Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Rising voltage leak to 100v AC - 220v 50Hz 35uF 1.5KW motors

  1. Oct 9, 2016 #1
    I noticed that I could get a minor belt off my 3 x 1.5KW compressor.

    I attached my meter to a mains supply 'socket earth', and to 'the chassis'.
    I watched the AC voltage gradually rise... perhaps increasing in speed slightly as the potential difference increased between the chassis and earth... as it approached 100v AC it was rising at a rate of 4v per second.

    At 100v AC (on the meter) I switched the motors off, and the voltage dissipated immediately.
    I presumed the voltage would keep rising until a failure occurred.

    When connected to Neutral (which is connected to a socket earth) - the voltage stabilised at 10v AC.

    The amps flowing between chassis and earth were around 1.4mA.
    Clearly, the chassis needs earthing to the Neutral.
    However, is a capacitor failing, or is it the motors?

    Can anybody explain what is happening?
  2. jcsd
  3. Oct 9, 2016 #2

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    In the US

    while earth and neutral are connected at the service panel you cannot use one for the other's purpose. What you describe, neutral tied to socket earth, sounds highly irregular.

    First thing i'd do is check that socket to make sure its earth connection is indeed earthed by your house wiring,
    These work fine, six bucks or so at Walmart

    The chassis of your compressor should be tied to the earthing wire in its power cord.
    In US that's the green one which is on the longest of the three prongs so when you plug it in earth makes first to prevent a shock.

    I'd say
    if your receptacle earth checks good you have a broken or disconnected earth conductor in the power cord. And likely some leakage inside a motor.

    It is not uncommon in US to find the earthing prong cut off a power plug when some unaware individual has used it with a two prong extension cord.

    Be careful with that thing, something is wrong. Verify receptacle is earthed, that is important.

    old jim
  4. Oct 9, 2016 #3

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    ps get the tester with a button for checking GFCI outlets, sometimes pressing the button helps flesh out high resistance connections. It's only about $3 more and worth it.

    If you're not handy around house wiring find an electrician. If you show interest he'll be happy to share knowledge with you.

    If you're gonna throw caution to the wind throw the breaker first.

    old jim
  5. Oct 9, 2016 #4
    Of course..... of course.... Sorry... how stupid of me.
    Sometimes the fog of war descends and.......

    Anyway thanks for the kick Jim.
    Yes Neutral is Neutral and earth is earth........ and the pillocks have cut off the earth cable.

    The lesson of course, is to imagine that this is not your problem, and look at it dispassionately.
    So I will be remaking the cable tails, and introducing an earth - chassis connection.

    Okay.... that said...
    Do you have any idea why the chassis is gaining a voltage as described above?

    I am sure that this is not normal.
  6. Oct 9, 2016 #5

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Only guesses at this point. My guess is a capacitive voltage divider between the windings of the motor and your meter.

    What kind of meter do you have ?

    The little el-cheapo analog meter that i just love for troubleshooting electrical stuff has a 1 ma movement. It measures AC by half wave rectifiying it and displaying average on a scale that converts average to RMS . Works fine so long as you keep in mind what it's really doing.

    The average of an AC sine wave is zero and that's what the meter reported initially. Rectifying it with a voltmeter charges the winding capacitance to some DC and that's what my meter would show.

    If your meter is not a rectifying type then that idea goes out the window.

    Now the 1.4 ma (i assume AC) that you measured is what 125 volts would push through about 0.03 microfarads, 30,000 picofarads which seems high to me.
    I hope you'll repeat that measurement after your motor has run long enough to warm up and dry out. Were it mine i'd take off the end bells , blow out the dirt and look for signs of corrosion..

    I nave fixed damaged motor windings with a good cleaning and application of varnish.
    GE Glyptal is great stuff if you can find it. It's an industrial product.
    For home hobbyist repairs i use good old polyurethane woodworker's varnish from the hardware store.

    old jim
  7. Oct 9, 2016 #6
    The worst thing about the physics forum is that you have to jump through hoops to upload an image.
    Just a little rant.... but this is one of the few remaining forums where you can't simply upload an image.... and it's 3.00am and I'm not gonna start looking at how to do it tonight.
    Tomorrow, i'll upload an image of my multi-meter.
    It's a reasonable piece of kit.... ohms, hfe, ac, dc, amps, uF, battery.
    It's old now.... but it's what I have.
    Whether it is the rectifying type..... I'm not sure.

