Connect with us

Ground fault clamp thingummy. Sorry, non-political post.

Discussion in 'Electronic Design' started by Keith Wootten, Sep 17, 2004.

  1. Mah fellow Armenians,

    Any advice on this test-bench safety problem?

    I have some metal-cased equipment whose body should be nominally at
    ground potential, held there by a cable shield. The cable is very long
    and I can't ground the equipment body locally as the resulting loop
    interferes with its operation.

    On the test bench, it's possible that a fault (say in the cable shield)
    could cause the equipment housing to reach an expletive-inducing high
    voltage.

    What I need is some sort of clip-on electrical clamp which grounds the
    housing pending a reset if it gets more than a few volts away from the
    local ground for a few milliseconds but otherwise leaves it more-or-less
    floating by at least a few kilohms. Phew, that's a lot of fews. A
    ground fault interrupter on the power supply won't be enough as there
    may be considerable capacitance in the supply and cable.

    It's not really a design query; I'm sure we could all come up with
    something workable, and Fred's probably already posted six designs by
    the time you read this, but this needs to be a commercial unit for
    liability reasons. Is there something available to do this? What's it
    called?

    TVM
     
  2. I read in sci.electronics.design that Keith Wootten
    'Ground fault clamp thingummy. Sorry, non-political post.', on Fri, 17
    Sep 2004:
    Since you are in UK, I can tell you that it is almost certain that any
    such expedient as you propose is illegal under Health and Safety at Work
    legislation. It appears also to violate the earthing requirements of BS
    7671, the IEE Regulations, which are legally binding in Scotland at
    present, but will become so in England and Wales in the foreseeable
    future.

    Can you explain how this 'loop' arises, and what effect it has? How long
    is 'very long'? This raises a 'remote earth' issue, possibly violating
    'equipotential bonding' requirements. What sort of cable shield fault do
    you have in mind?

    There is almost certainly another solution to your problem which doesn't
    fall foul of any safety requirements. But more information is needed in
    order to pin it down.
     
  3. John Larkin

    John Larkin Guest


    Git a big ole 35 or 50 amp bridge rectifier brick. Short the + and -
    pins, and use the two AC pins as a bidirectional clamp diode.

    If that's not enough of a drop, the next step is a big SCR or
    something.

    John
     
  4. Ken Smith

    Ken Smith Guest

    When testing, could you unground the far end and ground the unit right at
    the bench? This should remove the ground loop.
     
  5. I read in sci.electronics.design that Keith Wootten
    'Ground fault clamp thingummy. Sorry, non-political post.', on Fri, 17
    Sep 2004:
    My response to your e-mail was rejected. The rejection message rubric is
    copied below, with the e-mail address munged with (at) and exes.

    This message was created automatically by mail delivery software (Exim).

    A message that you sent could not be delivered to one or more of its
    recipients. This is a permanent error. The following address(es) failed:

    keith.wootten(at)xxxntlworldxx.com
    SMTP error from remote mailer after RCPT TO:<keith.wootten(at)xxntlw
    orldxxx.com>:
    host smtpin.ntlworld.com [62.253.162.40]: 550 Invalid recipient:
    <keith.wootten(at)xxxntlworldxxxx.com>
     
  6. Yes, but you may be working on either end of the cable and still relying
    on one ground. A clamp as described would be handy if one could be
    found - the test engineer would simply clip it on to the unit under
    repair knowing that it won't interfere with normal operation. It could
    even incorporate some sort of warning if not connected, but as I said,
    I'd like to find something commercially.

    Cheers
     
  7. Ken Smith

    Ken Smith Guest

    Write into the test procedure that the ground must be connected on the end
    that is under test. Make it clear that anyone not doing so will be fired.
    If some gets electrocuted, fire them before they hit the ground :)
    In the SanFansisco bay area is BART that had to solve this same sort of
    problem with making the return current go back to the right substation.
    Perhaps, you can try your local electric transit agency to see if they
    have something.
     
  8. Mac

    Mac Guest

    So, the equipment has a ground shield, but you want a local ground because
    you don't trust the shield. This part really makes no sense to me. If the
    equipment is designed correctly, why would the shield be faulty?

    Can't you just test it before you start messing with the equipment?

    You might object that testing the ground connection of the equipment is an
    extra step. Since extra steps can be forgotten, the technician might be at
    risk of shock if he or she forgets to test the ground. But connecting some
    special local ground is ALSO an extra step which could be forgotten, so I
    don't see how you win.

    Besides, how is the local ground more trustworthy than the remote one? Is
    it bare cable, bolted directly to a spike hammered into the Earth right
    at the test bench?

