Thread: Earth leakage testing and RCD's

With all domestic circuits having to be on RCD's these days, how does one go about troubleshooting a "non-fault" earth leakage issue on a machine in the workshop.
Can't run the machine as it trips the RCD, can't trace the source of the earth leakage unless you can run the machine.
Seems to be a chicken and egg situation, so I must be missing something !!

Steve

Steve how basic or complex is the wiring?
I would start at the plug and work my way towards the motor, testing for continuity between earth and active wiring.
Maybe first test the motor windings to make sure there is no continuity to earth...

If a continuity test fails to reveal the cause you may need a megger. It applies 500VDC at low current to test for deterioration of the insulation and leakage through contamination.

Thanks guys.
Shed, I was trying to stay away from being machine specific as I've got a couple of machines that will trip RCD's and will need looking into.
My old TOS mill which I'd put at about a 3/10 on the wiring complexity scale, and a CNC lathe of about the same vintage which is probably about 12/10 in complexity for me!!
I've got away with the mill so far due to the particular way I'm supplying it, but with the new wiring rules from a couple of years ago now requiring all domestic sub-circuits to be RCD protected its going to be an issue going forward. I need to arrange a new circuit for the CNC lathe, and as far as I can find out it has to be on an RCD (which I know will it will trip).
Also, getting too specific invites attention from the PC police.
For the record, I've already spoken to my sparky about options for a new circuit etc. Sounds like if it was a commercial installation there are exemption categories that may allow it to be done without RCD, but not for domestic.
Obviously the simplest way forward is the redneck DIY approach - but apart from it not being my style, my (ex sparky) old man is still alive and quite capable of kicking my butt if he found out I was messing in switchboards.
So - back to the issue....

Insulation breakdown etc on straight electrical installations is one thing, but what about noise filters etc that are designed to leak.
My understanding of general earth leakage tracing on those sort of systems is that you do it powered up, and identify the particular sub-circuits that are leaking.
Again, if my understanding is correct, the only way you can really do that if the main supply is RCD protected is to use the likes of an isolating transformer between the supply and the machine so that you can run it without tripping the RCD.
The RCD effectively doesn't see any imbalance in the secondary side of the transformer.

Steve

Originally Posted by apple8

If a continuity test fails to reveal the cause you may need a megger. It applies 500VDC at low current to test for deterioration of the insulation and leakage through contamination.

I should add that any electronics should be disconnected before meggering.

Normal process for fault finding is isolation of separate components until you find the cause.
Does not matter if you are doing that for a whole electrical circuit where you disconnect individual appliances by unplugging them one at a time or for appliance repair where you disconnect the individual component to find the problem.

It helps to note when the problem occurs, eg when main motor starts, but if it is as soon as it is plugged in it can be more difficult.
Disconnect the big ticket items, motors and heating elements is a good starting point.
Helps to keep in mind that the offending component does not need to be powered at the time, only that it is in circuit with earth and neutral at the time, current drawn on the active elsewhere can then find return path via a earth neutral fault.

Thanks Droog.
I get the concept of "disconnect things until the fault goes away", but in practice its not quite that simple/practical (well - not in my brain at least).

As a starting point, assuming good earth wiring etc, any leakage current *should be via the earth wire. Not always the case I know but have to start with some assumptions until you prove otherwise.
My original thinking was that there are a whole bunch of earth cables coming back to the main earth bus. Everything perfect there shouldn't be any current in them, but if there is an earth leakage issue then there will likely be some. Definitely if you're already seeing current on the main earth return wire.
So throw a clamp meter on each one and find the one/s that do have current and work back from there.
But if you can't run the machine due to the RCD tripping - its impossible to take that approach.
Also, you can't get any idea of the magnitude of the leakage. The 30mA to trip an RCD is (by design) sweet FA.
If you know you're dealing with 2A of leakage its a much different scenario than if you've got 35mA.

TBH its frustrating as hell. I feel like I'm trying to find a leak in a pneumatic system, but every time I try to pressurise and check for leaks I get tapped on the shoulder by the RCD and told - "sorry - can't give you pressure 'cos its leaking. Now run away like a good little boy and fix it before you come back and ask me again"
But Mr RCD - please tell me HOW BIG a leak am I looking for?
"More than you are allowed - now go away"
Grrrrrrrr.

Rant aside, FWIW the RCD trip on the CNC lathe is instant when you press the RUN button to boot the controller.
Old CRT screen, ancient DC power supplies, there's a few likely candidates.
I've confirmed that its definitely an RCD trip rather than overcurrent, as the controller boots up OK when powered from my old RPC which has an isolated secondary.

All good fun!

Steve

Steve can you bang in an earth rod beside the shed and run your machine earth to that for your testing?
Essentially bypassing the RCD but you still have the earth....?? or would the RCD still sense the earth??

Originally Posted by shedhappens

Steve can you bang in an earth rod beside the shed and run your machine earth to that for your testing?
Essentially bypassing the RCD but you still have the earth....?? or would the RCD still sense the earth??

Unfortunately not Shed.
The RCD senses imbalance between the currents in the active and neutral wires. In a perfect circuit they will be the same, but if there's leakage to earth then the neutral current will be less, and that imbalance trips the RCD part of the breaker.

As I said earlier, the only way I can see to "fool" the RCD on the supply side into thinking all is well is to electrically isolate the machine from the supply using an isolation transformer. That way the supply circuit is only connected to the primary side of the transformer, the active and neutral have the same current and all is RCD happy.
Since the secondary side of the transformer is only magnetically coupled its not part of the circuit the RCD is monitoring.

Steve

Originally Posted by OxxAndBert

Unfortunately not Shed.

bummer.....

if you could hire or get a loan of a 3 ph generator to do your fault finding you could save a few bucks from buying an isolation transformer, preferably an old gen with fencing wire across the ceramic fuses and no RCD

First thing I would look at, is there anything 240V on the machine and does it have a 4 or 5 pin plug?
If yes and 4 pin plug, there is a good chance the earth is being used as a neutral.

That's how my jig grinder and a chain hoist arrived.

Thanks Stu - good suggestion.
Unlikely to be the case with the CNC lathe as the supply to the machine is 200v 3ph - from a 415-200 3ph step down autotransformer.

General plan at the moment is to power it off the old RPC to try and find the earth leakage issue. Doesn’t help that our mains supply voltage is currently down around 216v, so I’m not getting the full 200v to the machine anyway

Steve

Originally Posted by OxxAndBert

Doesn’t help that our mains supply voltage is currently down around 216v...

Wow, that's right on the bottom of the tolerance window. Does it often get that low? Might be worth asking your supplier to investigate.

Yes, I know.
First time in a while that I’ve checked it TBH, and I only checked as I noticed my RPC output was low.
I had the opposite about 3 years ago when I had solar installed and the inverter was frequently shutting down due to over voltage.
Contacted Powercor, they investigated and changed the transformer tapping to bring it into spec.

I really need to set up something to log the voltage for reference.

Steve