Connect with us

Major PMT noise problem

Discussion in 'Electronic Design' started by Chris Carlen, Dec 10, 2003.

Scroll to continue with content
  1. Chris Carlen

    Chris Carlen Guest

    Hi:

    I have a PMT enclosure consisting of a 2" dia. x 4" long aluminum
    cylinder, with a BNC on one end. Inside, the BNC shell connects to the
    AL housing, and to the braid of the coax that leads to a potted
    Hammamatsu H6780-06 PMT module with integrated HV supply. There is also
    a 15V supply cable, but that is disconnected for the moment. The center
    of the BNC goes via the center of the coax from the PMT housing to the
    PMT anode.

    So we have nothing but a little can with a BNC.

    A 50R coax connects the can to a scope, which terminates the line in
    50R. A photodiode watches a pulsed YAG laser nearby, also going to the
    scope. Triggering from the photodiode light pulse, the scope shows a
    20-50mV p-p 250MHz or so burst of ringing about 50ns long coming from
    the PMT can, at about the same time as the photodiode light pulse. The
    rest of the signal from the PMT is quiet, except when the laser fires.
    There is no noise when the laser is off.

    There is another PMT housing from a commercial manufacturer (the
    Hammamatsu enclosure is custom made), that is also connected to the
    scope. It like the photodiode shows negligible noise of 1mVRMS or so.

    I have seen tremendous ground bounces and other EMI before from the YAG
    lasers. In this case, the photodiode as well as the other PMT seem to
    avoid picking up this interference.

    But the home-made PMT enclosure seems to amplify the EMI like crazy.
    Interestingly, if I simply disconnect the scope cable from the BNC on
    the little PMT can, the noise goes away. Shortening the coax from the
    scope to the PMT can down to 12" has no effect, still the same level of
    noise.

    On a theoretical level, what I think is going on is that there is a
    common-mode noise pulse carried by the scope cable, perhaps even the
    entire scope is carrying this EMI pulse, which is originating from the
    laser. This common mode signal is being converted into a differential
    signal in the can, then propagated down the coax back to the scope. How
    such a conversion could take place when the can is connected to the coax
    braid I dunno.

    I must somehow make this PMT not do this, and have a noise baseline
    similar to the other sensors around 1mVRMS, with no bursts when the
    laser fires. I am down to sticking the wires through a ferrite bead,
    which attenuates a bit, but doesn't solve the fundamental problem.

    Must I isolate the PMT from the output connector, by putting a buffer
    amp in the can? Or is there a better shielding and connection
    arrangement that will avoid the noise pickup?


    Thanks for comments.

    Good day!



    --
    ____________________________________
    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
     
  2. Sometimes such interference can be fixed by moving cables around.
    Filtering the power going to the scope (or using a battery operated scope).

    When you get desperate, try tri-ax or dual BNC's and coax, and use a
    balun transformer to make the single-ended signal differential to the
    scope, which you set for differential inputs A-B.

    --
    Scott

    **********************************

    DIY Piezo-Gyro, PCB Drill Bot & More Soon!

    http://home.comcast.net/~scottxs/

    **********************************
     
  3. Just remembered that a 2" x 4" cavity is just about the right size for a
    250 MHz helical resonator. Perhaps some ferrite or carbon could help
    spoil the Q of the cavity?

    --
    Scott

    **********************************

    DIY Piezo-Gyro, PCB Drill Bot & More Soon!

    http://home.comcast.net/~scottxs/

    **********************************
     
  4. Genome

    Genome Guest

    I have recently redefined myself as God. I use the term God because it
    appears to be Universal, but I'm just making a vacant claim there. I think I
    might be globuversal until I sign up for membership of the community of the
    intersparklythings you see at night. AKA, for some strange reason, the
    Cosmos.

    Is $50 a good deal?

    Unfortunately your part of the experiment means you receive the message via
    others in a distorted form. My other mate, called ground, insists that you
    think about things and refer all things to him (I'm a fucking sexist me) at


    Oh, I went for a job interview. The recruitment agent said I should know
    about EMC. The person at the interview asked about EMC.

    Hey, life's a bitch. If you're asking questions like that.... then you'll
    probably want to know about how I work in a team or perform under pressure.

