The determination T0 or TDC pedestal
  • The time spectra and pulse heights can be checked by using the codes BASIC and SIMPLE_XY_CLUSTER. The executable file is created by means of `make -f make_basic', then running this executable file with ` run_basic_ped' should be able to produce a histogram containing all necessary information for determining the TDC pedestals.  Before heading down the bussiness of determining T0, one should check all the plots avaliable in the histogram created, spoting any possible problem of cable swapping, etc. For example, one can checked  the plot of ADC cluster vs wire for each foil layer and see if there is any offset as shown in  this plots  (top two are the problematic ones with cathode cables swapping in the same foil layer, while bottom two are the ones after this problem is fixed.  Also shown here  are the cathode cable swaping between two different foil layers.  The cable swapping problem should be fixed before the calibration work can go further. Detail correction can be found in below  with titles of "UTC map".
  • With all the available time spectra, one should be able to get the TDC pedestal for each channel by means of a normal frequency function

    •  

     

                                           Amp* FREQ((X-T0)/Sigma) + Noise

    taking into account the noise contribution. To do so, an executable file should be first created with `make -f make_t0fit', linking the codes T0FIT, FITARRAY, LENSIG and KERROR. Running this executable file  with ` run_t0fit' to fit the time spectra produced above, wire by wire and foil by foil. A calibration constant file t0fit***.txt can then be obtained. As shown in this typical t0fit plot , to avoid a possible uncertainty originating from the TDC pedestal determination, the lower noise level, the more reliable TDC pedestal determination. The created t0fit***.txt should be checked before it can be provided for any further purpose. In some case, a negative value may show up in the file t0fit***.txt, indicating a fail fit to the T0 due to no entry or fewer entries in the plot or a `hot' wire with a potential noise. One can do nothing but assign a neighboring wire t0 value to this wire. More strictly, you can use more Kmu2 events, if they are available, to solve the no-entry or few entry problem.

    The determination of the space time relation

  • In the first step one needs to process the raw data and remove all unnecessary information, in order to reduce the size of the Kmu2 data file, because the STR finding needs an iterative procedure--a time consuming job. A special code DROPB is thereby used. Creating an executable file with `make -f make_dropb', one can use ` run_dropb' to produce a compact data file called run***.skim1.
  • Then the codes FINDSTR, WRITE_STR, UTC_ABIN, UTC_BOOK and TFITXY are used to create an executable file with `make -f make_findstr'. The procedure of finding the STR will start by means of `run_findstr_ang', which will use the TDC pedestal calibration file given above and perform a seven-parameter polynomial fit to the space time relation for each layer with left and right separation and four angle divisions as shown in these plots, giving a calibration file named findstr*.txt. To improve this calibration, an iterative procedure will be strongly encouraged to use with `run_findstr_ang_long', in which the starting value can br the above findstr*.txt. The last iteration will give a findstr*_6.txt used as the standard calibration file. The space time relation is very important in UTC track reconstruction. Before put it into CFM data base, it is better to check the resolution variation with the cell fraction as shown in the plots with good fit and the plots with bad fit . Since the seven-parameter fit may not describe the STR very well, using the histogram and setting a look-up table for STR can also be done. However, we can not find any improvement on the muon momentum (the result from the look-up table is shown in dotted).

    •  

     
     
     

    The ADC gain calibration

  • Once completing the TDC pedestal and the space time relation calibration, the last calibration involving the UTC is the ADC gain calibration. In practice, one needs the latest t0fit***.txt and findstr*.txt calibration files. In the first step, the ADC information should be filled into the histograms by using `make -f make_utc_nt', linking the necessary codes UTC_NT, CALC_PATHLEN and CALC_PATHLEN_B0. Using ` run_utc_nt', the histograms are thereby created.
  • To get the gain factors, all ADC information with respect to each wire and foil in the histograms is fitted by using `make -f make_adcgain'. The called codes are ADCGAIN, FITGAUSS and TRUNCMEAN. Then run run_adcgain, performing a fit to the ADC information as shown in these plots. The calibration files anode wires and cathode foils are produced with the names adcgain***_an.txt and adcgain***_ca.txt. Since there are some dead wires or strips in the used calibration sample, it is likely that the ADCGAIN calibration can not be performed, giving values like 1.0000 -1.0000. In this case, the best solution, of cause, is to use another calibration sample if it is available. Otherwise, the values from the neighboring wire/strip will be assigned to this channel.

