    • Ross
    • Jan : wants to use default jet finder (for isolation and away side veto cuts) and has 0 experience.
    • anyone else ?
    1. jet finder tasks (Jan needs help), .C macro
      • decide on jet finding params, detectors: TPC, BTOW, ETOW
      • decide on data massage:
        • TPC track PT saturated at 10 GeV/c
        • identify/reject hot unmasked towers
        • temporary BTOW abs calibration
        • ability to select prim tracks not from other than first primary vertex
        • compute not only jet ET but also uncertainty of jet ET (to identify & ignore huge but uncertain jet energies)
      • run jet finder over 600 runs, QA performance, uniformity per run, per fill
  3. TPC calibration, Gene: Could you describe realistic timeline for complete TPC calibration: PadrowT0, Twist, ShortedRing
    Next production should include all of the above
  4. ETOW gains - Scott/Justin
  5. BSMD relative gains - Willie
  6. Vernier scan - Ross, all 6 pp500 runs have been produced.
  7. AOB
  8. Announcements
    • Jan: the following runs will be tagged as Xx because there is 0 or <100 events with found vertex :
      R10084013, R10088103, R10081024, R10095032, R10085022, R10088060, R10085103, R10087007, R10093049, R10085118
    • corrupted muDst: /star/data33/reco/production2009_500Gev_c/ReversedFullField/P09ib/2009/097/10097044/st_W_10097044_raw_6180001.MuDst.root

August 20, 2009

Present: Rosi,Hal,Jan,Justin,Gene,Ross

  1. Production status
    • pilot production has been QAed . No major problems were found and Lidia is most likely executing production now.
    • expect St-W muDst to be ready in few days.
    • identified problems to be fixed before next production
      • explain why PPV sees no tracks crossing TPC CM
      • why 7-bit bXing has no abort gap for the fill on page 19 top plot ?
  2. ETOW absolute calibration to compensate for incorrect TCD phase used during data taking. Two alternative methods are used:
    • Justin is comparing ratio of slopes for correct & incorrect TCD phase from regular minB events from pp200 data:http://drupal.star.bnl.gov/STAR/blog/stevens4/2009/aug/05/run-9-eemc-tcd-phase-and-effective-gains
      Fig 1.1 show example for 1 tower
      Link http://drupal.star.bnl.gov/STAR/system/files/ETOW_SlopeSet0vSet1234.pdf
      shows slopes (top) and the deduced gain correction (bottom) for all towers and all crates.
    • Scott is analyzing slopes from calibration pp500 data taken and various TCD phases and predicts similar gain correction for the TCD phase value used during W-data taking. The comparison is here

      Error rendering macro 'viewpdf'

      com.atlassian.confluence.macro.MacroExecutionException: com.atlassian.confluence.macro.MacroExecutionException: The viewfile macro is unable to locate the attachment "scott-08-20-2009-all_ETOW-crates.pdf" on this page

      In particular pages 4-6 show error bars. Both methods lead to the same gain correction.
    • The plan is to compute weighted average gain correction and upload it to DB with a different flavor. So user will get non-sense gains for the default 'ofl' and needs to request those. (Very good plan, Jan)
  3. Status of Vernier scan was presented by Ross:http://drupal.star.bnl.gov/STAR/blog/rcorliss/2009/aug/20/absolute-cross-section-notes-5
    The rate of STAR events vs. time (fig 1) is described by the product of 2D gauss functions representing X,Y spread of ions in both (identical?) beams. We know from Angelika the relative displacement of both beams for every time period of Vernier scan. Ross fits 2 params: sigX, sigY - the width of gauss.
    Fig 3 ('counts vs. steps') shows anticipated result. The value of 'skfNN' is absolute cross section of STAR BHT3 trigger in units TBD and assuming no hot tower correction is needed.
    Ross add constant background to better describe the observed rate , yet still peaks are under-predicted.
    He will continue to try to model vernier scan better and is awaiting for muDst from 6 Vernier scan runs from ongoing production to be able to assess how stable is the value of skfNN.
  4. run QA : Hal will do QA of the first week of pp500 data within next 30 days, backward in time. (Thanks^2, Hal)

July 16

  1. Joe/Gene - spacecharge and gridleak - Having had Gene diagnose that there were issues with the gridleak in the current calibration I went back with his estimated numbers and re-ran it, but there may be problems with that now and it may take another iteration or two to find our way to the correct place to be.
  2. Hal - QA - Didn't call in, but he said at the analysis meeting that he would do more QA (the rest of it).
  3. Rosi/Jan - PPV - Rosi was working on the beam tilt part of it now and it sounds like she is getting good results and everything so it won't hold us up once we get the sc/gl ready.
  4. Willie - BSMD - Showed his progress (see the email he sent around)
  5. Ross - Vernier - See his drupal page
  6. Scott - ETOW gains - See his presentation. He started looking at the gains in each sector now with the data from the timing scans, but using "Scott's method" for calculating the gain (last slide). It looks like sectors 5 and 6 are actually different from 3 and 4 when you use "Scott's method". He thinks he will have a gain for each tower calculated by the end of next week though. So excellent progress there.