    DNA
     
  5. Damn, you're starting to sound like this person I know who gave up the piss
    and became a vegetarian. I hope you haven't stopped eating fish.
     
  6. John Larkin

    John Larkin Guest

    Do you have a TDR? It would be interesting to compare the various
    cans.

    John
     
  7. Chris Carlen

    Chris Carlen Guest


    None handy.

    Today I plugged the same cable into another PMT enclosure I made, this
    one make of a hogged-out block of solid aluminum, with 1/4" thick walls,
    about 2.5" x 2.5" x 6" interior cavity for the tube, and a smaller 1.5"
    square cavity for the connections. This one did the same thing.

    I think I may be getting into trouble with this: The BNC is an
    "isolated" one with the ground lead connecting to the enclosure via a
    short wire. This creates a loop of 1-2 cm^2 or so. Perhaps this is
    turning the cavity into a resonator.

    I will try to find a BNC with the shell metal going through the chassis,
    so there will be no little loop connecting the ground to the chassis.

    Argh!

    Thanks for the input.

    Good day!


    --
    ____________________________________
    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
     
  8. You do have both sensors, cans and cables mounted electically isolated from
    ground, I hope? A groundloop near a big fast driver is a great antenna.

    Try holding the outer side of the BNC cable part to the outer side of the
    chassis part. Do NOT connect the inner pin to anything. Does the noise stay
    away, just as when the cable is fully unconnected? If not, it's a ground
    loop.

    Regards,
    Arie de Muynck
     
  9. N. Thornton

    N. Thornton Guest

    Re: Major PMT noise problem

    A few men complain of this, the only upside is it will pass. I think
    the mericans call it PMS.

    Regards, NT
     
  10. I read in sci.electronics.design that N. Thornton <>
    You mean the nervousness of folk singers before they go on stage - Pre-
    Minstrel Syndrome?
     
  11. Robert Baer

    Robert Baer Guest

    For grins, treat the can like a microphone, with the window as if it
    was the diaphram.
    Try pointing it in different directions, and record pulse amplitude VS
    angle fro EMI source.
    The results may give a clue.
    Wild guess: you have a resonant cavity with coupling to the outside
    via the window, making it act like an antenna.
     
  12. N. Thornton

    N. Thornton Guest

    Yes, that does sound like a major noise problem.

    Regards, NT :)
     
  13. Try placing the PMT and scope on a common ground plane.
    Ferrite bead on the scope probe shield.
    Connect the PMT Shell to ground with a braid.
    Connect Scope to ground with a braid.
     
  14. Chris Carlen

    Chris Carlen Guest

    The little loop of wire connecting the BNC connector shell to the can
    was the culprit. Basically, you can't do it that way. Using a
    non-isolated BNC made the problem vanish.

    So I think I was correct, that the common mode noise on the ground was
    converting to differential, by being broadcast inside the can by the
    loop area between the cable ground and the can, then picked back up by
    the center conductor circuit.

    Thank goodness for EM theory. Much better to be able to think through
    problems than having to rely on rules of thumb and hocus-pocus.

    And thanks to Arie de Muynck for the sugestion to see if the noise is
    there when connecting only the grounds, and not the centers. For if
    that doesn't make the noise appear, which it of course didn't, then it
    is a problem like I have described.

    Good day!



    --
    ____________________________________
    Christopher R. Carlen
    Principal Laser/Optical Technologist
    Sandia National Laboratories CA USA
     
  15. I read in sci.electronics.design that Chris Carlen
    Although some EMC gurus preach that 'pigtails are anathema' (and justify
    that with the impedance figures of a 5 cm one) it is somewhat unusual
    for such a dramatic effect to be caused that way, but 'best practice' is
    always to connect connector shells to the enclosure directly, and
    preferably with a 360 degree contact.

    Two enclosures connected by a shielded cable should be like two water
    tanks connected by a pipe - no leaks!
     
  16. Roy McCammon

    Roy McCammon Guest

    John, I used to work out at McDonald observatory
    and had similar problems. When flexible coax did
    not work (even six inches), we would make "solid
    coax" out of BNC barrels and right angle connectors.
    I don't know why, but it made a difference.
     
  17. I read in sci.electronics.design that Roy McCammon <>
    Your flexible coax didn't have a dense enough outer conductor.
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-