Nuisance RCD Tripping - Diagnosis Advice PLEASE

[2nd Life]

That is quite a "special" CU to be looking at. and is it just camera perspective, or is that the biggest choc block I have ever seen?

Special is one word. And yes that is not a trick of the camera. That is the biggest connective block I have ever seen with two 10 mil bonding cable stuffed into it

 

[timhoward]

I want to do IR testing anyway as if I'm going to get someone to come and change the board.

I'd start with a 30ma ramp test plugged into a sockets circuit.
Then I'd turn everything off (at the service isolator not pictured), perform safe isolation check, and do a global IR test (Live and Neutral together tested to CPC bar) of each side of the board. This can be done with no disassembly if you have the right test leads, or minimal disassembly if you don't.

 

[2nd Life]

I'd start with a 30ma ramp test plugged into a sockets circuit.
Then I'd turn everything off (at the service isolator not pictured), perform safe isolation check, and do a global IR test (Live and Neutral together tested to CPC bar) of each side of the board. This can be done with no disassembly if you have the right test leads, or minimal disassembly if you don't.

Thanks again Tim. I'll do that. Incidentally here's a pic of the whole thing. We have a 30 amp Aga running off its own CU with a 32 amp MCB and isolator running off a Henley block, and I actually only tried to turn it back on for winter this morning and what do you know tripped the sockets RCD immediately. The mini CU has it's earth linked into the MET in the main box so doesn't surprise me

 

[timhoward]

A rather pertinent detail - have you noticed an earth rod anywhere?
At the moment I don't see how that lot is earthed and how fault protection for the Aga is going to work.

 

[pc1966]

The more we look, the more concerns we have

@timhoward has spotted the very important question of means of earthing, if it is a TT system or some other connection to the supply earth is in use but not visible (and possibly not acceptable, such as service pipe bonding).

There are a lot of circuits, and the Aga on a separate CU interacting with the main CU is troubling. Is this a big house, i.e. how many of those circuits are in use and necessary?

It might be a bit beyond you just now, but it really looks like a case for biting the bullet and replacing the main CU with a new all RCBO board and bringing the Aga in to that, as for the rest of the circuits. But before considering that in any detail, you need to establish the means of earthing and to do your basic checks on polarity, IR, Zs/Ra, and identifying the circuits in-use and seeing what is needing done ASAP and what can be part of a planned out change.

 

[pc1966]

Do USB sockets add to leakage as that is the only thing that has changed in that time.

Probably, but they are unlikely to be a big change as they ought to be sub-mA levels.

I want to do IR testing anyway as if I'm going to get someone to come and change the board. I want to know from my own benefit that the wiring is sound. What I did think was that if it all checks out on the board change goes ahead.

It might be worth seeing if anyone on this forum is nearby and willing to help you with the CU change. As you are learning to be a spark it would be good if you could be involved in that process but not all folks are happy to get involved that way.

Then at least having individual RCBOs will be able to share the leakage and it may just go away on its own?

If there is no fault beyond poor design / feature creep over time then that would (a) solve the problem of accumulated leakage, and (b) help pinpoint any future faults and reduce the impact they have until a fix can be done. But if you can safely do IR testing, etc, you can help pinpoint what is happening now.

 

[2nd Life]

I'd start with a 30ma ramp test plugged into a sockets circuit.
Then I'd turn everything off (at the service isolator not pictured), perform safe isolation check, and do a global IR test (Live and Neutral together tested to CPC bar) of each side of the board. This can be done with no disassembly if you have the right test leads, or minimal disassembly if you don't.

So bizarrely the ramp test at the socket did not trip the RCD, just get a reading of greater than 33mA where the tester I assume stopped. I tried the speed test too and same result. No trip. For an RCD that seems on the verve of tripping constantly not sure what to make of that. Could be the tester at fault?

 

[2nd Life]

A rather pertinent detail - have you noticed an earth rod anywhere?
At the moment I don't see how that lot is earthed and how fault protection for the Aga is going to work.

Yes. It's on a TT system (also terrible) I'll stick a pick in of the "electrode" but I have ZE it and I'm getting 165 ohms so well within the max miraculously

 

[timhoward]

Yes. It's on a TT system (also terrible) I'll stick a pick in of the "electrode" but I have ZE it and I'm getting 165 ohms so well within the max miraculously

Issue one is the Aga CU. That has a normal main switch (not RCD) and an MCB. So a live-earth fault would not trip the MCB as less that 2 amps of fault current would flow. It would stay on waiting for someone to touch it and that someone would become the path to earth.

So bizarrely the ramp test at the socket did not trip the RCD, just get a reading of greater than 33mA where the tester I assume stopped. I tried the speed test too and same result. No trip. For an RCD that seems on the verve of tripping constantly not sure what to make of that. Could be the tester at fault?

Does the test button on the RCD itself work?
If not, then this all becomes even more serious as things won't turn off under fault conditions.

Which make/model MFT is it? It could be a faulty RCD, faulty tester, or the tester being blinded by DC earth leakage.

 

[2nd Life]

Issue one is the Aga CU. That has a normal main switch (not RCD) and an MCB. So a live-earth fault would not trip the MCB as less that 2 amps of fault current would flow. It would stay on waiting for someone to touch it and that someone would become the path to earth.


Does the test button on the RCD itself work?
If not, then this all becomes even more serious as things won't turn off under fault conditions.

Which make/model MFT is it? It could be a faulty RCD, faulty tester, or the tester being blinded by DC earth leakage.

