RCD Tripping - long (sorry!)

[puffernutter]

I have a workshop a distance away from the house (approximately 60m). It is fed from an MCB in a consumer unit (with an RCD as a main breaker) through two x 2.5mm FTE. At the workshop it enters through another RCD then into a consumer unit with three MCB, two at 20A feeding two workshops and one at 6A feeding all lights. This consumer unit also feeds the underfloor heating and the immersion heater.

There is a second consumer unit that does the rest of the house - lighting circuits (protected by a single RCD) and upstairs and downstairs sockets on RCBOs and cooker, shower etc., also covered by an RCD. The incoming mains (pole route) is also protected through and RCD.

The reason for the RCBOs is that for a while, from what we believed to be an external influence, the "whole house" RCD would trip and we would lose everything. Now we have RCBOs protecting some circuits, so the losses are contained.

That's the background!

I am currently getting "nuisance" trips of the workshop RCD, although sometimes it also trips out unrelated RCBOs on the upstairs and downstairs sockets (a different consumer unit to the workshop supply in the house). At other times, just the RCD at the workshop trips.

I am trying to work out where to look. I have turned the workshop off totally at night (through the workshop RCD) and the house hasn't tripped. Not conclusive, but a start. I have removed circuits one by one from their MCBs and neutral bars (stopping the neural/earth fault as well) and this has shown that in two cases tried so far, disconnecting these circuits, leaving the others connected, doesn't stop the trips. I have two more circuits to work through (thank goodness it is cold as the freezer is on one of them!)

I haven't meggered the cables between the house and workshop yet (next weekend job when it is light) but my question is, could a fault in the house affect the "downstream" RCD in the workshop and how can the workshop RCD tripping affect "upstream" RCDs and RCBOs?

In other words, am I looking in the right place at the workshop end, or could I have a house fault reflected into the workshop RCD and I'm just wasting my time in the workshop?

Confused of Wiltshire!

[Andrew Mawson]

60m).

breaker)

fault as

disconnecting

trips. I

Peter,

We went through a long period of RCD trips - always in the small hours of the night. I isolated and meggered everything but found nothing. Eventally I changed the RCD (it was a Wylex 100A/30mA of 1996 vintage) and it totally cured the problem. The removed RCD would latch on ok, but the slightest tap with a screwdriver handle would trip it. It had gone over sensitive and I suspect triggered on perturbances on the (overhead) supply when the voltage was high.

(Incidentally I dismantled it and to my suprise it contained small electrolytic capacitors - so they must have a finite life)

AWEM

[John Rumm]

I think this is the crux of the question...

As a general rule, if you cascade normal RCDs (i.e. those without a time delay) then you will get no, or very little, discrimination between them.

If you have a 30mA trip RCD feeding another of the same rating, then a fault on the second RCD will be seen by both, and either one, the other, or both will trip.

If you have a 30mA RCD fed from say a 100mA trip RCD, and you have a fault on the far end, then again, both will see the fault and either or both can trip. The only exception to this is when the fault is a small magnitude leakage - say 40mA - and that is not enough to trip the upstream unit.

Generally to get cascading to work correctly, you need RCDs with time delays in the upstream positions. Then when a fault occurs downstream that potentially could trip both, the upstream one will wait for a bit to give the downstream one a chance to trip and clear the fault.

Needless to say, a cascade of more than two devices starts getting difficult unless they are high threshold devices (>= 100mA) there just for the protection of the wiring rather than protection from shock.

Depends a bit on the detail. You have not said which RCDs have what trip thresholds, and which are time delayed ones (usually with a S in the part number suffix). So its possible that faults in the workshop can cause trips in the house if the house RCDs are upstream.

Its less likely that faults in the house are causing trips in the workshop. However things like RCD sensitisation (caused by excessive normal leakage) can make things very unpredictable, and increase sensitivity to mains transients either carried on your supply or caused within your installation. So if up and downstream RCDs are close to their tripping point, a transient on an upstream only protected circuit will possibly affect both.

Also RCDs can go faulty. Normally this results in a failure to trip fast enough or at all when they should, but the reverse can also happen.

Not sure if you have read:

it may offer some additional points to check.

[ARWadsworth]

Photos? I would love a look.

[Andrew Mawson]

Sorry Adam but they've been binned. It was a small bit of pcb maybe

1cm square or less, with two diodes and an electrolytic capacitor - I assumed it was forming DC for the trip circuit

AWEM