    •  

     
     
     

    Check for the calibration

  • Before all above calibration files can be in use, it is necessary to run the test codes including KINEMA, RGSCIN,RD_TRK_ANAL, MY_TRKPOS and HBKINI, using `make -f make_kinema' and ` run_kinema' to process the KMU21 data. If the muon momentum (PTOT) is centered around 235.5 MeV/c, and the momentum resolution is around 3 MeV/c, that will indicates a good UTC calibration.

    •  

     
     
     

    Check the UTC MAP for anodes

  • From time to time, UTC cables are swapped during the maintenance period. Once the above UTC calibration is done, a check should be conducted to see if the UTC map is correct. Here is an example: the scatter plots of xy residual vs wire # show a gap at the common position in each sub-layer. Since a check on the wire TDC and ADC occupancy (from the ntuple produced by running run_basic_pedabove) show no problem in these anode wires, these gaps should be regarded as a result of using a wrong UTC map. Looking into the file for UTC mapping, usually named as ut_map.****, one should be able to find out the TDC cable for these gap and swap the cable as given here. And then, use the correct UTC map, redo the calibration until the gap is gone as shown here. Since anode TDC and ADC readout the same preamp and postamp, when correcting the TDC map, it is also needed to correct the corresponding ADC map, which is usually named as ua_map.****.

    •  

     
     
     

    Check the UTC MAP for cathodes

  • The check of the cathode map is not so straightforward as we did for the anode map, since a cathode hit cannot give a z position without knowing the anode wire number in the adjacent layer. To spot the problem, one should check the z residual vs strip number for each foil as shown here for single strip hit in the 5th foil layer. One can see there is a gap covering strip number from 109 to 114. Also observed is the funny behavior at strip number 48 and 56. By examining the cathode strip occupancy plots for ADC and TDC (also from the ntuple created by running run_basic_ped above), one finds no problem for these strips except for strip 48 and 56 with no ADC count. Since the gap covers 6 channels being consistent with a single preamp configuration, the problem may arise from either the channels order changed or swapped with other cable. Looking into the file for UTC mapping, usually named as uc_map.**** for cathode ADC and ut_map.**** for cathode TDC, changing the maps, one can see if the gap is filled by using the new map. For the problem shown above, we found it was swapped with a preamp cable in the 6th foil layer. The corrected map as given here. After correcting the strip ADC and TDC maps, the plot of z residual vs strip number for single strip hit in the 5th foil layer shows a normal distribution. In some case, one can debug the code 'utc_z_adc' and check all the strip hits. From the anode cell number, one should be able to determine the expected z position for the studied strip. Using the informatin from the code 'utc_setup', one may tell the correct the strip map. It is noted that the z residual for the neighbor strip will become worse when there is a dead strip as shown the above plot.

    •  

     
     
     

    Install the calibration files into CFM data base

  • The final step is install the above calibration into the calibration file management (CFM) system. To do so, log in on bnlku9 as e787, setup kofia and copy the calibration files to $CAL_DB. The naming convention for the UTC stuff is
    • Content  File name
    TDC  pedestal  ut_ped.nnnnn
    Space time relation  ut_sta.nnnnn
    Anode ADC gain ua_gai.nnnnn
    Cathode ADC gain uc_gai.nnnnn

    Usually nnnnn is set to 1 greater than what is already in $CAL_DB. Start up CFM by typing `cfm' and answer two questions

    Do you want to READ or WRITE the CFM database (R/W) ? w

    Enter your user name: XXXXX
    Update the database by

    modify 35000:40000 ut ped 14

    modify 35000:40000 ut sta 11

    modify 35000:40000 ua gai 6

    modify 35000:40000 uc gai 10

    exit
    The above example shows how the new calibration files are installed for the run number ranging from 35000 to 40000.

  • When all are set, send a mail to e787.pass1@viper.princeton.edu and inform people that the UTC calibration has been updated.