Test button works on the RCD in question it looks to be a newer one, but not on the other one after investigation.

It's a Di Log 9110 and had a calibration sticker that ran out in July. Got it for good money so planned to spend the savings on calibration but not as yet.

 

[2nd Life]

Issue one is the Aga CU. That has a normal main switch (not RCD) and an MCB. So a live-earth fault would not trip the MCB as less that 2 amps of fault current would flow. It would stay on waiting for someone to touch it and that someone would become the path to earth.


Does the test button on the RCD itself work?
If not, then this all becomes even more serious as things won't turn off under fault conditions.

Which make/model MFT is it? It could be a faulty RCD, faulty tester, or the tester being blinded by DC earth leakage.

On the Aga question the plan is to integrate it back into the CU once replaced.

 

[2nd Life]

Issue one is the Aga CU. That has a normal main switch (not RCD) and an MCB. So a live-earth fault would not trip the MCB as less that 2 amps of fault current would flow. It would stay on waiting for someone to touch it and that someone would become the path to earth.


Does the test button on the RCD itself work?
If not, then this all becomes even more serious as things won't turn off under fault conditions.

Which make/model MFT is it? It could be a faulty RCD, faulty tester, or the tester being blinded by DC earth leakage.

Could isolate the Aga for now and swap out the main switch for an RCD to cover us until the CU swap?

 

[2nd Life]

Issue one is the Aga CU. That has a normal main switch (not RCD) and an MCB. So a live-earth fault would not trip the MCB as less that 2 amps of fault current would flow. It would stay on waiting for someone to touch it and that someone would become the path to earth.


Does the test button on the RCD itself work?
If not, then this all becomes even more serious as things won't turn off under fault conditions.

Which make/model MFT is it? It could be a faulty RCD, faulty tester, or the tester being blinded by DC earth leakage.

Or swap the Aga MCB for an RCBO?

 

[DPG]

Or swap the Aga MCB for an RCBO?

Might as well just connect it to the main CU when you swap it. Unless it's going to be a good while before you can get this done?

 

[2nd Life]

Might as well just connect it to the main CU when you swap it. Unless it's going to be a good while before you can get this done?

Hi DPG, no we'll be getting on with it. Particularly given all the problems that the current setup can and is causing. Need to get rid Asap. But it could be weeks. And the RCD issue is daily so really want to check that nut if I can

 

[timhoward]

So bizarrely the ramp test at the socket did not trip the RCD, just get a reading of greater than 33mA where the tester I assume stopped.

Test button works on the RCD in question it looks to be a newer one, but not on the other one after investigation.

I find myself wondering if these two things are linked....
It might be worth double checking that the socket you used to test loses power when you turn the RH RCD off. There are a few cables that look bigger than typical lighting wires the left hand end.

It's possible to test the ramp test function of an MFT by putting a multimeter in series between tester earth wire and earth, set to the mA function, but it helps if the multimeter has a peak hold function.

 

[pc1966]

So bizarrely the ramp test at the socket did not trip the RCD, just get a reading of greater than 33mA where the tester I assume stopped. I tried the speed test too and same result. No trip. For an RCD that seems on the verve of tripping constantly not sure what to make of that. Could be the tester at fault?

It could be any of a number of problems:

• The MFT might be faulty
• The RCD might be faulty
• There might be some leakage that is either desensitising the RCD ("blinding" by DC) or is in opposite direction to the L-E test used, cancelling some of the test current.

So you can look at trying:

• For the 1st point can you check the MFT on someone else's socket? Or get a plug-in RCD adaptor and see if it tests/trips normally.
• Once you know the MFT is working OK then test the RCD at the CU with all MCBs off, and see if that is any better.
• Get a basic clamp meter that allows you to check for leakage, and see what is happening on the tails with all off, and then one MCB at a time on.

 

[2nd Life]

I find myself wondering if these two things are linked....
It might be worth double checking that the socket you used to test loses power when you turn the RH RCD off. There are a few cables that look bigger than typical lighting wires the left hand end.

It's possible to test the ramp test function of an MFT by putting a multimeter in series between tester earth wire and earth, set to the mA function, but it helps if the multimeter has a peak hold function.

Hey Tim. No I mapped every circuit in the house going round with a socket tester etc it's Definitely on the tripping RCD. I'll try the multi meter trick that would help to know the MFT is working. Cheers again

 

[2nd Life]

It could be any of a number of problems:

• The MFT might be faulty
• The RCD might be faulty
• There might be some leakage that is either desensitising the RCD ("blinding" by DC) or is in opposite direction to the L-E test used, cancelling some of the test current.

So you can look at trying:

• For the 1st point can you check the MFT on someone else's socket? Or get a plug-in RCD adaptor and see if it tests/trips normally.
• Once you know the MFT is working OK then test the RCD at the CU with all MCBs off, and see if that is any better.
• Get a basic clamp meter that allows you to check for leakage, and see what is happening on the tails with all off, and then one MCB at a time on.

Thanks for that PC! I think the clamp meter is a must I'll be picking one up asap. At very least this is a fantastic fault finding theory session to build up my knowledge. Thanks to all for the help it's much appreciated

 

[DPG]

Thanks for that PC! I think the clamp meter is a must I'll be picking one up asap. At very least this is a fantastic fault finding theory session to build up my knowledge. Thanks to all for the help it's much appreciated

Make sure it can read accurately at low mA figures.