W45 Analysis and Software Page

This page is under construction.

Contact Mario Leberig for more information.

How to calibrate W45

1. Produce MegaDST

   To calibrate a detector it is necessary to get access to the raw information collected by this detector. As this information is not stored in the standard DSTs the raw data has to be reprocessed by CORAL and Phast. The output files obtained should contain the detector digits as well as all tracking information. These file will be called in the following description megaDST or MDST.
   • The installation procedure for phast as well as the principal special settings needed to produce megaDSTs are described here.
   • The set of shell scripts which can be found here allows to produce megaDSTs for a full run. Copy the tar-archive to your phast.X.YYY/coral/test/ directory unpack and read the README-file.
   • After having done so you must produce the template traffdic.opt file which should be used in the production. Here you have to take into account the following steps:
     • Add a line like:
       mDST hits DW
       mDST digits DW
       to store hits or digits in the megaDST
     • Remove your detector from the tracking by adding the line like:
       TraF DetNameOff VO VI DW
     • You might to add smoothed helixes to the tracks at the position of your detector by adding lines similar to:
       TraF SmoothPos [0] 2860. // Smooth track at
     • Make sure that no "Data file" line is given. Otherwise in every run the same data file will be added.
   • Launch the script msubmit.sh and cross your fingers. If everything works right you will submit a bunch of jobs to LSF and after some time you will have your megaDSTs.

2. Produce ROOT Tree

   After creation of the megaDST file phast can be used to further reduce the data and to create an easy to use ntuple which contains the necessary variables to do rt-calibration. My phast function can be found together with the class (W45P.cc and W45P.h) I have written here. My ntuple contains the folllowing variables:
   • extra: flag to mark tracks which were extrapolated to the detector
   • alpha: track angle ( alpha = 0 ==> perpendicular to wire plane)
   • x : x-ccordinate of track at detector position
   • y : y-ccordinate of track at detector position
   • z : z-ccordinate of track at detector position
   • u : u-ccordinate of track at detector position
   • v : v-ccordinate of track at detector position
   • d : track distance from wire (measured in the detector plane)
   • dd : distance of closest approach between wire and track
   • ex : track precision in x-coordinate
   • ey : track precision in y-coordinate
   • eu : track precision in u-coordinate
   • ev : track precision in v-coordinate
   • t : hit time
   • tt : track time
   • tsig : signal propagation time
   • chn : channel of the hit
   • dID : dID ( 0 - 15 ordered by z-coordinate)
   I have writte a shell script which allows a fast processing of a full set of megaDSTs stored in one directory, which can be easily adopted for your needs. This script produces one root file for each megaDSTs file and produces the chain.cc root script which can be used to merge all this root files in a single one (just type root -b -q chain.cc to execute the script).

3. Analyse ROOT Tree

   The final analysis is then done by a root script which does the calibration in three steps starting with the so called V-plot (a TH2D).
   1. Project the histogramm on the time axis and fit to first order polynoms to the edge of the spectrum to determine T0.
   2. Project Slices of the histogram on the r-axis and fit two gaus to these 1-d histos. These two gauses should be centered around zero, if the detctor was correctly aligned. I fill a histogram with (mean1+mean2)/2. The mean of this histogram is the misalignement R0.
   3. After applying both the T0 and the R0 correction finaly the rt-plot is fitted by projecting slices to the time axis and fitting a gaus to each of these slices.

4. Calculate Residuals

   After having done the calibration it is necessary to check the success by determining the residuals of the different detector planes. This can be done by another root script which reads the calibration and alignement values determined by rt.cc and loops over the ntuple to determine residuals.

5. Possible Improvements