Photomultiplier threshold detector: need advice

Discussion in 'The Projects Forum' started by ElectricAye, Apr 6, 2008.

  1. ElectricAye

    Thread Starter New Member

    Apr 5, 2008
    3
    0
    Greetings,

    I have a photomultiplier tube (PMT) that I hope will work to output pulses on the order of 10 nanoseconds wide and which will occur approximately every 100 nanoseconds. The PMT will also output some noise, so I would like to build some kind of threshold detector that will give me a countable output pulse only when the output pulse from the PMT exceeds a threshold. I would like the level of the threshold to be adjustable. I'm not very experienced with building electronics but I'm guessing I would need something like a comparator chip. But I'm not sure. I'm also not sure I can solder together a circuit with regular "Radio Shack" tools that can give the kind of frequency response I described above. Can anyone suggest a chip and/or tell me if building a circuit like this by hand is even feasible for an amateur?

    many thanks,
    Mark
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Hi Mark,

    I haven't experimented with photomultiplier tubes before.
    Wikipedia has some very useful information:
    http://en.wikipedia.org/wiki/Photomultiplier
    There is a link in the article to a Photomultiplier Handbook, that contains loads of information. I'm still wading through it at the moment.

    One of the requirements for a PMT is high voltage, on the order of a couple of KV.

    Which PMT do you have? A 931A?
     
  3. beenthere

    Retired Moderator

    Apr 20, 2004
    15,815
    282
    Take a look at the Linear Technology LT1016 comparator. They may have an application note about it as well as the data sheet. Now I can't recall the company that made the counter equipment for the research reactor, but you may need to get properly designed instruments to work with your PMT.

    You get 290,000 hit on Google using "single photon counting" as the search term. I notice there's lots of silicon detectors mentioned. Are you stuck with the PMT?
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Along with Beenthere's suggestion:
    There are a couple of Application Notes (AN13, AN72) on Linear Technology's website that you'll find very helpful. They're available for download on the right side of this page:
    http://www.linear.com/pc/viewCategory.jsp?navId=H0,C1,C1154,C1004,C1012

    The LT1016 and LT1394 high speed comparators are discussed in these AN's, along with avoiding numerous pitfalls.
     
  5. ElectricAye

    Thread Starter New Member

    Apr 5, 2008
    3
    0
    Thanks, Sarge... thanks, BeenThere,

    To answer your question, Sarge, I am planning on using a Hamamatsu R5984HA tube, which is maybe about 7 years old and might be too old to be useable anymore due to helium leaking through the glass. I already have a socket that has a built-in power supply and voltage divider but it's the signal processing I need to learn about. Thank you both for the links to the literature on components. But I guess what I really want to know is: can an amateur build the kind of amps and threshold detectors that are needed to interface the output of a PMT into a standard frequency counter? or does this require customized printed circuit boards, etc. to handle the approximately 10 nanosecond wide pulses, etc.? I simply don't know what sort of level of competence is needed to build something like this from scratch, so I'm wondering if I should even try.

    thanks again,
    Mark
     
  6. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    10 nanoseconds corresponds to 100 Mhz, which is within the reach of standard logic chips with an S or AS suffix. thus 74##S. (not 74LS, they are not fast enough)

    You rightly surmise that you need to turn the short 10ns pulse into a longer one, say 40ns if they are going to occur every 100ns. To do this feed the output of the comparator to a monostable. The output from this will drive any modern counter, 10mhz (the pulse repetition frequency) is not at all fast.

    And yes you should be able to construct at this frequency it is the same as the FM radio band (88-108Mhz)

    If you need more post again.

    http://www.interfacebus.com/Speed-Power_Chart.html
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, 100MHz is kind of on the bleeding edge of 74F-type logic (F=Fast). For example, TI guarantees their SN74F163 binary counter to have a minimum Fmax clock speed of 90MHz, but with careful planning you could probably get 120MHz out of it.

    ECL (Emitter Coupled Logic) might be worth considering. ECL uses a LOT of power, therefore it's good for northern climates where an extra room heater can come in handy. Seriously though, the big advantage of ECL is the very low propagation delays that are possible.
     
  8. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    That's why I suggested 74S series.

    74S is guaranteed to 120 Mhz
    74AS is guaranteed to 200 Mhz

    Note these are switching frequencies.

    I don't think propagation delays are an issue here.

    The interface is simply to stretch the pulse from the photomultiplier to make it accessible to the counter. There is a trade off between length of the stretched pulse amd the gap before the next one, which might be a bit sooner than 100ns.

    This is where I started electronics, with counters in a in a radiochemistry lab, but we had to roll our own with transistors in those days.
     
  9. ElectricAye

    Thread Starter New Member

    Apr 5, 2008
    3
    0
    I was going to shop for a frequency counter that was good up to about 1 GHz. I thought that would be fast enough to capture pulses that are 10 ns wide. Am I right about this or am I missing something? I'm afraid if I stretch any pulses, then I might lose good pulse counts due to signals overlapping too much. True?

    many thanks again,
    Mark
    :)
     
  10. studiot

    AAC Fanatic!

    Nov 9, 2007
    5,005
    513
    When counters were first produced numerical counting (which is what you require) was the principal function. Others such as frequency or period were a bonus.

    Nowadays many have limited or dropped the counter function in favour of frequency so check before you buy.

    Aye that is exactly what I meant by saying it's a trade off. However I was using the figures you originally provided and you stated you had a counter.

    The specs of a new counter will tell you if it will do the job or not, look at the minimum latch or capture time.
     
Loading...