Home » Groups » STAR Protected

EEMC Calibration Notes -- Summer 2010

Updated on Tue, 2011-01-25 18:12. Originally created by rwm46383 on 2010-06-01 13:29.

5/24/201

Alice Posted page here of helpful comments to run Jan's code:

http://drupal.star.bnl.gov/STAR/blog/aliceb/2010/may/27/eemc-calibration-tips

Scott Wisink:  Link for 2007:  http://www.star.bnl.gov/~wissink/cal2006/run7Slopes/

Other Links referred to by Alice:

http://www.star.bnl.gov/public/eemc/calibration/run4/calibUnits/

http://web.mit.edu/rhic-spin/public/eemc-bnl/software/gainCal/mips_1.html

http://www.star.bnl.gov/public/eemc/calibration/run5/absMipCal/

http://web.mit.edu/rhic-spin/public/eemc-bnl/calibration/

 

 6/1/2010

Questions for Scott on our WED 2JUN Call (1:00 pm EDT)

  1.  Is this to be an absolute calibration using expected dE/dx of MIPS?

    2. No, we start with relative gains, matching slopes in SMD, nothing in TOWERS

    Then after one or two iterations, we worry about absolute gains--towards the end.

  2.  MIPS--  What?  Are these mainly hadrons or electrons?  MIPS are likely 0.5 to 1.5 GeV charged pions.  dE/dx expected to be for this p-range.
  3.  Is the Sampling Fraction still an issue to get absolute calibration correct? Only for the Towers, not the SMD.  
  4.  Will Justin Stevens be involved?  Has he worked with Jan's code?
  5.  Alice was suggesting using ZDC's.  Do we want WEST ONLY Zdc Trigger ?  (Endcap is in the West -- Correct?)  Endcap is on WEST side;  we want to use both E and W ZDC triggers.  There is little correlations of what goes towards the Endcap and the neutron headed towards the ZDC's.
  6.  Eta- dependence of Energy deposited -- is this tuned so that there is a known eta dependence?    The towers only have eta-dependent gains. This is so that E_transverse is constant in Eta, not E.  I also note in the routine StEEsmdCalMaker.cxx, there is a comment as to this effect:   "ideal calibration usedby Fast simulator has eta dependent gains for towers and plain energy deposit for SMD, pre, post."  (Note fro Jan I assume) b
  7.  Related to same sensitivity to p_perp for all eta?  -- I am confused about something else here I think, but there
  8.       is an eta dependence to gains.  WHY?   "Ideal gains" -- How defined?  Are they eta dependent? ==> See above:  only Tower gains are eta dependent.  Not pre/post/SMD. 
  9. Describe iteration.  How many cycle?  Do we start with the relative gains used at the beginning of the 2009 run?
  10. Where do we start?  It seems difficult to get a grasp as to how Jan's code works. Is this sequence on track?  (From STAR EEMC Calibrations Workshop summary):
    • i) use relative gains in place,
    • ii)  Cut:  Select MIPS with cut demanding hit above pedestal in PRE1+PRE2+TOWER+POST

      •     + SMD with 1/2 slats above threshold and 7 (or 11) slats not hit on each side

    • iii) fit Gaussian to the ADC distribution of interest

      •      iiia) For SMD strips, fit Landau Dist and use MPV from Landau fit. 

           But to what does one "adjust MPV" to?  Is there an ideal value of it for MIPS? 

      iv) modify gains such that desired energy is deposited (from known MIP dE/dx)

      v) repeat   

                    ==> WHAT AM I MISSING?

  11.   What sequence of code/functions accomplishes this?

 Summary from Phone Conference 6/2/10 with Scott and Alice:

     We start with NO gains/no gain tables from previous years, and first try to get the SMD strips performing so the compare relatively using uncut MIPS.      The ZDC triggers should be fine.

One fits slopes of the tails beyond the 1-electron photopeak, and trys to make all smd's look about the same. 

The Code that does this is ______________. 