    I'm a mechanical engineer..... with a general grasp of electronics...... the earth embarrassment was just that.... it just shows that we are all human.

    Re the fault...... the rising voltage appears as if the chassis is behaving like a capacitor.... gradually gaining charge.
    The capacitors themselves are big beasts: CBB60 35uF +- 5% class B 450v AC 50/60Hz.

    The motors are new.... only a few minutes work.
    The conditions are dry..... and the motors got hot.

    To get to 100v on the chassis, was done with the air outputs open.
  8. Oct 9, 2016 #7

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    snip or copy image ,paste into paint, save(i use jpg they're smaller files than png ) use upload button


    nothing should be connected to earth, neither motor windings nor the 33uf capacitors

    i dont know why it drifts up a 4 volts AC per second

    it sounds like capacitance charging through a rectifier.

    It's that new ? I think replacing the ground wire will fix your trouble.

    If there's insulation failure in progress it'll either dry out and be fine or it'll speed up and self announce.
    I'd run it on a GFCI outlet so that fault current will de-energize it before current gets large enough to do a lot of damage.
    1 milliamp wont trip a GFCI but 50milliamps will and that's not much of a spark.
  9. Oct 10, 2016 #8
    Nice one Jim.
    I don't know how many posts I've made on this forum, and never clicked that the image upload button is hidden amongst the 'post reply and preview' buttons.
    I've been stupidly clicking the image button in the menu :confused:
    It's very funny :biggrin:
    Here's a pic of my meter:

    Surely the chassis should be connected to earth?
    The earth cable should never have been cut off.

    Anyway, that's the job I'm doing now.
  10. Oct 10, 2016 #9
    Job done!
    The motors seem to run better.

    I tested the voltage between neutral and earth.
    When cold the results were variable.
    But once the motors were warm, they settled to around 2.5v each.
    With no motors running, there is 0.5v between earth and neutral.

    At any rate.... that's the basic safety problem solved.

    RE the original voltage build up on the chassis.... I just don't know enough about how the leakage occurs in the first place on motors.
    Apparently, some leakage is to be expected.

    The motors are rated at 35uF, but could they be too big, and the energy is not being used by the motor, while it is running...
    ... so they ultimately discharge more and more voltage to the chassis?
  11. Oct 10, 2016 #10

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    EDIT i see you're ahead of me , well done !

    yes, poor wording on my part.
    Neither of the two power wires at plug should be connected to chassis except by the small capacitance of the windings and the high resistance of the insulation.

    Curoiusly on your meter i dont see a scale for AC milliamps, only DC.
    So i don't know what to make of your current reading.

    Check the resistance from both power prongs to the ground prong of your cord to before you plug it in. Both should read megohms.

    old jim
  12. Oct 10, 2016 #11

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    Reciprocating compressors present an uneven load to a motor. On the compression stroke they're hard to turn so motor current goes up. Motor current actually cycles low to high on every revolution.
    If your compressor is large enough to have an unloader instead of a motor cutoff, motor current will steady out when unloader actuates.

    Voltage from neutral to earth is the drop along neutral wire caused by motor load current. A couple volts sounds okay for three big motors. It'll fluctuate with that cyclic torque presented by the compressor as motor current cycles along with torque.

    Sounds like you had a success.

    Sorry i don't know more about the inner workings of your meter. I'm an old analog guy who uses DMMs only when i need high resolution (as opposed to accuracy), like on automobile charging systems.

    old jim
  13. Oct 10, 2016 #12
    Thanks Jim for those insights.
    Nicely spotted re the DC amps.

    The truth is... I use my multi-meter a great deal.... but it has always been to establish voltage, connectivity, or battery state.
    This compressor adventure... by refusing to accept what was given; has added much more to the learning curve... but it is all still a bit foggy.

    Like the mega ohms test between the socket feeds and earth.
    It sounds like a swimming pool heater test..... I don't have it.

