For detailed description in german please see my diploma thesis .
Here you can see the setup for accurate measurements of pulse height distributions.
A solid state Laser called NanoLED by
HORIBA Jobin Yvon
emits photons with 375 nm weavelength and has about 200ps FWHM pulse duration.
The number of photons is reduced near to single photon events. The branch with the splitter
and trigger PM is not needed.
This setup is not noise free. This means the higher the emission rate of the source is,
the better is the signal to noise ratio.
With this setup you can get the information about puls heigth spectra and about the signal
shapes. In this case one can use also other light sources like simple LEDs. The emission duration is
less importand. If one discriminates the PM signal, one can also derive the efficiency by counting
the events and comparing to other PMs. But a better setup is the following.
In this setup we get all informations like in the first setup (splitter with good Bandwidth is needed).
In addition we get the information about transit time and other timing behaviours of the PM.
The fraction of detected events in coincidence with the trigger PM gives us a comparison of
efficiencies. All results has to be compared to a reference PM, that was in our case a comlete
new XP2020Q by photonis .
The real setup can be seen in the following picture. The heart of this setup is the box with
the optics inside. The box could be much smaller, but this way we have a box with a lot of
possibilities ;).
The most important results can be found in the download area. But what everybody wanted to see is the following comparison of counting rates normalized to the XP2020Q.
One expects from the datasheeds 1.2 to 1.3 better performance of the CEDAR PMs Thorn Emi 9820 UVA compared to the XP2020Q. One PM was noisy and will be replaced (C2PM2). The other PMs seem to be OK.