June, 25

Present: Gene, Scott, Joe, Jan, Willie, Ross,Rosi

  1. TPC, Gene:
    • TPC alignment used now is old, from run 8
    • Joe produced non-linear SC based on 10 runs , not yet in DB.
    • Units for SC are V/cm^2. Conversion to actual correction used to shift clusters is SC of 0.01 V/cm^2 --> 8.5mm of shift in the X-Y plane for track at eta=0.
    • Few days later Joe generated SC ver 2 based on ~60 runs and differences between SC ver 1 & ver 2 are of 2 mm at the highest ZDC rate of ~170 kHz
    • we need to check if after applying SC the residua are scattered around zero, using different runs
    • Gene will evaluate DCA to the vertex to see if we achieved 2mm accuracy of global tracking
    • Jan investigated chi2/dof for tracking (before SC were applied) and found its average is of 1.5. We discussed how to rescale TPC hit errors to get avr chi2/dof=1.<br> There are 3 critical parameters deciding how big are errors of clusters: *dip angle, * crossing angle, and *track momentum(?not sure?) <br> Few days later Jan check chi2/dof w/ Sc and it was much closer to 1 - perhaps rescaling of hit errors is not necessary.
  2. BSMD was never aligned with TPC for real data - brought by Scott. We could (should) use high pT electrons reco in TPC for alignment. L2W events will be ideal. Jan asked Matt about the change in BTOW geometry last year to add dead material at eta=0. Matt said this change did not affected BSMD geometry model, so we have in reco what we believed was true ca. 1998. It needs to be verified with actual 2009 data.
  3. ETOW gains: Scott got from Alice timing scan data and will analyze them. We discussed reliability of gains for towers with significantly wrong timing when gain correction factor is of 10 or more. It is important to look also at derivative of d_gain/d_timing and flag as bad towers with large correction or large derivative. Details TBD.
  4. beam line constrain: Rosi was finalizing 3D fitter based on Minuit. 3 days later: code is running but convergence is not great, few problems are identified, e.g. too many fit params. Rosi & Jan are working on this - top priority, needed for production to start.
  5. preparation for data production:
    • we want to run over 1.4M of L2W events vs. ~900 M of all events taken in run 9
    • Joe will give presentation motivating this production to PWGC
      #*Before we start production following must happen: <br>- confirmation non-linear SC does improve TPC accuracy<br> - mean chi2/dof should be ~1 <br> - beam line algo by Rosi, Jan must be developed, tested, vetted <br>- beam line algo must be executed over full pp 500 data set and beam line params loaded to Db<br>- problem w/ ADC=0 for ETOW for certain runs needs to be resolved , Ross reminded me it was discovered already in the past, if # of corrupted 'sth' was >40 the whole block was set to 0 - need half day of attention <br>- run list needs to be generated, including sub-list of daq files for the 5% sample-production - this is half day of my work
  6. run QA: ver 3 of run list incorporating peoples run QA prepared so far has been assembled: pp500 run list ver3c , 80% QAed, updated June 29

June 18, 2009

Present: Gene, Scott, Jan, Willie, Hal, Ross

  1. TPC, Gene:
    • non-linear SC(lumi) should provide TPC track accuracy of 2 mm at both ends of the track. We hope Joe will work on this next week, should take a week.
    • We could also use Gene's PCA method of finding a linear combination of scalers (each at first order).
    • since linear SC(lumi) was tuned to lumi equivalent to ZDC base rate of 70kHz, any pp500 run taken at the ZDC rate has current best TPC calibration. Jan: I'll use it for beam line constrain study, see below.

    • beam line constrain : plan A is Rosi & Jan succeed with 3D beam line fit to global tracks w/o intermediate vertex finding. Here are [[Media:Beamline-v3.pdf ]] as intended and not implemented. If this works Gene needs a ~day to produce beam line constrain for all 39 fills. Plane B is to use Minuit vertex finder which may require significant stats to overcome jitter due to pileup tracks. In such case we would need to provide manpower for ~week long job, Gene would only supervise it. Jan: lets go for plan A.

    • conclusion: once both items are resolved we can ask for production, perhaps 2 weeks from now.
  2. BTOW, ETOW gains - no progress.

  3. BSMD : Willie is working '''slopes for individual strips'''. He run over 10k VPD-minB events and got this [[Media:BSMD-strips-Picture 16.png]] and spectra [[Media:BSMD-strips-Picture 15.png ]]. The available 1.5M vpd-minB events should be enough to fit every strip. If not we will merge strips.

  4. BSMD geometry: Scott asked how well we know '''relative position of BSMD vs. TPC'''. It is followed up on emc2-hn . <p> Scott once did [[Media:Conversions-distace-Scott.pdf | simple calculation for photons that convert to e+/e- pairs]] in the SVT, attached. <br>The relevant equation is on the bottom left. For the W > e, we have r_1 = 0, r_2 = 2.3 m (approximate radius of BSMD planes), and B = 0.5 T. With these numbers, you get the simple relation<br> dPhi (cm) = 40 / pT (GeV)<br>For example, an electron of 40 GeV pT will deflect in the phi direction by 1 cm relative to the same e of infinite pT. This is probably good to better than a few %.

  5. run QA : Scott is working on his fills.

no meetings for few weeks

May 21, 2009

Agenda:

  • Joe/Gene - TPC calibration progress, touch upon expected beam line constraint accuracy
  • Rosi/Jan - upgrade of PPV to reconstruct vertex in the transverse plane (for better beam line constraint) diagram, blog from Rosi
  • Scott/Alice (question) - ETOW calibration
  • Willie - correcting BSMD pedestal residual vs. fill , example mod 25, all 120 modules, webpage
    *Joe/Jan - production of L2W-stream could start ~June 15
  • AOB

May 14, 2009

no meeting

May 7, 2009

Present: Gene, Scott, Joe, Jan, Willie, Hal, Dave

  1. TPC:
    • determination of twist & sector alignment in progress, another 10 days needed
    • times of Anode HV on/off will be uploaded to DB in a day
  2. ETOW calib:
    • few dead towers needs to be added to DB
    • slopes from timing scans will be used to determine gains for towers, once Alice will come to IUCF and work on this, in ~10 days.
  3. BTOW: pedestals and status tables uploaded to DB. BTOW status tables done & QAed
  4. Hal run QA comments , spreadsheet
  • No labels