Explain types of RCD socket, pleeeez.

WIAWSNB

I fitted a metal-clad double socket with RCD near my garage door many moons ago, to provide power to outdoor thingies, like mowers and stuff. It doesn't have switches.

Recently fitted another properly-outdoor socket in a box, also with RCD, but also with switches, specifically to supply a Granny charger.

Anyhoo, both have test buttons. When I test the external one, it trips as it should, and I reset it. When I test the inside one, it trips - but so does the garage RCD switch in the house that supplies the garage - quite a pita.

So, why?

It's annoying, because the internal socket doesn't have switches, so I need to physically remove anything that's plugged in to remove power to it. I cannot leave my second Granny plugged in, and use the 'test' button to cut power.

I'm replacing it with a switched type, but curious why one test trips just the socket, and the other trips both the socket and garage supply?

Ta.

vacheron

RCD's monitor the current in the phase (live) and neutral conductors to look for an imbalance suggesting an earth fault (basically someone being electrocuted from live to earth). Just 30 milliamps will trip an RCD, but this is the maximum allowed, so it usually trips at around 26-27mA.

When you press the test button, the RCD will simulate an earth fault of 30mA to see if the unit trips, but this imbalance will also be detected by the upstream RCD feeding the garage.

In your case it looks like the the inside unit is pulling slightly more current during the test, or taking slightly longer to trip when you press the test button than the outside RCD, and so this is causing the upstream (garage feed) RCD to trip too.

As just the garage is tripping, I assume this is fed from a consumer unit using RCBO's rather than a conventional Circuit Breaker with shared single (or double) RCD?

grumpy_codger

So, how many RCDs do you have? Do you have two connected in sequence - 'internal' and 'external'? If so, why?

Section62

https://forums.moneysavingexpert.com/discussion/comment/81972678#Comment_81972678

When you press the test button, the RCD will simulate an earth fault of 30mA to see if the unit trips, but this imbalance will also be detected by the upstream RCD feeding the garage.

Normally the simulated imbalance would only exist internally to the RCD - created either by a short across the toroid (e.g. from the 'live' output side to the neutral input side) or by a separate 'test' coil on the toroid which simulates a fault current. The traditional RCD only has 'live' and neutral connections, so there is no 'earth' present to simulate a real earth leakage, and it is only if there was a real 'earth' leakage (simulated or otherwise) that this would be visible to RCD devices upstream.

It is possible the test button on WIAWSNB's metal-clad socket does create a real 'earth' leakage, because an 'earth' should be present in the fitting and it would be possible to do so, but I've always understood that wouldn't be an acceptable design for the very reason that the test button would cause upstream devices to trip as well.

vacheron

https://forums.moneysavingexpert.com/discussion/comment/81972700#Comment_81972700

Good point (and an assumption on my part) which makes perfect sense as the earth isn't generally connected to the RCD. A quick google corrected that assumption.. which as a 30 year electronic engineer I probably should have done first! 🤦‍♂️ 🙃

Now my engineer brain has had a sleep, how about this for a theory?… 😋

For clarity I will use/assume "RCBO" for the garage feed and "RCD" for the unswitched socket with the test button.

As it sounds like the OP is using the RCD test button as a contactor to disconnect the local load (which is not really the intended practice) my next theory is that any capacitive elements in the energised load(s) which are present between live and earth (such as input filters) are discharging when the mains power is removed.
The high dI/dt when the mains is removed outside of the mains zero crossing points would cause these charged capacitors to appear as a low impedance current source.

Normally the dual pole isolator of the RCD being would prevent any return current in either live or neutral when tripped, but if the test button was still being held in after the RCD had tripped, a reverse current path from the load capacitors will exist via the held in test button creating a current limiting shunt to the RCBO neutral in the consumer unit.
The common earth between the CU and the load will remain via the earth pin and complete the circuit.

As this flow will be in the neutral line only (the live being already disconnected by the RCD) this could cause a current imbalance in the Consumer Unit garage RCD/RCBO toroid which could trip it.

If the OP were to disconnect his load devices first and then press the test button, this could prove or disprove this theory. My guess is that most people won't generally press the test button with operational loads connected, and the existence of a further upstream 30ma RCD.

vacheron

https://forums.moneysavingexpert.com/discussion/comment/81972681#Comment_81972681

The op mentions that the socket in question was fitted "many moons ago", so my guess is that the consumer unit has been upgraded since the original sockets were fitted.

That is certainly the case at my parents house who still have a couple of RCD sockets installed which I fitted specifically as these were used for connecting outdoor garden tools / pond pumps etc. when they still had a "fuse board" which would have happily barbecued them with "amps for hours" had they encountered an earth fault when mowing a damp lawn or something. 😧

WIAWSNB

Thanks everyone. I'll clarify the full setup later.

I know the test button isn't a substitute for an on/off switch, but it says to test regularly, and since there's an absence of switches on one, I thought 'why not'...

The garage has its own separate CU with main RCD switch in the house, as the PV system was causing nuisance trips to the house CU when it was just fed from one of its MCBs. That sorted the issue.

WIAWSNB

The garage was built around 15 years ago, and wired by, er, me. On the inside, near the main door, I fitted a Powerbreaker H22 - a double, metal-clad socket with RCD, but (stupidly) no switches.