[Not sure what functions are used, but rdEztMuSmdCal.C reads in the MuDst files and begins the process.]  I need to get insight from Alice on this.

Once the SMD's are all relative the same from fit to Landau tails, we set up a MIP's Cut demanding Pre1&2, SMD u+v,Post, Tower.  Then leave out the element who's gain is to be determined.  One will need 2 -3 iterations for SMD, maybe 2 for other elements.

RUNS:   Alice had problems with the list file for muDst.  But thinks she figured out the problem.  Scott gave her an EXCEL spreadsheet.  CAN I GET?

At the beginning of each FILL, there is a ZDC Polarimeter run of some 100k events.  That would be a good set of MinBias triggers. These are spread out in time -- not so good. But the benefit is that the setup is identical for these.

One needs the STATUS TABLES on the run.  [How do I point to the status tables that have been done???] 

When working with the SMD's after the relative gains have been set by using the slopes of the ADC spectra, due to the triangular shape of the strips, the MIPS will not depsit same energy.  One gets around this by adding adjacent strips.

EDGE CUTS:  Throughout this, even for the SMD, one does not want the u+v intersection to be too close to edge (say 1 cm inside).  So one needs to impose edge cuts. [WHERE  is this in code?] 

GAIN:  Two senses of "gain" are used.  TOWER:  set "gains" such that  60GeV of E_transverse will saturate the ADC, hence fall in channel 4096.  Also E_tower is scaled by the sampling fraction  [ Is it E_transverse = gain*ADC*Sampling fraction (.05)?]

                                                            SMD,Pre,Post:  no sampling fraction enters.

Early On:  We should compare SMD ADC slopes to the 2006 Calibration results for several strips, just to see if things are reasonable.  Scott could not remember where these values were, but I think they are available here: 

                  http://www.star.bnl.gov/public/eemc/calibration/run2006/cal2006/

 

NOTE:   Here is an example of fit to Landau Distribution:

 

   

Jun 3,2010 -- Setting up Alice's code

Copied Alice's code here:    /star/u/rwm46383/work/EEMCcalib/2010-may-18

Setup link on this disk for output:   BNL>ln -s /star/institutions/valpo/rwm/EEMC_calib/iter-0/ .

Created two subdirectories:  ./histo and ./log

 Jun 16:  Reading through Zach's code.  

Zach has nice results of the fit to slopes. 

To get gains, he has used the strange hardwired constant

relative gain g = scale factor/slope where scale factor = (850/1.4) ~ 607.14. 

We need to understand this!

 

23JUN,2010

Scott Wissink's histograms from 2006 in raw form are here at RSSH:     /star/u/wissink/cal2006/iter4-pp/

There are several iterations in the cal2006 directory.  I assume the later iterations are likely more final results.

25JUN,2010    Note from alice: 

Database at STAR:     http://www.star.bnl.gov/Browser/EEMC/put EEMC pix gains in Data Type and 2006 to 2007 in the years you can find the preshower/postshower gains from the last real calibration. Look at flavor ofl for an element.  The things with P in the name are preshower1, Q are preshower2 and R the postshower.  FYI: to get tower gains browse EEMC Pmt gains instead of pix gains.  The naming scheme in these files is the same as in the code (or really really should be).

Alice

================

25JAN2011:

Bob sent note to John S:

John,

I looked back in my notes and dug out where Scott's 2006 code and histograms are:

/star/u/wissink/cal2006

I had a list set up defining what the codes were and did, but have not yet found it.
I think Alice and Zach did something similarly.

However I think I moved a copy of all the code (not hists, etc) here:

        /star/u/rwm46383/work/EEMCcalib/2010-may-18/StRoot/StEEmcPool/StEzSmdCal

Here is my log of last summer on the STAR Drupal server:
http://drupal.star.bnl.gov/STAR/blog/rwm46383/2010/jun/01/eemc-calibration-notes-summer-2010
Feel free to poke around there.
Alice's code is also on the /star/u disk.
Let me know if you want me to work through some of this with you.  I can do it after my classes tomorrow.

Bob

 

 

 

Groups: