Čerenkov Photo-electron amplitude
determination
| This document briefly describes the way to measured the Photo-electron amplitude Čerenkov PMTs. |
J.S. Real, LPSC
Grenoble
Updated July 17th 20067
Cerenkov PMTs detect signal at the
photoelectron(PE) Level. The cerenkov trigger is 2PMTs/4PMTs, the
threshold is setting to 1/2 of PE. The gain (the mean amplitude of the
photoelectron) should then matched for all the PMTs and has to be
checked regulary.
There are 3 ways to measure the
Photoelectron amplitude:
1) with a scope, the measurement is subjective and wrong by few %
2) the ADC, but it does not seems very precise for photoelectron level.
3) the rates of the photomultiplier noise versus the threshold (the
Threshold scan).
The PE measurement should be done without
beam, after several minutes (hours) after setting the HV On. The
cerenkov's PMTs are 5 inches tubs which need some times to stabilized.
After changing a PMTs, calibration should be done several
times a day for the first 24/48 hours .
The threshold
scan
When there is no beam, the noise is mostly
1 PE. The amplitude of the PE response of a photomultiplier is a
gaussian distribution with a sigma of 1/3.
(see here for more detail). The threshold scan (looking at the
rates versus the threshold) is then the integral of that gaussian. The
mean value of
the photoelectron amplitude is the amplitude where the rates is half of
the one a 0mv.
To
do a threshold scan:
- Ask for beam off.
- log in as gzero on any cdaqlx machine [x stand for 1 to 7]
- execute ~/g0backward/scripts/cer_threshold_scan.pl ( this
change the configuration file to be load at prestart of the run). The
setup is maximum Fastbus prescale, no coil modulation and a flag in the
cerenkov discriminator file.
- Start the run.
- Once the run is started, just go in the g0daq_config and click submit
to restore normal running mode. (You don't have to wait for stopping
the run to do that).
- wait until 1 minutes (~2000 events) with beam or 4-5 minutes with no beam (~ 10000 events).
- Stop the run
To Analyse the run :
- On any cdaql computer (login gzero), wait for few minutes to be able
to acces the data file, then run g0analysis with mpsntuple on
("g0analysis --mpsntuple all --nobeamcuts -r [run number]").
- run the scripts $G0SCRATCH/scripts/Cer_Threshold_Scan.C runnb
The script plot the rates versus the threshold for all 32 cerenkov PMTs
with one canvas per octant. For 1 PMTs the plot seems like :
 | -If we don't take into account
the first point at 10 mV (which is not relevent for the PE), we can
assume that the amplitude at 0 mV is the one at 20mV. Then searching
the amplitude for which the curve is half the value at 0 mV we find the
mean of the distribution of the PE amplitude. - In this example the value at 0mV is 18. The x value for which the curve is 18/2=9 is 80 mV. Then the PE for this PMTs is at 80 mV. |
Once the PE amplitude (Am) is measured, we can then adjust the HV of each PMTs for the wanted PE amplitude Aw (set to 100 mV for french octants) with the formula:
HV_new = ( HV_old - 157 ) . ( Aw / Am ) ^ 0.153162 + 157