Electricians: measuring an single appliance's earth leakage current?

[Howardlong]

Indeed.  When I Googled the manufacturer of the breakers shown in the posted image above, the breakers I found only had a single physical screw terminal on the line and load sides.

Yes, as I mentioned, they're just the circuit MCBs, they're not the RCDs. There is only one RCD, and that is only on the socket rings and cooker (and the Joker, a 16A circuit that, as far as I can tell, only supplies the masthead amp on the TV antenna... I'd left it turned off for weeks, and only finally realised what it was for when I turned the TV a long time afterwards and there was no signal).






The RCD is either a GE 604241 or 604243, can't tell the last digit as it's hidden by the switch, but they're the only 80A RCDs starting with 60424#.

The way I've seen sparkies sniff out leaking circuits is to open up the CU, remove each L & N connection from each of the MCBs, and test the insulation resistance of each. Now I could do that with the high voltage tester designed for the job, but frankly I think this would probably be the time to get a proper sparks in. Unfortunately the last sparks I had in when this originally happened was a come-and-go guy, he wasn't interested in actually trying to resolve the problem as it wasn't tripping when he turned up.

[Andy Chee]

The way I've seen sparkies sniff out leaking circuits is to open up the CU, remove each L & N connection from each of the MCBs, and test the insulation resistance of each.

To apply this strategy to your scenario, turn off all MCBs, leave RCD on, measure L+N of your 80A RCD.  Should be zero imbalance. 

Whilst continuing to measure L+N of your RCD, turn on each MCB one by one until leakage is observed.

[Howardlong]

The way I've seen sparkies sniff out leaking circuits is to open up the CU, remove each L & N connection from each of the MCBs, and test the insulation resistance of each.

To apply this strategy to your scenario, turn off all MCBs, leave RCD on, measure L+N of your 80A RCD.  Should be zero imbalance. 

Whilst continuing to measure L+N of your RCD, turn on each MCB one by one until leakage is observed.

Yes, this depends on me having a clamp meter that gives anywhere near consistent and reliable results: the two I have with milliamp resolution don't (UT210E & ACM81). Mike convinced me enough to purchase the WinApex 268B, so we'll see what happens there. It's not tripping at the moment since I removed the surge protector.

One further bit of information that I didn't think was relevant, but in retrospection it might be: on about 1 in 5 of the RCD trips, one of the 32A MCBs protected by the main 80A RCD tripped too, as well as the RCD itself. Can I infer anything from that?

As I understand it, the problem of using the strategy of "turn on each MCB one by one until leakage is observed" while measuring the L+N is that, as only L is switched on the MCB, any N-E fault won't be isolated.

Thanks for your input.

[adam4521]

You’re right, don’t discount N-E faults, they can happen all over the place. Check your cooker, immersion heater, toaster. At my parents house years ago, the electrician fitted a new consumer unit and discovered a latent N-E fault in a ceiling rose, been there for decades probably, really hard to find.

I think your original idea of using your isolation transformer is a useful avenue for ‘individual appliance’ test. It will stop your house RCD from tripping while you are trying to measure. You’ll need to ensure/know that the secondary winding is earthed and connected to the DUT CPC to recreate the protective conductor return path, and to keep your equipment enclosures safe. Take all the measurements you want on that local system. Note that unless your transformer is one of the medical grade ones that has an interwinding screen, there will be a small but not negligible amount of leakage current through the transformer.

[mikeselectricstuff]

One further bit of information that I didn't think was relevant, but in retrospection it might be: on about 1 in 5 of the RCD trips, one of the 32A MCBs protected by the main 80A RCD tripped too, as well as the RCD itself. Can I infer anything from that?

It takes quite a lot to trip a 32A MCB so maybe indicative of something more serious - an intermittent short to earth - things likeoven or immersion heater elements are common culprits. Current sufficient to trip the 32A could well have blown out what was causing the short, at least temporarily.
The problem going away by removing a surge protector could simply be that the leakage was already close to the limit and the surge protector pushes it over the edge. 

[Monkeh]

One further bit of information that I didn't think was relevant, but in retrospection it might be: on about 1 in 5 of the RCD trips, one of the 32A MCBs protected by the main 80A RCD tripped too, as well as the RCD itself. Can I infer anything from that?

Was it always the same MCB?

[themadhippy]

if its the 32A device next to the rcd id consider swapping the circuits around  to see if the fault moves,its rare ,but occasionally the magnetic field in one device can interfere with its neighbor .