The garage had/has a small CU, but a non-RCD main switch. In the house, the garage was initially supplied from a 32A MCB in the house CU which has a main switch with RCD. So, yes, from the beginning this garage socket had two RCDs - itself and the hoosie one.

Following nuisance trips, the PV guys wired in a separate one-way CU with MCB and RCD for the garage, set alongside the main house CU, and that cured it. Still, that means the socket has two RCDs.

I recently fitted an external double socket inside a metal wall-mounted box for use with a Granny. This socket, too, has RCD, and switches.

I now have a weird setup with the Granny in the box delivering 10A charges when the sun is high, but a second Granny in the garage socket which has been tuned down to 6 or 8A to match duller days. (The idea is that car charges are almost completely fed from the PV, regardless of how bright or dull the day…)

The irritation is that the garage socket doesn't have switches, so I need to either physically unplug the charger when not in use - a pita and also wear on the contacts, or use the 'test' button to turn it off (well, it does say to test 'very time' before use…), but this trips the garage RCD in the house - a pita - or to leave the Granny plugged in, which I understand is 'ok', but just not something I would ever do.

Or, to swap the unswitched RCD socket for a standard switched double - and I've just ordered one from SF.

So, the issue will be resolved, but I was curious about why one RCD test button would also trip the house RCD, but not t'other. Seemingly there are a number of possibilities, from differing sensitivities, to there being different 'types' of RCD.

The unswitched garage one is a Powerbreaker H22, and the exterior one a BG WP22ARCD which is a Type 'A' RCD. I can 'test' this one without it tripping the house.

Not a biggie, and now resolved, but just curious :-)

Thanks all.

Section62

When you've removed the Powerbreaker H22, check to see if the markings show it to be a DP one.

Is the garage on the main house (PME?) earthing system, or does it have an independent (e.g. TT) system?

How/where is the PV connected? If possible, what happens if you use the inside RCD socket test button with the PV completely isolated?

Given the previous issues with tripping I'd fear there is a fault somewhere which the fitting of a garage specific CU has masked rather than solving. Did you post about this before, or am I thinking of someone else with a similar issue?

TBH rather than getting a replacement RCD socket I'd probably have gone with a new CU in the garage and have the sockets on a RCD/RCBO protected circuit so there were no need for the RCD in the socket(s) themselves - that way you could just have standard switched socket(s) inside the garage. With the PV added to the mix this is probably a job for an electrician rather than DIY. I also expect the new rush for plug-in PV to start throwing up similar issues more widely.

WIAWSNB

https://forums.moneysavingexpert.com/discussion/comment/81972970#Comment_81972970

Thanks.

I'll be replacing the Powerbreaker in the next few days, so will check.

Some memory you have! I believe I did post about nuisance tripping waaay back (10 years?!) when it was a regular issue, but the separate garage CU in the house sorted that.

I think all will be well once I replace the PB22 with a switched socket - only the house-located garage CU will then have an RCD covering it.

Not really keen on replacing the garage-installed CU if not essential. There's no real issue other than the test button also tripping the garage CU, but that will soon not be the case.

Cheers. 🙂

Section62

https://forums.moneysavingexpert.com/discussion/comment/81973073#Comment_81973073

No, it must be someone else as the one I'm thinking of is more recent. You may have commented about your problem on that thread? IIRC the recent one was solved when they got a new inverter.

vacheron's theory about capacitive elements across 'live'/'earth' makes sense, but only if the RCD is faulty, or SP, or very poorly designed. The 'test' button 'short' would be across the toroid, not across the contacts of the breaking mechanism. If there is a trip then (assuming DP) the breaking of both 'live' and neutral connections from upsteam should be virtually instant and simultaneous - so the only remaining connection (if the garage is on the same earthing system as the house) would be via the CPC. Thus no circuit and no imbalanced current to be detected by the upstream device.

That's in theory though - in practice there could be some momentary arcing, or one pole might break fractionally before the other, or there could be a fault including contacts welded together, or incorrect wiring, or (which would be surprising) the RCD only operates on one pole.

If the RCD in the house is 'sensitive' (some are more than others) then the unwanted tripping could be caused by that kind of thing.

The subject interests me because my parent's place was fitted with a whole-house RCD (upstream of the consumer unit) when RCDs were still fairly novel in domestic homes. Having lots of outdoor electricity use and outbuildings we were regularly getting nuisance trips, despite the whole-house RCD being a 100mA type. Nothing we or the electrician did ever stopped the problem completely.

However, (when it was still allowed) I eventually upgraded the wiring to the outbuildings so each one had its own CU and was disconnected from the house PME earthing system. By coincidence, probably, the whole-house tripping more or less stopped.

Using vacheron's theory, the disconnection of the earthing systems would have meant that the only electrical connections were the 'live' and neutral to the outbuilding RCD, plus a path through the ground from the TT system back to the supplier's PME grounding. That ground path is probably far too high resistance for any theoretical capacitive discharge-type effect to generate a trip current.

I'm not suggesting you DIY alter the garage earthing system, nor temporarily disconnect the CPC to test vacheron's theory, because both are potentially very dangerous. However, if you still have a problem with tripping it might be worth asking an electrician to quote you for converting the garage to TT if they think your circumstances are suitable for doing so.