RCD nuisance tripping.

Risteard

fenwick458 said:

Scrounger said:

I agree with FreeBear, I would change the RCD.

Those Mem A100HE RCD's sometimes develop a fault where they become sensitive to spikes on the incoming mains causing nuisance tripping.  Sometimes even your neighbours' electrical activity will then induce the tripping.

Scrounger

well I won't say it's not going to work, because there's a small chance it might. but in reality, it probably wont work. in all my years of faultfinding and electrical testing i've only found a faulty RCD once I've tested many RCD's that are twice as old as me and absolutely enormous in comparison to todays technology,  and guess what, they still work.  it's very common for a customer to phone up telling me that they need me to replace the trip switch that keeps on tripping and just assume that the trip switch is the problem (it's not that btw, its being caused by a fault)
an RCD is rated at 30ma, but by design most will trip at only 23ma when you ramp test them (this is because NO manufacturer wants to be liable when they DON'T trip at 30ma, so just keep erring on the side of caution and now it seems like they should just call them 23mA RCD's)
also most electronic device now leak a few mA of current to earth as standard. so lets say you have 17 circuits, and each one is leaking 1ma to earth, thats 17Ma already. you're 5ma away from the RC disconnecting already and there's not even a fault yet...
I've also lost count of the times I've been to faultfind a nuisance RCD tripping scenario, and after testing every circuit (with all loads disconnected) I've found no fault. but yet the RCD nuisance tripping persists later on. the customers always seem to think that they need a "better" electrician, one that can "actually find the fault", in reality the fault is somebody has put more that 5 circuits on one RCD!
thankfully I've also lost count of the amount of times I've been proved right too, and have changed many many 1990's upfront RCD consumer units to either dual RCD or RCBO's and *mysteriously* the problem has vanished
(as the earth leakage had been divided over multiple devices)

An RCD must not operate at half its rated residual operating current, but must operate at its full rated residual operating current within 300mS (or 200mS for older BS type devices). So 23mA is perfectly correct for a nominally 30mA device. (I find around 25mA is common on devices.) But 16mA would comply, as would 29mA or 30mA.

Jeepers_Creepers

Thanks everyone.

RCDs certainly do go faulty - when I moved in 16 years ago, I found the existing one wouldn't trip when tested, so I replaced it at that time. Ok, a different type of fault, but faulty none the less :-)    

When I look at the cost of replacing all the existing MCBs with RCBOs, I'm now thinking I'm best to see about having the whole board replaced, and hopefully the circuits rationalised a bit - 17 is crazy, as it's only a three bed house...

Thanks everyone.

Risteard

Never suggested that they can't fail, generally due to stiction from a lack of periodically pressing the test button - but that they don't tend to operate without a fault. Sometimes neutral/Earth faults in the installation can disable them, though it's more usual for this to trip them. DC leakage currents through a Type AC RCD can also blind it. It's why most of the world has prohibited Type AC RCDs for decades. 

Jeepers_Creepers

Thanks Ris. Tho' I didn't understand most of it :-)   

Jeepers_Creepers

I'm guessing sparkies will come recommending their preferred CUs and RCBOs?

Are there any makes which are considered the best balance betwixt quality and price? Are there any that should be avoided?!

I should be able to get the RCBO number down to a more sensible 12 without compromising anything significant, so I think this is the way to go. Would it be more sensible to go for a complete new CU or to have the existing MEM refitted with RCBOs?

Cheers.