Test equipment to identify cause of RCD trips: Update

[Caecilius]

An update to my search for the elusive cause of the intermittant RCD trips that I posted about in March.

I bought a cheap megger off eBay for just under GBP 30. It's a Smart Sensor Insulation Tester model AR907.

the price it seems to work fine. I've checked it against some resistors and it gives the expected readings, and a multimeter shows it's supplying the expected voltage.

I've tested the plugged appliances but not found anything with low resistance. The lowest I found was 95M at 500V for the microwave. I have found that anti-surge sockets show about 0.3M at 500V, but infinity at

250V. Apparently this is a known problem and doesn't indicate a fault, but it does make me wonder whether anti-surge plugs could cause a trip if a mains transient caused them to conduct to earth.

I found one fault, which I already suspected. Some armoured cable used for garden lights measured 0.04M L-E at 500V after some rain, so obviously damaged. I had observed that switching this on usually caused a trip if the ground was damp. I've now disconnected this, as I don't really use it, but I don't think that it's the main cause of the nuisance trips as it's normally isolated from the supplying ring by a DP switch. But maybe having the earth connected to "real earth" in the back garden might cause a problem?

I've had a look at changing the MCBs for RCBOs, but am trying to avoid it for now as the inside of the CU is quite crowded and difficult to get to as it's behind the rails for the up-and-over garage door. Here's some pics:

?action=view&current=P1000072.jpgI'm going to check out the fixed appliances (boiler Etc) and the fixed wiring next to see if there's any problems there.

[The Natural Philosopher]

of course they could. Thats almost what they are designed to do.

I cant quite see how..

Check away, but my solution was to put 100mA RCD in et mains.

I ought to have 30mA RCBO's when I get a round tuit as well.

Note this has blown every time there has been a genuine fault, but has cured the nuisance trips.

[tony sayer]

In article , Caecilius scribeth thus

And the neutral is isolated from earth as well as live in the system?...

[Ron Lowe]

> I'm going to check out the fixed appliances (boiler Etc) and the fixed

I may have replied to you before, not sure. I also had nuisance tripping of RCDs.

The only method of isolating the issue was 'disconnect stuff, wait and see'. Totall useless, unless the trips were occouring within 5 mins or so.

Megger is not that usefull either, nothing conclusive.

This is a DIY group, so I built my own Residual Current Meter. Based on an old Merlin Gerin RCD, I used only the Current Transformer, with a 10k load resistor. Calibrate the thing with a 12vAC transformer, and induce test leakages with resistors. Measure the actual mA test leakage, and plot against actual mV output. Voila, calibrated earth leakage meter.

Use this to measure total leakage, then per-circuit leakage. This will instantly shed light onto where the problem lies.

Otherwise, you're groping around blind, and guessing. Start making actual measurements.

[Caecilius]

I'll check that when I test the fixed wiring. I'm planning to join L&N and then test between the joined L&N and E to check for leakage, like I did for the appliances. That should show up any L-E or N-E insulation problems.

[Caecilius]

I tried that too, but like you I found it useless. Sometimes my house can go months without tripping, and sometimes it will trip twice in a week.

So far I've found it useful to prove or disprove things, which is useful. But it's got a couple of limitations that I've discovered since using it:

a) It only uses DC, so won't show any capacitative leakage; and b) Many devices don't supply power to the internal workings unless they are powered up, so a megger on these only shows leakage from the permanently live bits.

For simple things like the cooker I've manually checked the elements, but for more complex stuff like the fridge that's not an option.

Now that's a good idea. That will actually measure the L-N inbalance on a working circuit just like the RCD will.