Cause of RCD not tripping with test switch

I don't think that an explanation is required. If the RCD has a proper supply and the outgoing circuit/busbar is detached, does the test button work or not? If not, the problem must lie in the RCD so it has failed.

Whether it amounts to C2 or C3 depends upon the circuit.

The cause may be corrosion of the test button contacts if the devices have not been tested on a regular basis. Or failure of the S type electronics internally. Or perhaps the internal test resistor has burnt out. As Chris says, if the R.C.D.s do not work on test button chuck 'em out. What are these 100mA R.C.D.s protecting? Will you need 30mA R.C.Ds?

Are you keeping the test button pressed long enough for the devices to trip off?

Do these devices date back to the 80s or 90s?

circuitbreakersonline.co.uk/wylex-wstm-1002-100ma-time-delayed-rcd-circuit-breaker-wstm1002-1644-p.asp

Z.

Hello Zoom, have you seen this?

IET Forums - RCD fails tests.. (theiet.org)

Z.

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?

If that's 0.84 MEG Ohms, i.e. 840,000Ω then I suspect so - I make it 0.286mA - as a quick sanity check 30mA at 240V is about 7.6kΩ

Usual cause of RCD not working is a N-PE fault of some kind - a portion of the test current can then find its way back via N (instead of PE) and so the RCD doesn't see the whole imbalance (or for test buttons connected Lin/Nout instead of to the c.p.c. then some of the test current can return via the c.p.c. rather than back through the RCD's N, so again the RCD doesn't see the whole imbalance). Switching off single pole MCBs doesn't disconnect N-PE faults of course. Just your IR results don't seem to jive with that theory.... DC is another possibility but would be unlikely I would hope.

    - Andy.

Need to have another think.

I got confused about whether you were still referring to the fault.

some things to try

1) RCD in the board - is there 240V between N and L on the inbound side. When it is on is there 240V between L and N outbound Does this vanish on both L and N when the RCD is in the off position.

2) Does the RCD test button really put a resistor between outbound and inbound L-N crossed - I know that it should, but does it - a general purpose DMM of the £20 kind with an ohms range will do. With the RCD out and on the table put one meter lead on N-out the other on L in ,  confirm the meter is on ohms and press the button. If still reads infinite  try the other way, L-out and N-in. Before that  if it is a new meter and you are not confident with the meter on ohms read a few other things first, filament lamps, heaters pencil lead etc just to get in the swing of the right readings and not measuring wet fingers by mistake.

10K would imbalance by 23mA, 1k by 230mA so expect something between 4 and 7K on a 30ma and 1 and 2K on a 100mA one.

(some test resistors fail open circuit, but the rest of the RCD stays OK.)

3) with RCD out, wire it between a plug and socket, as if an extension lead. Now test it...

If it really is failure of just the test button, and everything else is up to spec, but be really, really sure about that, perhaps your customer can always test them with something like one of these

plug in tester with 'trip my RCD' button until funds permit replacement.

Of course if it is not that then come back - it is interesting.

Mike.

Could the electronics in the time delayed R.C.D.s have been damaged by insulation resistance testing or a high Voltage lightning strike nearby? So that is why the electronic part does not work with the test button.

Z.

I have just had a look inside an R.C.D. that is of the same make and design as your Wylex S type unit. The cover is missing but I am sure that it is the same. It contains a small circuit board that contains several diodes, a couple of small capacitors, a resistor and a transistor. Two of the diode like devices are marked ITT 2PY 51. They are very small and I needed a magnifying glass to read them. The transistor is marked GE2N4992 and will be the most delicate and vulnerable component. The resistor is marked yellow, violet, brown gold.

I have installed these R.C.D.s myself and they are of very good quality. Mine boasts that it was made in Austria. In fact I have one installed here at home and it trips reliably on test button push.

A small printed label shows that the test button circuit is connected to the bottom (load) terminals 1 and N. Also is says that the incoming connections can be top OR bottom.

The diagram is very simplistic but seems to show that the test button circuit is independent of the fault sensing circuit.

The manufacturer is F&G.

Z.

Thanks al for your thoughts

I agree I was mixing up my decimal places, not thinking straight at all when I wrote the post.

I have 3 RCD's to work with (original and 2 replacements as I blamed the RCD's to start with)

The devices do date back to 80's/90's. The RCD in question is the main incomer switch to the CU, further 30mA RCD in CU to protect socket circuits, this means that there is domestic lighting and cooker protected by a 100mA time delayed RCD. Ideally should be replaced but that would involve a considerable cost which is going to be challenging for the customer and I would prefer to avoid unless really necessary.

I have been holding the test button for 45 seconds when installed and they trip more or less instantly with a bench test

I don't think it's the test button, 3 devices same issue and all 3 work on bench test connected to 13a socket (unless we are looking at cumulative errors) but I will test the resistance to double check.

On Andy's explanation, I think the explanation assumes test resistor is connected live out to CPC. In this case the resistor is connected live out to neutral in. All I can think of is that there is some pre existing leakage current neutral to earth which then balances out the current introduced on the live side. I would assume most RCD's work like this as most don't have an earth fly lead?

My KT65DL will not run an RCD test unless it sees reasonable voltages and correct polarity on the test leads

The downstream 30mA RCD works correctly after replacement. Presumably this means I have less than 30mA imbalance on the socket circuits and they are completely isolated from the 100mA RCD when the 30mA RCD is turned off. Therefore my issue is likely to be in the lighting or cooker circuits.

One thing that comes to mind is that to my knowledge insulation tests with a meter are DC, maybe I have much less impedance with AC due to capacitors or capacitance.

Looks like this is going to come down to lots of testing as nothing is obvious.

I think I am going to do the following 

   As a sanity check, test installed with line and neutral disconnected on the load side.

  Use clamp meter to test imbalances on supply cables with one circuit energised at a time

  As above with the neutrals disconnected, along with trying the test button and insulation tests again 

My device of the same make and model disconnects in under one second when I press the test button. Please see my post above. Were your replacement devices new or from fleabay? I have bought old faulty devices from fleabay in the past. They may test o.k. by tester but not work on test button.

Z.