UPS Neutral/Earth Short

[Huge]

Much investigation of the intermittent RCD trips this afternoon (I'm shattered after spending all afternoon unplugging things and plugging them in again.

One thing we discovered, which may well be a red herring, was that my APC UPS, when switched off, with no power being fed to it, exhibits a Neutral/Earth near short circuit on its input. (0.4 Mohm)

Is this normal? It cannot be permanent, else the RCD would trip the moment the UPS input was powered on. Does the UPS perhaps strap the earth and neutral together when there is no input power to provide a false earth to the hardware being protected? Or is it just faulty?

(Looks like the trips are, as expected, N/E leakage on the external feed to the Klargester and workshop. [Forsees large bill for replacing the workshop supply. Sigh.])

[Dave Liquorice]

A 30 mA RCD won't trip with 0.4 M ohm (400 k ohm) neutral earth. The earth loop has to be around 7 k ohm or lower (do the maths...)

You haven't mixed up your M's and m's have you?

[Peter Parry]

Have you managed to completely isolate (L,N and E) the feed to both the workshop and Klargster? Is the fault still there? Have you tried isolating it at the source end and is the problem still present.

Some years back I helped a friend with a tripping RCD (eventually traced to a nail through a floorboard neatly but intermittently shorting the neutral and earth).

However, one of the things which could cause the RCD to trip was taking a long extension lead out into the garden. As soon as it was put down on the ground the RCD tripped with nothing connected to the cable at the far end. The cable reel was fine with no leakage whatsoever. I can only assume there was some transformer coupling to earth through the extension cable which was just enough to kick the RCD already stressed by the N/E leakage over the edge. There was certainly no physical leakage from the cable or cable reel.

The range of weird trips in this particular case included switching the lounge light on (sometimes), putting the toaster into one kitchen socket, but no others and using the socket in the main bedroom for the Hoover but not the one in the spare room- except sometimes it was the other way around. It was this last symptom which led to finding the fault, we realised that which upstairs socket tripped the RCD depended upon whether someone was in the toilet - their weight on the floorboards moved the nail fractionally and shorted the bits together.

You probably didn't want to hear all that :-).

[Caecilius]

400K ohm is not a near short circuit. Even between live and earth that wouldn't trip an RCD on its own as it would only pass 0.6mA at 240V.

But did you check that with a multimeter? I had nucience trips a few years ago, and found there was sometimes a big difference between the N-E resistance shown by a multimeter (which uses a few volts) and a megger (which uses a few hundred volts).

When I had a similar problem, I found a few problems with a megger and sorted them out. But I only cured the problem for good when I changed out all the RCDs for RCBOs.

[Bob Minchin]

I recently posted on a related topic where the UPS I purchased to run my gas boiler did NOT have any connection between the power output and earth which meant the flame detector failed as it expects the neutral to be at or around earth potential. The manufacturer finally admitted that this was the case but would not give me the circuit so I could devise the connection of a relay to effect an earth neutral short when the mains failed. In the end I fitted a resistor of a suitable value in situ all the time what kept the flame detector happy and yet did not trip the RCD.

Maybe APC have included the relay that I would have liked to add to mine? I would suggest that you could fit a double pole switch on the power input to fix your problem maybe?

Bob

[Chris Whelan]

[...]

Insulation resistance has to be tested with a source that is twice the working voltage or more, hence why those intended for domestic or other single phase supplies are 500v.

Chris

[Huge]

[26 lines snipped]

It all sounds very familiar. :o(

[Andy Burns]

Is that mega or milli? I've got a couple of spare APC units kicking around I could compare ...

[Huge]

Yes. But we can't leave it off too long, else there'll be raw sewage coming out of the Klargester, which Isn't Allowed.

Who knows. Such is the joy of intermittent faults.

[John Rumm]

For starters could you separate the workshop/klargester feed off on to its own RCD separate from the house?

(I have a dedicated CU here with its own RCD for all the outbuilding stuff. At least that way if there is a fault I do not import the consequences back into the house)

[tabbypurr]

or his m&ms.

NT

[Andy Burns]

With an ordinary multimeter on 200 MOhm range, E->N reads as infinity.

With a hand-cranked analogue "megger" it wobbles about somewhere between infinity and 100 MOhms, the wobbling is more likely from the action of cranking it while trying to hold two probes in contact with the plug and avoiding zapping myself with 500V.

See, I knew that schuko to C19 cable I didn't throw away the other week would come in handy!

[Dave W]

As the mains is AC, there will be a difference between leakage measured on a DC meter, and the greater actual leakage due to electrical capacitance between cable and ground in the garden.

[Andy Burns]

But perhaps having so many C14->C13 cables which strung nose-to-tail they reach over 95m, is a couple too many?

[Huge]

Yes, thank you, I am well aware of this.

[Mike Tomlinson]

En el artículo , Andy Burns escribió:

I've seen it, believe me. One reason not to leave boxes of C13-C14 cables where the lusers can get at them.