Strange RCD tripping problem

[Mjbarnard]

I have a very strange repeated RCD tripping scenario, which appears to suggest that two unrelated circuits are connected. Any problem solving thoughts would be appreciated.

I have an electrician who has done several jobs for us (new lighting circuits, appliance installation, outside lights, changing sockets etc). I am very pleased with him, and he has visited around 5-10 times over the last year to try and fix this - but admits he doesnt understand it. He is going to bring in colleagues who specialise in troubleshooting. So we both just thought it might help to put the problem out there for any ideas. I do not touch anything electrical myself.

New build detached house, finished 2012. Dual split load consumer unit with RCD on both (picture). Everything worked fine until last summer when Air COnditioning installed by very experienced professional firm. They installed 5 split units in 5 rooms, and put two outside Mitsubishi units each with new dedicated supply with MCB in CU in two spare slots.

Ever since then we have had intermittent RCD tripping, which my electrician has investigated extensively. Coincidentally we have also replaced all appliances since then with new Miele ones and new Liebherr high end fridge, so unlikely they would be involved.

More recently the RCD repeatedly trips unless the MCB for one of the air con circuits is switched off . Both the electrician and air con company have investigated this - but cannot find a fault with the air con circuit. However my electrician when checking all the possibilities discovered a very strange anomaly. First there is a 6 gang double pole grid switch in the kitchen for isolating various appliances (picture). He found that one of them (for plate warmer) had incorrectly had a simple (light) switch installed - so live and neutral were connected when on. He replaced that and checked all the others.

He then made a discovery which is hard to understand. The RCD immediately and repeatedly trips when the air con MCB is switched on. However it does not trip when air con MCB is on but dishwasher grid switch is off. Turning the dishwasher grid switch on immediately trips the RCD. The air con works fine and no RCD trip when dishwasher isolated, but dishwasher works fine and no RCD trip when air con isolated! We have repeatedly tested and found the same results.

My electrician is seeking advice but does not understand how a new air con supply can somehow be linked to the dishwasher which comes off the ground floor ring mains which has no connection to the air con circuits. Not only has he inspected and tested the CU several times but so have two very senior technicians from the air con company.

This all happens consistently including when all other appliances switched off, so I do not believe the supply is insufficient.

Hope this makes sense. I am not sure if posting pictures is allowed from new members, so I will try to do so in the next post.

 

[Wilko]

Yes if I isolate the air con external units or turn the air con mcb's off OR leave the air con on and isolate the dishwasher (irrespective of whether it is plugged in or not).

Edited this - as just read the DW bit. Please confirm - the DW is unplugged and just operating the DW benchtop isolation switch affects whether the RCD trips?

 

[Ian Mills]

Hi new to forum I work in France so work with aircon units, I think problem is capacitors in AC units because when turned on they store small amounts of electric they make the rcd think there is a fault, the dishwasher will also have capacitors and when you turn dishwasher on it will have same effect, you may also have some small earth faults and the capacitors just push it over edge, would use combined rcbs /mcb, to give less chance of tripping, can not remember what they are called in UK regards Ian

 

[Octopus]

Yes if I isolate the air con external units or turn the air con mcb's off OR leave the air con on and isolate the dishwasher (irrespective of whether it is plugged in or not).

Can you post a better picture of your fuseboard?

 

[DPG]

Welcome to the forum Ian.
In the UK we call them RCBOs.

 

[marconi]

I agree with all the advice on thorough inspection and electrical wiring tests - including earth leakage current measurements. It's a good idea too to swap to an RCBO populated CU for the convenience and least disruption to the family.

I reckon moisture, water or damp air could be troublesome inside the external a/c units although our 4 split a/c mitsubishi units on top of a flat roof have been happily energised permanently and fed by an mcb on one side of a dual rcd board and never caused a trip.

What puzzles me about this problem is the involvement of the dishwasher circuit post the double pole grid switch - the dishwasher itself has been ruled out by unplugging it. I have a suspicion it has a relatively high impedance N-E somewhere between the switch and socket including in the socket or its back box.

I also observe that in the photo of the grid switches the neons for the dishwasher and cooker hood look noticebably dimmer than the other two - maybe a feature of the camera angle but seems odd to me. I'd be itching to open that grid switch box and CU and to check the correctness and tightness of all the connections.

