Connect with us

Ground Loop Problem!!

Discussion in 'Electronic Basics' started by Adrian, Jan 12, 2004.

Scroll to continue with content
  1. Adrian

    Adrian Guest


    I have two devices:

    1) Power is supplied from a 120VAC to 12VAC Adapter (not grounded)

    2) Power is supplied from 120VAC (grounded) to an internal switching power

    I would like to have the two devices physically connected and since they
    both have metal chassis', this entails electrical connection as well.

    The problem is that there is a potential difference between the two chassis
    and whenever they touch, they conduct current through a ground loop causing
    a multitude of problems, from blowing transistors, to malfunction of the

    Is there any way to solve this other than to physically separate the

  2. CFoley1064

    CFoley1064 Guest

    If your description of the situation is correct, I think you should first look
    at replacing the 12VAC wall wart. Before replacing it with an adapter or
    transformer that's guaranteed to be isolated, use your DMM to check if one side
    of the 112VAC is grounded or has high leakage to ground. Place a 47K 1/4W
    resistor from each leg in turn to GND, and measure the voltage across the
    resistor with the wall wart plugged in. One leg of the AC from your adapter
    may be grounded, or you might just have sufficient leakage current to read a
    significant voltage and cause problems. Wall warts are usually made just good
    enough, but that might not be good enough isolation for your application.

    You haven't given a complete description of your AC adapter, but I would start
    by looking at the Stancor 4112 ([email protected]). Unless you have a "medical
    isolation"-type application, that should be sufficient.

    Another thing you might want to try is using "cheater plugs" to get the ground
    out of the equation, at least for a test to see if the electrical problems stop
    or change.

    If that doesn't solve the problem, your easiest route would be to contact the
    manufacturers of the devices and ask them for assistance. Without knowing more
    about the devices or your application, it's difficult to give advice. And
    sometimes there's a "gotcha" that isn't obvious.

    This type of thing is a "common" problem (sorry ;-), but there's usually some
    kind of solution. Luck is the residue of hard work.

    Good luck.
  3. Adrian

    Adrian Guest

    Hi Chris,

    Thanks for the reply.

    I checked the wallwart with a 100K resistor and measured 2V potential to
    ground. I also measured the open circuit potential to ground, and on the
    center pin of the adapter it is 12V!

    This seems to indicate the leaky adapter as you suggested, correct?

    If so, is there any way to check an adapter for isolation before purchasing
    it? Using this same method perhaps? What is an acceptable potential for


  4. CFoley1064

    CFoley1064 Guest

    Hi Chris,
    Wall warts made by reputable "name brand" manufacturers are tested for
    compliance to UL specification for hipot and leakage before shipment. These
    specifications can usually be obtained by looking at the manufacturer catalog.

    You could do worse than look at a Stancor product. Their wall warts are
    available through standard distributors such as Newark and Allied. Since there
    isn't any information available as to the type or size of wall wart, you can
    choose one yourself which has a current rating greater than or equal to the one
    you're using:

    (By the way, if the two devices are permanently affixed together, you probably
    don't need a GND pin on the plug).

    If you don't mind possibly having a few return trips, just buy one and test it
    as above. If it doesn't work, take it back and buy another one somewhere else.

    Good luck
  5. Jim Thompson

    Jim Thompson Guest

    You have a serious problem. Please get professional help.

    ...Jim Thompson
  6. I suspect capacitive coupling between the primary and secondary circuits
    of the power transformer. If the transformer is not shielded this leakage
    capacitance couples common-mode signals to the secondary.

    Euthymios Kappos
  7. Capacitive coupling in a working transformer is only going to be a few pF,
    certainly much less than 1nF. At 120V 60Hz, a 100pF capacitor will pass
    5uA, if my math is correct. That's not enough to blow transistors (as the
    OP claims is happening). If he's getting that kind of current between the
    chasses of two mains-powered devices, then as Jim said, he's got a serious
    problem and he should seek the help of someone who's got test equipment and
    knows how to use it.
  8. Genome

    Genome Guest

    Mr OP doesn't describe his set up too well, oh dear......

    I think....

    C = Eo.Er.A/S

    Take a small switch mode transformer with a width of 25mm and a winding
    length of 25mm for an area of 25mm x 25mm.

    A = 6.25E-4 M^2

    Eo = 8.85E-12 Something or another

    Er = oh..... call it one

    S = 1.5E-4 M

    S is the winding separation. 3 layers of 2mil (mixed units!!) yellow stickey

    C = 8.85E-12 x 6.25E-4/1.5E-4
    C = 37pF

    Turn 120VAC into 170VDC for the supply. Switch at 100KHz and ignore the
    harmonics. That's about 4mA.

    Not totally insignificant...... but, sort of, not enough to blow up


    It's an SMPS with feedback from secondary to primary via some thing,
    probably an opto-coupler. If you don't pay attention to a bit of judicious
    capacitive/resistive coupling between primary and secondary then the
    resulting common mode hash is likely to send the feedback loop to apesville
    and things might well go bang.

    If you rolled you own then you'd make sure it wouldn't.

    Then you'd run it up and see what the nasties were. Couple primary to
    secondary wiv a bit of Class Y. Preferably, primary to ground and then back
    from ground up to secondary.... using two of them and taking care of
    regulatory requirements. Depends on where you live.

    See how it rings.... wiv your 20MHz analog(ue) scope, do some sums.... add a
    bit of appropriate R in series with the Y to settle things down.

    If you've bought an El-Cheapo from wing-tong-tiddle-aye-po then life might
    become a bit harder.

    There you go. I do know about EMC type stuff, next question.

  9. GPG

    GPG Guest

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