RSSC calibration

Each pair (or wire) consists of two straw tubes in RSSC.  They are connected on the upstream end. In E787/E949, the straw tube on the left side from the center of each sector (or chamber) is connected to the early channel and the one on the right side is to the late channel. t_late is the timing from the early channel, while t_early is the one from the late channel. Since Fe55 sources are sticked on both sides of chambers,  four peaks should be observed in the time difference distribution (t_late- t_early) as shown in  this histogram  for sector 8 and the 1st pair (or wire) in the inner layer. The first two peaks correspond to the  straw tube to the late channel and the last two peaks are for the partner straw tube to the early channel. The first step of RSSC calibration is to determine the four peaks position for each pair (or wire). Pulser trigger data (start from the leading edge and stop by the trailing edge with Fe55 source) with only RSSC on are used for this purpose. Since there are 2496 straw tubes (or 1248 pairs) for entire RSSC, 5k triggers from each straw tube will lead to 10 million pulser trigger events needed in total.

The code  (ana.F)  is used to fill the histograms of time differences for all the pairs. Using make_calib  to create executable file "rsscpeak.exe", one can then run the script "run_peaks"  to produce a hbook file "pulser.paw".  For each peak position, a Gaussian plus a constant term is used in the fit. A paw .kumac file  "peakfit.kumac"  is then used to fit all the time difference spectra and determine the peak positions, which are then written into a calibration file "peaks.dat".  This PAW .kumac file is also capable of handling with two separate hbook files if one would like to get more statistics in the peak positions determination. When running this PAW .kumac, error messages such as

         "No enough statistics, skip the histgram with ID=2223."
or

        "Use the previous peak positions for the histgram with ID=2256."

may show up occationally.  Be sure it is not due to a flaw in the pre-determination of peak positions.  After the fit to the all chambers are done, the code will compare the new peak positions and corresponding sigmas to the previous ones ( see peaks_old.dat . Warning messages such as

       "The peak shift by more than 20 counts in"
       "sec=         1lay=         0cham=         0"
       "Please make sure the fit is correct in"
       "histgram ID=       100"
or
      "The sigma changed by more than 1 counts in"
      "sec=         1lay=         0cham=        13"
      "Please make sure the fit is correct in"
      " histgram ID=       113"
or
      "The sigma is less than 2 counts in"
      "sec=         3lay=         0cham=        10"
      "Please make sure the fit is correct."
      "histgram ID=       310"

may indicate a problem. An "peaks_err.dat" is created, listing all the possible problematic fit. Please check the histgram and make sure the fit is correct.  If the fit is bad, please do it manually. The whole PAW job will take 15 minutes at maximum.

2. Calibration of velocity and tube length

With the peaks positions given above, one can determine the velocity and tube length by means of

Z = 0.5 × ( L  +   V × dt )

3. Check for the calibration

Since the Fe55 source positions are exactly the same for each chamber, the validation of above calibrations can be checked. The code  (z_check.F)  is first compiled with  "make_z_check"  and run through  "run_z_check" . The hbook file "z_check.paw" will provide the z information for each tube. If there is any z offset for one straw tube spotted in the same chamber (48 tubes in inner RSSC layer and 56 tubes in outer RSSC layer), it will indicate the input Z position used above is wrong and should be corrected in the code  "calibrate.F" .  The calibration is done iteratively until everything looks ok. Since there are always some dead channels, it is better to use the calibration constants from the neighbor tubes in stead of leaving the channel dead all the time. Usually, when a dead channel is spotted, it will be fixed in the later run.

4. Initial timing calibration

Initial timing is important now for the chamber installed with new electronics which are designed to get rid of the time walk effect found in E787. The pair with earliest hit is treated as the true hit, isolating the ones produced by cross talk. By doing so,  the Z position resolution  is seen to be improved  from a study of Kµ2 sample. An initial timing calibration needs  2.5k x 1248 /4  ~ 0.7 M Kµ2  sample.

The calibration for the initial timing for each pair is rather strait forwards:  compile the code  cal.F  with  "make_t0"  and create an executable file "cal.exe".  The script of  "run_cal"  will created a hbook file called "cal_t0.paw" .  Be sure that the  Kµ2  raw data is available and its name is correctly written in the script.  The readin initial t0 are all set to be  zero by adding "use sc tof  sc_tof.00000 " in the script of  "run_cal" .  When the hbook file "cal_t0.paw" is ready, a paw .kumac script "t0_cal.kumac"  can be used to perform fits to the TDC histrograms. The results are written into a file called sc_tof.dat, which should be added in CFM.
 

5.  Check the mapping of RSSC

Occasionally, there is a cable swap problem. To check it, one has to wait a standard pass2 job running. In the pass2 ntuple, the residuals of RSSC xy position measurement are available for both layers labeled as "rsxy1d" and "rsxy2d". A look at these residuals vs sector can immediately tell if there is the cable swap as shown for the xy residuals vs sector   for the 2001 Kµ2 sample, in which there were two cable swapping in outer layer sector 8. As this result, the  the xy residuals for inner sector 8 and 9  are also affected.  These can swap is found due to  the problem which the cables to early channels and the late channels  are swapped. A geometrical file for identifying the early and late channels is called "sc_geo.xxxx" in which the position for early channels and late channels are distinguished by the sign for the coordinators. The above problem can be solved by exchanging the coordinator values as given  here . There is also another RSSC mapping file called sc_map.xxxx.  If there is any cable swapping in the same sector or between sectors, this file should be updated.

6. Add all the calibration files in to CFM data base

The RSSC calibration files include sc_dtt.xxxxx, sc_tof.xxxxx, sc_map.xxxx and sc_geo.xxxx. Usually, the first two are very likely  the ones that should be updated.  http://sitka.triumf.ca/e949/ gives a detail of how to update CFM data base.