Smart meter tripping RCD anymore than anecdotal evidence?

I find it hard to believe the boiler would trip the sockets unless there is some really weird cross connection that I am sure you have already eliminated.

In terms of how many of those leads on one RCBO, as many as you fancy pretty much - so long as they are in good condition - and less than one where the MOVs are starting to fail after a few oversize surge events.

The data sheet says 2 x MOV + GDT  which tells us the sort of surge suppression -  two of the semiconductor type and a gas discharge tube (like an enclosed spark gap in a special atmosphere.) I'd expect a star config of Mov L-centre, N-Centre and the GDT from E to centre.  There may even be a fusible resistor in series with the live to prevent a fire when the MOV is end of life..

370 joules sounds like a lot for a single MOV disk, so there may be a few in parallel, or one large one of about inch diameter in each position (example)

When new the capacitance is a few hundred pF, so the cable will be more than that - we can discount capacitive leakage through the MOV being an issue - but only if it  is in good order. The MOVs should not leak on a 250V test between L-E  or L-N . (or indeed on anything up to the crest of 253V RMS, about 330V DC, but that is not an easy test voltage to generate.)

Mike.

Whatever it is was taking out the 300 mA upfront RCD that trips at 210 mA, so I replaced it with a main switch having over sheathed the SPD phase cable with heat shrink.

Is that referring to a Type G RCD or a Type G SPD, I am assuming it is a RCD but it seems it could be either or both.

I think a G-type RCD is one with a very small deliberate delay - around 10ms - so it still provides additional shock protection (overall it's still within the 300ms at 30mA and 40ms at 150mA limits) but is less sensitive to very short duration residual currents/surges.

https://www.eaton.com/content/dam/eaton/products/electrical-circuit-protection/circuit-breakers/xeffect-rccb/eaton-rcd-application-guide-br019003en-en-us.pdf mentions them.

If the (presumably non-delayed) 300mA device is tripping and but the downstream 30mA ones aren't then that to me points to a large magnitude (>>150mA) but short duration (<40ms) pulse of residual current, where both devices are in their 'may or may not trip' regions.

Modern boilers do normally ignore the gas by a generated spark - galvanically that should be all contained within the boiler but I suppose some strange fault might let the surge escape. If the boiler is close there might possibly be some radio interference (the original radio transmitters were a spark gap with an aerial/earth connection just to help the effect propagate longer distances) but again the original design should have prevented such a problem.

Other thought is the classic DP switch in front of a high resistance N-PE fault - with the switch fully open or fully closed there's little leakage, but if the N opens slightly before L or closed slightly later than L, then there can be a momentary surge of current to PE.

   - Andy,

"Modern boilers do normally ignore the gas by a generated spark".   I like that. 

And in someone else's comment in another recent thread there's "hypnotic" instead of "hypothetical".  

Does this new site have spelling "correction" that I haven't yet been victim of, or at least haven't noticed?  These seem unlikely words to have come by accident, and I don't remember such gems from the earlier site. Or perhaps it's mobile devices doing clever AI-based inference of what's meant. 

And now I remember also a "route 3" versus "root 3" that might have had the same reason.

Inside the modern boiler, as well as the spark ignition is also a micro-controller and a fair amount of delicate sensor electronics to keep the fan speed mass air flow and so forth all on the efficiency knife edge - designs where the ignition was bosky enough to fire an RCD at 40 paces by EMP would almost certainly at the least re-set its own processor if not put itself unequivocally beyond use. I think we can rule that out.

Something like a shared neutral that only comes into circuit when the pump is on and the 3 way valve is in the hot water position might just about sneak in under the net as it were, but I think even that would have already been found during other tests.

Beware of the autoKrikketer and double check your spellings before committing to witchcraft as a career.

Some errors phrase are funnier then others

M.

A cute clemon!

The plan was to have a set up like in this video, but this is the post-Brexit UK in the middle of the Coronavirus pandemic at Christmas and whilst we are told to use Type A RCDs the 100 mA S-Type time delayed RCDs are something that I have never actually ever seen or used and are not readily available, I am on a promise that I will get some, but I do wonder if they will actually arrive as they are on back order and I suspect they are not even in the UK.

So I went for plan B, a 300 mA RCD in the tails, which was not the best of plans, however the IET Onsite Guide tells us we can omit the upfront RCD if it is an RCBO consumer and is well engineered, but there is the additional issue of the SPD connections, I have ended up over sleeving the SPD phase connection to its MCB with heat shrink and replacing the upfront RCD with a main switch.

Youtube video

It seems Youtube videos cannot be embedded into posts anymore.