    Also also, unless the power cable is really ridiculously long, the
    resulting ground loop shouldn't cause a problem. If it does, it probably
    reflects a flaw in the equipment.

    Sorry for being so negative, but the whole thing sounds a bit hokey to me.
    If all of your fears are well founded, it seems as though the piece of
    equipment you are talking about is very poorly designed.

    --Mac
     
  9. Ken Smith

    Ken Smith Guest

    I've seen the following:

    (1) Because the shield has blown open due to an incorrectly applied mains
    voltage.

    (2) The thing did not get built according to the drawings.

    (3) The ground of the outlet it is plugged into is not actually grounded.

    (4) The other end of the cable is 4000 feet away and it really is
    grounded but due to IR drops this end has +60V on it.

    Usually (where I work), all of the outlets have a "ground" pin and all of
    the metal of the bench and the grounds of the equipment are all connected
    to it. It really doesn't matter if it is ground or +500V so long as all
    voltages are referenced to it. If someone brings a real ground connection
    into the picture it really starts to matter.
     
  10. Rich Grise

    Rich Grise Guest

    I've always wished someone had come up with a name of "Frisco Area Rapid
    Transit."

    :)
     
  11. Jim Thompson

    Jim Thompson Guest

    [snip]
    They're trying to establish PHoenix Area Rapid Transit here (PHART),
    but I think it will fail at the next election.

    ...Jim Thompson
     
  12. John Fields

    John Fields Guest

     
  13. John Larkin

    John Larkin Guest

    Lake Tahoe has TART.

    John
     
  14. Didn't the old voltage-operated earth leakage trip
    operate in exactly the way you require? It sensed
    the voltage difference between true earth and the
    'earth' wire in (say) domestic mains wiring, and
    if the voltage difference rose above a few volts
    it would trip. There was one in this house when
    we first moved in, mfr'd by Chiltern afair.
     
  15. Tim Shoppa

    Tim Shoppa Guest

    The standard solution (well, other than dyking off the ground pin
    from the mains cable): two channels to scope, one of them the signal in
    the tested equipment, the other the ground from the tested equipment,
    and scope in differential mode.

    Other than that, I can only recommend isolation transformers etc. as
    a way of getting around existing ground loop problems, and even these
    have safety issues if you don't properly do single-point grounding to
    the clean side.

    Tim.
     
  16. Mac

    Mac Guest

    Maybe I misunderstood the original poster. I guess I was thinking that
    the piece of equipment under discussion was test equipment, and that is
    why I couldn't understand the reason for the doubt. But maybe the
    equipment under discussion is actually something that is in his shop
    for repairs. In that case, it sort of makes sense that the ground
    shield could be blown. Still, it seems as though it would be reasonable
    to just check the ground shield with an ohm meter prior to plugging the
    thing in.
    Same comment as above.
    Easily tested and fixed if it is your shop. Besides, you could say the
    same of the local ground.
    It seems to me that the equipment under discussion should have a very
    secure bonding of the case to ground by way of the three conductor power
    cord. The ground conductor in the power cord should have a low enough
    resistance that it will sink a LOT of current in a fault condition,
    thereby causing a fuse or other protection to shut off the circuit and
    remove the hazard. So the 60 V should disappear almost as fast as it
    appeared.
    Right. Where I work, almost all the equipment is grounded to the green
    wire ground for the building. Some equipment uses two-wire plugs, but that
    is relatively rare. Our benches aren't metal and aren't grounded except
    weakly due to the static mats.

    The building code requires the green wire ground to be connected to an
    earth ground, though I've never actually checked it. (We don't use any
    high-voltage DC equipment).
    I don't know, the whole thing just sounded a bit weird to me.

    --Mac
     
  17. In message <>, Tony Williams

    That's almost it, thanks, but I really want the tripped condition to
    switch a ground to the unit under test; I presume the Chiltern device
    removed power when tripped.

    The cable is very long and even when operating normally, the IR drop
    will ensure that the unit's body is not at local ground potential. There
    is no possibility of using a different cable, and grounding the unit
    locally stops it from working correctly. It's not a question of poor
    design, rather of something working close to the limits of feasibility.
    Sorry I'm short on detail, but you know the score.

    Thanks for the help everybody, it pretty much confirms that there's not
    an off-the-shelf solution to this. A bridge rectum-fryer, a zener and a
    sensitive NC relay to detect the voltage and a second double pole relay
    to self-latch and switch the ground connection through when de-energised
    should do it, packaged in a wall-wart with test and reset buttons and
    status lamps. Whether it's worth doing is a question for Ron.

    Cheers
     
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

-