Two RCD's on same circuit - Test failed

[vwes59]

Hi, quick question before I start ripping a customers Consumer Board apart.

Called out to a simple job of installing electric supply to a summer house. Nothing major being used in the garage or planned to be in the summer house and is only 10m from the house so was happy to extend the radial circuit in the garage which was on a 16A MCB in a garage consumer unit protected by an RCD incomer with a supply going back into the main house connected to another 16A MCB which was protected by another RCD in the main consumer unit. The main house was rewired recently and the paperwork shows the main RCD tested fine - no paperwork for the electrics in the garage and the owner tells me it wasn't part of the rewire.

Plugged in my Kewtech tester into a socket in the garage. 1/2x tested ok, but as soon as it tried to test at 1x both RCD's tripped with a >50ms on the display and FAILED underneath. I checked the circuit (R1/R2 and IR) and both checked out fine.

So, my question is, could having two RCD's on the same circuit affected my testing ? Is it possible the RCD in the garage is faulty ? I've tested circuits in the past where two RCD's are present, and the nearest one to where I'm testing is the one that always trips and I've never had a situation where its failed or where both trip at the same time.

 

[James]

2 rcd’s in series is a poor design if they are both the same trip characteristics.
because of other leakage in the system, more often than not the rcd closest to the supply will trip first but not always.
I would start by testing the rcd’s at source withe all outgoing breakers switched off.
it is possible that some device is generating some dc leakage that is slowing the rcd trip time by partially saturating the sensing coils.

 

[telectrix]

it's pot luck as to which RCD trips. either one or both. if the supply to the outbuilding needs RCD protection then remove the RCD in the outbuilding. if not, feed from a non-RCD way in the house (if available).

 

[Pretty Mouth]

As above, it's pot luck as to which RCD (or both) will trip when under test. The upstream one may be inclined to trip quicker, as it's more likely to see more leakage current from the other circuits that it protects.

You can test the downstream RCD independently of the upstream one, by testing at the device, and putting the earth probe onto the supply neutral rather than the earth. This will leak current to the neutral when under test, so the upstream device won't see any imbalance.

 

[vwes59]

As above, it's pot luck as to which RCD (or both) will trip when under test. The upstream one may be inclined to trip quicker, as it's more likely to see more leakage current from the other circuits that it protects.

You can test the downstream RCD independently of the upstream one, by testing at the device, and putting the earth probe onto the supply neutral rather than the earth. This will leak current to the neutral when under test, so the upstream device won't see any imbalance.

Never thought of that. Thanks for the suggestion, I'll give it a try. There's quite a bit of work at this particular house on top of the summer house supply, so the owner is quite happy to replace the garage consumer unit if it becomes necessary, or at the very least switch out the RCD incomer to a bog standard main switch.

 

[pc1966]

Plugged in my Kewtech tester into a socket in the garage. 1/2x tested ok, but as soon as it tried to test at 1x both RCD's tripped with a >50ms on the display and FAILED underneath. I checked the circuit (R1/R2 and IR) and both checked out fine.

Having two 'instant' RCD in series is, as already said, poor design due to the lack of selectivity.

But that should not cause them to trip slowly!

As above, test in a methodical manner. First check the up-stream RCD on its own and verify it works as expected and never trust presious measuments - they may have been worng, or something might have failed / been surge-damaged in the mean time.

Then test the down-stream one as @Pretty Mouth explained.

If both test OK individually but somehow fail in cascade, which I would be surprised to see, then look at replacing the summer house with a simple switch.

 

[EricMark]

Two 30 mA RCD's in series is common, nearly every caravan or narrow boat is wired that way, but normally we work on 1/3ths or times 3 so each RCD is 1/3 of the size to last one.

There is some debate about testing, one school of thought is the RCD should be tested with no circuits connected, however the main cause of failure in early days was strain caused by the cables, and dropping cables to test means it could pass then be disabled due to strain.

The same applies to DC causing desensitising of the RCD, clearing if everything switched off, one would be unaware of any desensitising, although it seems to be under some debate as to exactly why or if DC desensitises type AC RCD's.

It has also been raised that it is the time when tested as a type AC which matters, if the 40 mS is achieved when tested as type AC but exceeded when tested as type A it should still be a pass, which raises the question why test as a type A, F or B then?

Not so long ago I watch a video where representatives from manufacturers were complaining about returned RCD's which were found not to be faulty, due to testing methods, and I must say I did not agree with a lot of what they said, however they seemed to be able to support it with reference to BS 7671 so left a little confused.

I have had a RCD fail when supplied from a ELCB-v it seems the resistor in the earth from using the ELCB-v was stopping the RCD tripping, and I looked at the ELCB-v and must admit not a clue how to test it. I have nothing old enough to tell me.

I fitted a new earth rod and used the ELCB-v purely as an isolator and fitted another RCD. However it did point out how a faulty earth can stop the RCD tripping, this raises the idea of using the neutral for the test in question, as whole idea is to test it works in a real fault situation in time, to use the neutral to test the second RCD does allow one to test second RCD, but at the 5x setting one should trip within 40 mS.

I had a problem with a 10 mA RCD socket type A with the test button, pressing it would trip the 10 mA, 30 mA, 100 mA and 300 mA RCD's all would trip. What we must remember is BS 7671 says over 200 ohm can be unstable, so in real terms up to 200 ohm can be accepted with a TT supply, it should be easy to select a resistor which will not trip the 100 mA, but there is also back ground leakage, however it seems MK did not do this. They simply did not expect to have a series of RCD's.

To me it seems strange that two 30 mA RCD's both failed to trip in 40 mS. Other than with that ELCB-v I have not had a problem with earth rods and tripping time, it may have been because not a simple resistor but a coil of wire?

I no longer have a tester to test earth rods, tend to compare with DNO earth, but I have only once tested MCB's on tripping current, it was no a caravan site, and it seemed too much power being used, all 5 or 6 amp MCB's and tested with 3 kW fan heater, if did not trip, changed, but in the main if the manufacturer says it's a B6 we accept it is a B6 without testing, so why test a RCD?

I would say we test as history has shown they often fail, but how can we both test which nothing connected and test in a real situation? We can't. So test as it would be in real situation, as type AC with all connected.