The other aspect which might be a factor is that the tripping occurs when an MCB switch is made to power up - so some form of transient effect - which makes me ponder about the current/voltage disturbances in L, N and cpc caused by any input filters the a/c systems have - they are largely electronically controlled these days so will need protecting. I'd be keen to observe the L-N, L-E and N-E voltages and cpc currents during these switching events with DW switch on and off.

I note that the oven and hob are supplied from the same side of the CU - the former often being a villain for element earth leakage and N-E when the element controls are off.

Some electrician is going to have a fascinating time investigating and putting right!

 

[Ian Mills]

Welcome to the forum Ian.
In the UK we call them RCBOs.

Thanks

 

[marconi]

ABB Classification of RCDs upon detectable wave form - Residual Current Devices RCDs (Modular DIN Rail Products) - http://www.abb.co.uk/product/ap/seitp329/4bd0ec0528a84420c1256f5500557772.aspx

http://www04.abb.com/global/seitp/seitp202.nsf/0/740f1caaeb62137fc1257fe900545279/$file/Guia+Técnico+ABB+-+Diferenciais+Tipo+B_EN.pdf

I cannot get this problem out of my mind. At the moment the advice is to switch to an RCBO board (or to fit 2 new RCBOs for the two a/c systems). The question for me is what type of RCBO? I think there is a risk that using type A or AC RCBOs will not stop the problem of nuisance trips because of the nature of the inverter drive technology and its effect on the load current waveform. I cannot find any recommendation by Mitsubishi on what type of RCD device to use with their inverter a/c systems. During the course of this on-line research I have read some folk (and manufacturers) recommend not using an RCD to feed a/c equipment because of the problem of nuisance trips - sometimes when the a/c equipment has been powered down for some period and then powered up.

Looking at the references above it seems to me that the correct RCD type to use is Type B which does all that Types A and AC do. The Type B can detect those 'true' faults which have more complicated current waveforms generated by earth leakage from inverter electronics and supply filters and yet avoid nuisance tripping in the presence of 'normal' low level earth leakage currents as the supply filters do there job 'dumping' harmonics, noise and leakage to earth including power up/down transients.

Does anyone have any practical experience of selecting the correct RCD type for a/c systems using inverter technology?

On page 6/12 the second reference you will find:

With these properties, Type B RCDs turn out to be RCDs with high immunity to nuisance tripping caused by grid surges, electronic loads and EMC filters. Therefore, Type B circuit breakers are the ideal solution for all “difficult” loads, not only from the point of view of protection, but also in terms of service continuity

 

[Ian Mills]

The advantage of 2 or more RCBOs is that it means if there are other small earth faults on system they do not add to what ever the a/c units are doing and the a/c units might not take a trip out on there own, and if they do you can go back to the a/c company and say it is your system sort it out

 

[marconi]

Picking the right protection - https://www.electricalreview.co.uk/features/9258-picking-the-right-protection

In this reference I have just read:

Now let’s turn to RCBOs. In principle, as these are functionally an MCB and an RCCB in the same device, it would be possible to produce versions with a huge range of combinations of MCB and RCCB characteristics. In practice, manufacturers limit themselves to the most commonly used combinations, such as a Type B or C MCB characteristic with a Type A or AC RCCB characteristic. Where other combinations are needed, it often means that an RCBO can’t be used and that the necessary protection must be provided by using a separate MCB and RCCB. Some manufacturers, however, offer field-fittable residual current units that can be used to convert an MCB into an RCBO, and this makes possible a wider range of combinations.

 

[Ian Mills]

Here in France we would use Hpi type but only if a problem as expensive but normally will not trip if not a earth fault, do not know if you have same but would think so

 

[marconi]

Hpi type:

Modular devices Hpi type RCCBs - http://www.export.legrand.com/EN/modular-devices-hpi-type-rccbs_63.html

Hpi seem similar to UK type B RCCB - not sure yet if they are the same.

 

[Mjbarnard]

#61 - correct. We had DW unplugged all day yesterday and could replicate tripping.
Coincidentally we had the DW replaced a few months ago and the original one had been hard wired in, so as part of the installation the electrician changed the fixed wiring to a single socket. So this is also new and presumably tested.

 

[Mjbarnard]

iPad error

 

[Mjbarnard]

iPad error

 

[Mjbarnard]

#63 will try to post better image CU when I finish rounds.