I am currently working on the timing calibration of the CLAS Electromagnetic Calorimeter (EC) for the e1f run. This is done with the use of a program named ec_time, the use of which is described by the following README file:
Initially, I used the code to analyze run # 038493, which consists of 25 files. The code produces an hbook file which can be viewed using PAW to determine the quality of the calibration. I primarily want to look at histograms displaying the difference between the EC time and the SC time for electron events. A sample histogram is shown below. For a good calibration all such plots would be centered to zero to within 100 ps (0.1 on this scale), and they must have a sigma value of less than 300.
If after the code is run on a particular run number, the desired results are not obtained, the data should be recooked with the new calibration parameters. The ec_time program should then be run on the newly cooked data. This is an iterative procedure that should quickly converge on good values given a good set of starting parameters.
When the desired results are reached for the initial run number, the entire data set should be cooked using the new values and then examined in several places spaced evenly throughout to search for any changes that need to be made to the calibration. If any discrepancies are found that region of the run should be calibrated until the desired parameters are obtained. Once the calibration coefficients have been satisfactorily determined for the entire run, the data should be recooked for one last time using these parameters.
I have produced an initial set of calibration coefficients for the following run numbers:
37742
37743
37744
37745
37746
Click
here
to see the plots from the initial calibration. These plots show the difference between the EC time and the SC time, which should be zero. We know that we have a good calibration when all of the sigmas are below 350 ps and the means are within 100 ps of zero.
A second iteration was completed. The application of the new constants showed a great improvement over the initial set. A comparison of the two can be seen at Maurizio's monitoring page.
I also wrote a kumac to compare the histograms produced by the ec_time code. Click here to see these plots. There are 2D plots of mean and sigma vs strip and sector for each view/layer combination, as well as 1D plots comparing the histograms displayed in the above link.
Below are two plots showing values taken from the monitoring histograms produced in the latest pass0. They show the mean and sigma vs sector for ec_time - sc_time.
Current Progress: Please see: Update