    I did feebly respond to your response, but I got no readings till I switched to -►|- when I got a reading between neutral and earth of .300... nothing from live.
    ... but I live in france... so the cabling here is entirely hit and miss.
    Power outages remind me of growing up in the sixties, in the UK.
    In fact, IT IS the sixties where I live..... it does have its advantages.

    re the unloader.... not sure what that is.
    The compressor stops when the diaphragm pressure switch hits 8 bar (adjustable).
    At that point, the air-feed hoses are unloaded by an N.O. solenoid.

    I take on board the observation on fluctuating power strokes, explaining the variable readings.

    All this is grist to the mill.... but it seems that the Sword of Damocles remains.
    Vis a vis the continual rising potential difference between the earth and the chassis.

    Let us not forget that I chose to cut the power at 100v AC.
    We do not know what would have been the final outcome.

    Also, let us not forget, that the 'cause' is still current (chortle :) )
    All that we have done is to ameliorate the problem, by feeding the growing voltage leak to earth.

    The pic of the 100v is real.... and I have another pic at 96.3v taken 4 seconds earlier.
    For the free gamblers.... the answer on a postcard would be.... what would have happened if I had left the motors running?

    Any takers?
  14. Oct 10, 2016 #13

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    That's curious.
    Was that on the compressor, or on your house wiring side ?

    DMM;s that i've encountered when set to diode function do this
    attempt to push one or two milliamps out the + lead and pull it back in the - lead, report the voltage it takes to do that.
    0.3 volts says it saw 300 or 150 ohms between neutral and earth. Strange it didnt find it on ohms scale.

    Do you have another meter?
    To the end of "knowing my test equipment" i like to use another meter to measure what current mine applies on diode scale, and how much voltage it generates to push that current through. Second meter when set to milliamps will report the current, when set to volts will measure volts your meter applies to an almost open circuit.
    You can make the same test to see what your meter does on ohms scale

    The more you know about your meter the less apt you are to get fooled by it.

    It seems freaky at first to hook two meters together and people will laugh at you, but it's really informative.

    old jim
  15. Oct 11, 2016 #14
    I was measuring the supply at the main compressor switch.
    It is the only multi-meter that I own.
  16. Oct 11, 2016 #15

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    if it was plugged in at the time you measured through the house wiring to the panel and back, round trip resistance plus ant stray voltage picked up en route..

    If you measured that on the compressor not plugged in then it's a mystery.

    I wonder if something curious is happening inside a capacitor. Are they metal or plastic cased ?

    if it's running well might just leave it alone..... I'd put in a GFCI receptacle though, for it'll tattletale on leakage current. I'm obsessive over figuring things out.
  17. Oct 11, 2016 #16
    They are plastic cased with a stud.
    I was wondering whether there could be leakage to the stud.

    It seems likely that the problem lies with a capacitor, simply because the chassis is gradually gaining charge.... as a capacitor does.
    The compressor is on plastic wheels, so there was no way for the charge to run to earth.

    The control box is plastic.
    This leaves the three metal blow off valves, actuated by N/O relays.
    I don't see these being the problem.
    The pressure switch contacts are encased in plastic... so no problem there.

    That leaves the capacitor studs and the motors.

    The key to this mystery is the 4v per second rising potential difference, while the motors are running.
    Something must be capable of creating this situation.

    The answer may pop into the head two weeks from now.
    However, it's unlikely to come from me because in this area of expertise, I don't have sufficient fundamental knowledge floating around my head.

    Let's leave it a while, and see if it comes to mind.
    At least the compressor is now running safely.

    Thanks for all the help so far :)
  18. Oct 11, 2016 #17

    jim hardy

    User Avatar
    Science Advisor
    Gold Member

    I = C X ΔV/Δt

    were that voltage to earth DC i'd say
    to charge two 33uf capacitors at 4 volts per second requires 66X 10-6 X 4 = 0.26 milliamps
    and you measured 1.4 milliamps

    but you reported it's AC

    so i admit im puzzled. There's a time to declare victory and leave for the time being.
    Mother nature will show us whatever lesson she has in mind when she feels we're ready for it.

    old jim
  19. Oct 15, 2016 #18
    Yes... it was definitely AC.
    The photo shows AC was selected.

    ... we'll solve this problem in due course :)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted