Do MCB's go bad in time?

[Radian]

I thought MCB's were pretty dumb so wouldn't expect them to trip without due cause. RCCB's are a tiny bit more sophisticated so I would be less surprised to have one trip out for no good reason.

 

The thing is, last night at around 8PM the MCB for the kitchen end of the house tripped and on investigation the 2.5kW heater in the dishwasher had just come on for 2 minutes prior to the MCB tripping. Very suspicious and I was bemoaning the fact that we'd literally just bought a replacement fridge, installed today, and that all our appliances were bought at the same time (24 years ago) when we built the house - so they were all failing at the same time!

 

However, tonight at 8:27 there was another outage only this time it was the RCCB common to all ring finals that went. This gives no clue as to what caused it as resetting it made everything come back on and stay on. The only suspicious thing is that precisely 8:27PM was Amazon's idea of sunset and a routine was in operation that would turn on a bunch of smart bulbs. A huge coincidence, except it doesn't really fly as a theory because all those bulbs were already permanently connected to the AC. Only the LED's would be turned on and they're so far removed from the simple kind of switch that could suddenly connect Live (or Neutral) to earth as to be extremely unlikely as a cause. Plus the earth connection is very remote from the bulbs, not physically connected to anything.

 

So I'm struggling to diagnose the problem, and started wondering if it was possible that the MCB's and RCCB in my CU are yet another '24 yo. appliance' at or near EOL?

Edited August 18, 2022 by Radian

[markc]

Last nuisance tripping problem I had turned out to be electric UFH in bathroom

[dpmiller]

I'd be checking for loose terminal screws in the CU. L,N and E...

[Radian]

Last night I flipped all the MCBs off before resetting the RCCB in an attempt to see which circuit might have an earth leakage fault. I put most circuits back on one at a time but it didn't trip. However I didn't switch back on a 16A radial circuit in my workshop that has five double 13A outlets. When I switched that on today the RCCB tripped immediately. Simple? Not quite...

 

The five outlets had nothing plugged in, so you'd expect an insulation breakdown somewhere in the wiring. My insulation resistance tester at 500VDC reports 1.2 megaohms (both L->E & N->E) which would be suspicious for plain old wiring and sockets but these are all USB equipped outlets, so some leakage might be expected. 1.2m equates to 200uA which is a long way off the 30mA that's supposed to trip the RCCB. Each individual outlet could just be leaking 40uA so not an earth shattering amount.

 

I see two possibilities here, the 200uA just pushed it over the trip point by adding to a substantial existing leakage elsewhere and/or the RCCB is faulty and tripping on too little current.

 

2 hours ago, dpmiller said:

I'd be checking for loose terminal screws in the CU. L,N and E...

 

A very sensible check but if any of LNE became high resistance why would this create a residual current fault?

Edited August 19, 2022 by Radian

[ProDave]

What do you get for insulation test on other circuits?  I would not expect a USB socket to make much leakage to earth, it is a tiny little power supply connected between L and N.  Damp wiring could be an issue?

 

A lot of loads, e.g. a cooker or washing machine probably won't show much of a problem until they are actually turned on and the faulty element is connected to the circuit so not always easy to find.

[MikeSharp01]

Can you not measure and log the leakage current and see what's going on as you bring the breakers in. If it trips immediately does that not suggest something starting up oddly rather than staying odd.

[Carrerahill]

1 hour ago, Radian said:

Last night I flipped all the MCBs off before resetting the RCCB in an attempt to see which circuit might have an earth leakage fault. I put most circuits back on one at a time but it didn't trip. However I didn't switch back on a 16A radial circuit in my workshop that has five double 13A outlets. When I switched that on today the RCCB tripped immediately. Simple? Not quite...

 

The five outlets had nothing plugged in, so you'd expect an insulation breakdown somewhere in the wiring. My insulation resistance tester at 500VDC reports 1.2 megaohms (both L->E & N->E) which would be suspicious for plain old wiring and sockets but these are all USB equipped outlets, so some leakage might be expected. 1.2m equates to 200uA which is a long way off the 30mA that's supposed to trip the RCCB. Each individual outlet could just be leaking 40uA so not an earth shattering amount.

 

I see two possibilities here, the 200uA just pushed it over the trip point by adding to a substantial existing leakage elsewhere and/or the RCCB is faulty and tripping on too little current.

 

 

A very sensible check but if any of LNE became high resistance why would this create a residual current fault?

Do you have RCCB's? I would suggest you do not and that you have an RCD - if it was an RCCB then it would add another slight level of complexity to your fault, so for the avoidance of doubt, what do you have?

 

The Workshop circuit could simply be damp or something as simple as some creepy crawlies in amongst them, it could be external lighting or a lamp. I once had external lights with slugs on the junction box. The lights would reset and work fine for a day or two then the MCB or RCD would go, the RCD usually went when the lights were off, the MCB when the lights were on. Simple low impedance offered by the slug over N and PE was enough to make the RCD upset, similar event on a live circuit acted as a dead short and the exploding slug cleared itself, circuit reset until another slug had a shot!

 

This is where full RCBO boards are great!

 

 

[Radian]

1 hour ago, ProDave said:

What do you get for insulation test on other circuits?  I would not expect a USB socket to make much leakage to earth, it is a tiny little power supply connected between L and N.  Damp wiring could be an issue?

 

Yes, you're absolutely right. The SMPS in the USB sockets shouldn't really present any resistance breakdown even at 500V.

 

I think I may have found the problem though. One thing that seemed odd was the coincidence between an MCB trip one day and two apparently random RCCB trips subsequent to that and I think I now understand the link. When I unplugged all the equipment in the workshop I had in fact left a 6-way extension lead in one outlet. Just now I was removing and testing each USB equipped outlet in turn when I found the one hosting the empty 6-way extension was the leaky one... It wasn't the USB outlet, it was the empty 6-way block!

 

 

That was a surprise, and the reason soon became clear on opening it up...

 

 

It was an old 'surge protection' block. With a very helpful pair of MOV's between L->E and N->E and a third between L->N

MOV's degrade when they absorb voltage spikes so the apparent coincidence with the MCB tripping could be explained by them becoming leakier after taking a back EMF hit the day before.

 

There is one slight oddity in that when the RCCB tripped first thing this morning, the 6-way extension was isolated from Live via a smart socket that was switched off at the time. But N would still have been connected and the MOV between N->E and could still present a current imbalance that might trip the RCCB. 

 

53 minutes ago, MikeSharp01 said:

Can you not measure and log the leakage current and see what's going on as you bring the breakers in. If it trips immediately does that not suggest something starting up oddly rather than staying odd.

 

Hmm, that would be an interesting thing to log for sure - but how? Obviously the equivalent of a pair of antiphase CTs like in a regular RCCB would be what's needed but achieving this in practice sounds more than a little tricky.

 

47 minutes ago, Carrerahill said:

Do you have RCCB's? I would suggest you do not and that you have an RCD - if it was an RCCB then it would add another slight level of complexity to your fault, so for the avoidance of doubt, what do you have?

 

It's an old plastic fantastic single RCCB breaker unit:

 

 

[MikeSharp01]

1 hour ago, Radian said:

Hmm, that would be an interesting thing to log for sure - but how?

The RCCB must be doing it? Would just current measuring the earth return on the consumer unit not be a good proxy. Fluke have a good article on it https://www.fluke.com/en-gb/learn/blog/clamps/leakage-current-measurement-basics

 

[Radian]

13 minutes ago, MikeSharp01 said:

The RCCB must be doing it? Would just current measuring the earth return on the consumer unit not be a good proxy. Fluke have a good article on it https://www.fluke.com/en-gb/learn/blog/clamps/leakage-current-measurement-basics

 

 

Excellent;

Quote

These are very much like the clamp meters used for measuring load currents, but deliver significantly better performance when measuring currents below 5 mA. Most clamp meters simply won't register such low currents.

 

Which is why I discounted that approach but a highly sensitive CT would be just the ticket. Could be fun to plot!

[Onoff]

Just has some nuisance tripping here. Narrowed down to the upstairs ring main causing it. Old Crabtree C50 board with an upfront 30mA RCD covering all circuits so one out all out. Only thing I've done very recently is added another USB equipped double socket making two in total in that ring. 1st one was MK and this latest one was Click.

 

Did all the end to end tests on the ring legs and all as expected. Then tested at all sockets between L&N then L&E. Got the expected nom (r1+r2) /4. (There's 2 double sockets I can't get to as behind full wardrobes).

 

IR was interesting. At 500V I was getting circa 300 MOhm  between all conductors. Holding the button in and this crept up to infinity. Maybe some damp I burnt off?

 

For now I've left both USB sockets out and replaced with ordinary 2G socket outlets. Had no more issues.

 

Did make me revisit getting an SBS all rcbo board. Got as far as using the online board builder and it's in my basket! 

 

[Radian]

1 hour ago, Onoff said:

For now I've left both USB sockets out and replaced with ordinary 2G socket outlets. Had no more issues.

 

Quite possibly a MOV surge protector in the mains facing side of the USB PSU. These things are all over the place and get increasingly leaky with age (cue the humorous quips). But your IR test suggests otherwise. What it won't reveal though is the transient current through EMI filtering capacitors (AC to earth) so providing an additional leakage path. The more devices with switch-mode PSUs on the ring final, the more chances an impulse somewhere will trip an RCD. Maybe the additional USB socket was the straw that broke the Camel's back.

 

I meant to update this thread with a 'funny' that happened a few weeks ago. I had yet another nuisance trip. Went hunting again and eventually found low IR on the same circuit that caused trouble back in August. It was even more like groundhog day when I finally located the culprit: an identical 6-way mains distro with built-in surge protection. I totally forgot I had two the same, both lurking at the back of a workbench.

[Onoff]

38 minutes ago, Radian said:

But your IR test suggests otherwise.

 

Forgot to say I swapped out both USB sockets before testing in case they were sensitive to the Megger. 

 

Tbh I don't like them. Had a previous MK one cook a power bank I was charging. The replacement MK, well the USB side simply died. That's why I went for Click this time.

 

Having all circuits on one RCD really is at the mercy of accumulative leakages across the circuits. Really must change mine. 

[Radian]

1 hour ago, Onoff said:

Forgot to say I swapped out both USB sockets before testing in case they were sensitive to the Megger. 

 

 

Just join L & N and test between that connection and earth as well as the USB pins. I'd want to know if the hf transformer insulation had broken down. If it can't take 500V it has no business having peoples phones plugged in to it!

 

I agree they're a bit of a worry being on 24/7 and potentially on a 32A breaker. A PCB interfacing directly with a 7680kW capable source seems like asking for trouble.