Detector Characterization Uncertainty -Poisson

[kitepassion]

Hi to all,

I have the following problem: I want to characterize two lots of detectors in order to retrieve their mean cross section and standard deviation for a proton beam. For cross section I mean the number of counts respect the fluence of particle (fluence = integral of the particle flux over the time).
Now for each lot of sensor I have 12 detector. For each detector I made some measurements at different energy. For istance 3 measurements at 30 MeV, 10 measurements at 60 MeV and so on..
For each measurements more then 6000 counts were collected.
Now I want to calculate some meaningful parameter to characterize the detectors.

What I did was to collect all the measurements carried out at the same energy but on different detectors belonging to the same lot and calculating their mean and standard deviation as a gaussian distribution. Now the 1-sigma standard deviation could be called uncertainty?
Is the process correct? Am I missing something?
Should I use a Poisson distribution?

 

[mfb]

What I did was to collect all the measurements carried out at the same energy but on different detectors belonging to the same lot and calculating their mean and standard deviation as a gaussian distribution. Now the 1-sigma standard deviation could be called uncertainty?

Looks reasonable. The uncertainty comes from statistical fluctuations and differences between your detectors. It would be interesting to see if those differences (within your groups of 12 detectors) are significant.

Should I use a Poisson distribution?

For 6000 events, a Gaussian distribution is fine.