TPC Calibration & Data-Readiness Tasks:
Notes:
* "Run", with a capital 'R', refers to a year's Run period, e.g. Run 10)
* Not all people who have worked on various tasks are listed as they were recalled only from (faulty) memory and only primary persons are shown. Corrections and additions are welcome.
To better understand what is the effect of distortions on momentum measurements in the TPC, the attached sagitta.pdf file shows the relationship between track sagitta and its transverse momentum.
Δ(r-φ) vs. r and φ | Δ(r) vs. r and φ | |
---|---|---|
Padrow 13 | ||
Padrow 13 lines indicating locations of rows 13 and 14 |
||
Padrow 40 | ||
Padrow 40 lines indicating locations of rows 40 and 41 |
This is a recipe of RunXI dEdx calibration by Yi Guo.
In 2012, the procedure documentation was updated, including global T0 calibration:
Below are the older instructions.________________
The procedure here is basically to calculate two beamlines using only west tpc data and only east tpc data independently and then adjust the XTWIST and YTWIST parameters so that the east and west beamlines meet at z=0. The calibration needs to be done every run for each B field configuration. The obtained parameters are stored in the tpcGlobalPosition table with four different flavors: FullmagFNegative, FullMagFPositive, HalfMagFPositive and HalfMagFNegative.
To calculate the beamline intercept the refitting code (originally written by Jamie Dunlop) is used. An older evr-based version used by Javier Castillo for the 2005 heavy ion run can be found at ~startpc/tpcwrkExB_2005, and a version used for the 2006 pp run that uses the minuit vertex finder can be found at ~hjort/tpcwrkExB_2006. Note that for the evr-based version the value of the B field is hard coded at line 578 of pams/global/evr/evr_am.F. All macros referred to below can be found under both of the tpcwrkExB_200X directories referred to above, and some of them under ~hjort have been extensively rewritten.
Step-by-step outline of the procedure:
1. If using evr set the correct B field and compile.
2. Use the "make_runs.pl" script to prepare your dataset. It will create links to fast offline event.root files in your runsXXX subdirectory (create it first, along with outdirXXX). The script will look for files that were previously processed in the outdirXXX file and skip over them.
3. Use the "submit.pl" script to submit your jobs. It has advanced options but the standard usage is "submit.pl rc runsXXX outdirXXX" where "rc" indicates to use the code for reconstructed real events. The jobs will create .refitter.root files in your ourdirXXX subdirectory.
4. Next you create a file that lists all of the .refitter.root files. A command something like this should do it: "ls outdirFF6094 | grep refitter | awk '{print "outdirFF6094/" $1}' > outdirFF6094/root.files"
5. Next you run the make_res.C macro (in StRoot/macros). Note that the input and output files are hard coded in this macro. This will create a histos.root file.
6. Finally you run plot_vtx.C (in StRoot/macros) which will create plots showing your beamline intercepts. Note that under ~hjort/tpcwrkExB_2006 there is also a macro called plot_diff.C which can be used to measure the offset between the east/west beams more directly (useful for pp where data isn't as good).
Once you have made a good measurement of the offsets an iterative procedure is used to find the XTWIST and YTWIST that will make the offset zero:
7. In StRoot/StDbUtilities/StMagUtilities.cxx change the XTWIST and YTWIST parameters to what was used to process the files you analyzed in steps 1-6, and then compile.
8. Run the macro fitDCA2new.C (in StRoot/macros). Jim Thomas produces this macro and you might want to consult with him to see if he has a newer, better version. An up-to-date version as of early 2006 is under ~hjort/tpcwrkExB_2006. When you run this macro it will first ask for a B field and the correction mode, which is 0x20 for this correction. Then it will ask for pt, rapidity, charge and Z0 position. Only Z0 position is really important for our purposes here and typical values to use would be "1.0 0.1 1 0.001". The code will then report the VertexX and VertexY coordinates, which we will call VertexX0 and VertexY0 in the following steps.
9. If we now take VertexX0 and VertexY0 and our measured beamline offsets we can calculate the values for VertexX and VertexY that we want to obtain when we run fitDCA2new.C - call them VertexX_target and VertexY_target:
VertexX_target = (West_interceptX - East_interceptX)/2 + VertexX0
VertexY_target = (East_interceptY - East_interceptY)/2 + VertexY0
The game now is to modify XTWIST and YTWIST in StMagUtilities, recompile, rerun fitDCA2new.C and obtain values for VertexX and VertexY that match VertexX_target and VertexY_target (within 10 microns for heavy ion runs in the past).
10. Once you have found XTWIST and YTWIST parameters you are happy with they can be entered into the db table tpcGlobalPosition as PhiXZ and PhiYZ.
However - IMPORTANT NOTE: XTWIST = 1000 * PhiXZ , but YTWIST = -1000 * PhiYZ.
NOTE THE MINUS SIGN!! What is stored in the database is PhiXZ and PhiYZ. But XTWIST and YTWIST are what are printed in the log files.
Enter the values into the db using AddGlobalPosition.C and a file like tpcGlobalPosition*.C. To check the correction you either need to use files processed in fast offline with your new XTWIST and YTWIST values or request (re)processing of files.