Testing PMT's

Discussion in 'Test & Measurement Forum' started by Davall, Apr 5, 2010.

  1. Davall

    Thread Starter New Member

    Aug 29, 2009
    Does anybody know of a way to testing PMT's so degradation can be detected?
    What I have is a PMT, an LED, a scope and software to activate the PMT with the LED to read the gain with the scope. Under normal conditions the PMT's are used under X rays......of course, I can only test without the X rays.(I have to test without X rays to find a balance between those that are under the rays, so I can relate/compare to PMT's not under the exposure of x rays and prepare for breakdowns) Recently, I've pulled a PMT out of a machine as it was getting noisy. I've compared this to a new one and cannot find any differences in the output gain...actually if truth be told the new PMT was worse than the one I took out of service. Any suggestions, procedures or references would be greatly appreciated.
    Thanks guys.
  2. bertus


    Apr 5, 2008

    Take a look at the attached PDF.

  3. justand

    New Member

    Dec 31, 2017

    OMG. The machine you work on may hurt people.
    I know this was very long time ago, and you probably forgot the whole thing but from the way you described it - it is outright scary.
    PMT degradation happens on the anode - if the expected input signal exceeds the allowable threshold and the anode overheats.
    This can happen. I experienced that. But then you mention noise and that is entirely different matter. So you have 2 separate issues really.

    From the way you described it - it sounds like you are experimenting with either a Rapiscan or some cadscan machine and you don't know much what you're dealing with. If you don't know what you are doing - then the best course of action is to leave it alone. Just don't play with it.
    PMTs are very complex devices and there are issues with them, that are not easy to deal with. I worked with PMTs in imaging applications and I made some startling discoveries so I know you are stepping on a minefield.

    Noise in PMT has a lot of sources. First, you get Shot noise (Poisson noise), then you have noise from electronics - the inverter, its power supplies, and then the PMT power supply, voltage divider, and finally the dynodes. Then you get gain distortions from the divider under load. If you happened to have a perfect power source for the PMT and a perfect electronics for it - say - within a single LSB of an attached 16Bit ADC converter (doable) - then you would be able to measure the dynode noise amplification factor - which for Hamamatsu tubes is aproximately 3%. This means that the noise you get on the PMT output for a particular input light signal will have a 3% dynode noise component. This is very low. It makes sense to use the PMT gain and limit outside circuits gain. There are methods to increase the capture efficiency and lower the noise of the system.

    But then there are issues with the HV. For instances - the Hamamatsu PMTs have gas inside and the maximum allowable HV value as stated in their manuals is sometimes much lower. I am not giving you a specific advice, but a general area of search. Say for instance you have an 8 dynode setup and the voltage at which the leakage start to occur is 70V. This will exhibit a large noise source - that would be isolated to 2 dynodes only. So you may be looking at entirely different problems and obviously not being aware that you are looking at a failure really.

    70V is quite high for a dynode to dynode area. They aren't perfect. They are made manually. By hand. People aren't that perfect.
    That is why it is so important to take time and understand the problem. But having a problem in a lab with a gamma camera or a counter is entirely different matter than a problem in a working X-ray system.
  4. spinnaker

    AAC Fanatic!

    Oct 29, 2009

    The thread is 6 years old. The TS has only 3 threads and hasn't posted in 4 years. You just wasted a whole lot of effort.
  5. justand

    New Member

    Dec 31, 2017
    People sometimes use search function for older threads. I have heard this question in the past over and over again. So let this be for any future inquiries. One day - someone will come and want to play with a scanner and maybe this will stop them.
    kubeek likes this.