I have a strange fault where an RCD operates as expected when tested on the bench but when installed the test switch does not cause it to trip.
I have asked this on another forum but not really managed to get an explanation of what could cause it and why.
I have 3 RCD's giving the same fault, they trip satisfactorily when tested with an MFT when installed but test switch does not cause a trip with the downstream MCB's on or off.
My thoughts so far.
Unlikely to be the RCD as swapping has no impact.
MFT test creates fault current to earth, test switch creates fault current from live to neutral prior to sensor coil, how much current flows??
Insulation tests on installation gives about 1.04m ohm live conductors to earth at 250V on two circuits, 0.84 for installation at 240v this gives a leakage current of 285mA. Why doesn't this trigger the RCD, or do I have the decimal point in the wrong place?
The above is leakage current at 240V, but problem exists with CB's of, therefore presumably very limited voltage across circuit and current flowing.
Others have said its likely to be a neural earth fault. I can see the logic that the line is disconnected by the CB's and therefore it must be related to the neutral plus I could see that if there was a fault current in the neutral side it would balance the test switch current, preventing a trip. But what would drive the current as the potential difference between Neutral and earth is close to zero, and why would testing with a meter give a different result.
I thought about the RCD being bypassed in someway but I couldn't see any unexpected electrical connections.
I have considered DC leakage current, but would you get DC leakage with everything switched off? Maybe batteries feeding back through the charging circuit? There are led lights, one TV and quite a few gadgets plugged in, but very limited compared to most, no car chargers or computers.
The house was rewired about 20 years ago with black / red PVC cable that looks reasonable. There is quite a bit of dodgy DIY, with fixed lights on flex, plugged in in to sockets, lots of extension leads and adaptors; unplugging these will be a good starting point.
One strange thing I did find was that the house has a TT earth but when the main earth cable is disconnected at both ends I get an earth impedance of 0.3 ohms from the cable. Presumably a hidden connection to steel frame of the semi detached Cornish house, maybe with next door on TNC-S. Water pipes are bonded but isolated on entry to the house.
As some background this is on an old Wylex WSTM 100/2, 100mA type S rcd.
I realise I am going to have to disconnect circuits to break the problem down but would very much appreciate any thoughts on possible causes and theory behind the cause. I have a horrible feeling it's gong to come down to cumulative effect of multiple issues which are going to take considerable time and money to sort out. I won't be able gaurantee making 1 or 2 fixes will solve the problem and the customer has a genuine lack of money.
Any guidance on possible faults to look for and why they would cause the symptoms would be much appreciated, I am not a fan of just trying things without understanding the background science.
On a slightly different tack, I am considering failure of test button test as a C2 that needs resolving, am I over coding it and would a C3 be more appropriate? I am certainly not that comfortable with it as the owner has no way of knowing if the RCD is working and presumably there is one or more faults in the house.
