MCBs tripping randomly?

[Lewis Curle]

Hi all,

So I’ve been having a strange fault in my house which has just appeared over the past few days. About a month ago I moved my pc set up into the sun house to make space for my second kid.

Inside the sun house/garage I have a DB that is fed of a 40A type B breaker fed by a 6mm core cable.( it’s not a SWA it’s a wierd cable with loads of little cables as a neutral) but I have insulation resistance tested that feed cable and it’s all good. On the garage DB I have a 32A ring main and a 6a lighting circuit. I have had my pc setup in there for a few weeks now and it’s been totally fine up until a few days ago when the 40A MCB is tripping of at random times. It can go off when I’m there and gaming or off through the day when I’m at work.

The circuits in the garage have all tested out totally fine and the feed cable going to the garage DB has tested out fine as well. But for some reason the 40A MCB in my board keeps tripping off. I have changed the MCB as of yesterday and it has went off again today whilst I was at work. There is nothing on in the sun house when the MCB trips. 40A is huge MCB for a house and I can’t get my head round what is causing it because there isn’t even anything above 10A going through that circuit.

Now this would be fine if it was just this circuit going off. As I could slowly work my way through what is wrong with it. But over the past 4 days as well as my garage MCB tripping I’ve had both upstairs and downstairs rings trip off, and as of today the upstairs lights have also tripped off.

They sometimes trip at the same time as the garage but they have tripped off just by themselves. I have tested all the rings and they appear to have no issues with the cables.

What makes it even more confusing is I have a split RCD DB, not once has the RCDs ever tripped on either side. But I have had the garage MCB trip on the left RCD and lights on the right RCD trip at the same time. This has also happened with the ring mains in the house.

I can’t get my head around how a single circuit on its own like the garage can appear to have an effect on another single circuit that are on separate RCDs.

Could it be an issue with the DB itself? Because as far as I can tell after testing each cable they haven’t got any issues. The MCBs can stay on for hours and just trip off.

Any help into this would be great as I’ve been pulling my hair out for days now.

Thanks, Lewis

 

[timhoward]

It might help to stick a photo up of your main house consumer unit and the consumer unit in the sun house.
Before going any further please also press the test button on both RCD's and check they actually work.
(By the way the weird cable is called split concentric)

 

[Lewis Curle]

It might help to stick a photo up of your main house consumer unit and the consumer unit in the sun house.
Before going any further please also press the test button on both RCD's and check they actually work.
(By the way the weird cable is called split concentric)

Thanks for the info on the cable, I’ll get some photos when I get home

 

[pc1966]

Very strange indeed.

When you say you have tested the insulation resistance was that using an electrician's style of meter that applies 250V or 500V DC during the test, or a multi-meter that applies a couple of volts at most?

Several circuits having similar faults is worrying as it suggests some common problem and if they are unrelated circuits that could be mice/rats chewing cables.

If you can do as suggested, try the test-buttons on the RCD to verify they are able to trip on a simulated fault, and get some photos it would help. Also if you can get a photo of the inside of your CU safely that would be good. So only if you can totally power it off at the main switch and are OK to remove it, photo it, and replace cover again without touching anything or disturbing the main 'tails' feeding it (large cables from the supply meter, etc).

 

[Lewis Curle]

Very strange indeed.

When you say you have tested the insulation resistance was that using an electrician's style of meter that applies 250V or 500V DC during the test, or a multi-meter that applies a couple of volts at most?

Several circuits having similar faults is worrying as it suggests some common problem and if they are unrelated circuits that could be mice/rats chewing cables.

If you can do as suggested, try the test-buttons on the RCD to verify they are able to trip on a simulated fault, and get some photos it would help. Also if you can get a photo of the inside of your CU safely that would be good. So only if you can totally power it off at the main switch and are OK to remove it, photo it, and replace cover again without touching anything or disturbing the main 'tails' feeding it (large cables from the supply meter, etc).

So the RCDs are working fine but just tonight the garage has went off twice and the down stairs ring has went off just now. What do you mean by CU?

For the record also my house is 7 years old and there isn’t many places rats or anything could get at the cables,

I tested the cables on all the circuits that were tripping and they have all passed a 500v insulation resistance test.

It’s just confusing me that if something in the garage was causing an issue on the sockets it would trip the now 16A MCB that I changed today from the 32A in an attempt to get it to trip. So as far as I can see there isn’t an issue in the garage circuits. They have all been tested as well and showing no issues.

But to trip a 40A MCB instead of a 16A is the most baffeling part to me. It’s also not consistent at all when it decides to trip

 

[Lewis Curle]

So the RCDs are working fine but just tonight the garage has went off twice and the down stairs ring has went off just now. What do you mean by CU?

For the record also my house is 7 years old and there isn’t many places rats or anything could get at the cables,

I tested the cables on all the circuits that were tripping and they have all passed a 500v insulation resistance test.

It’s just confusing me that if something in the garage was causing an issue on the sockets it would trip the now 16A MCB that I changed today from the 32A in an attempt to get it to trip. So as far as I can see there isn’t an issue in the garage circuits. They have all been tested as well and showing no issues.

But to trip a 40A MCB instead of a 16A is the most baffeling part to me. It’s also not consistent at all when it decides to trip

 

[UNG]

Do you have a fridge or freezer in the garage

 

[Lewis Curle]

Do you have a fridge or freezer in the garage

Only thing plugged in right now is my pc set up mate!

 

[timhoward]

It’s just confusing me that if something in the garage was causing an issue on the sockets it would trip the now 16A MCB that I changed today from the 32A in an attempt to get it to trip. So as far as I can see there isn’t an issue in the garage circuits. They have all been tested as well and showing no issues.

But to trip a 40A MCB instead of a 16A is the most baffeling part to me. It’s also not consistent at all when it decides to trip

It's hard to follow what you are saying. But exchanging breakers is definitely not going to help add clarity.

So we are looking for something that
a) is not leaking current to earth (the RCDs are functional)
b) is either 4 independent faults on different circuits or is somehow causing sufficient current to flow to cause any permutation of 4 different breakers to trip.

A similar photo of the inside of the smaller CU (consumer unit) might be worth doing.

(The neutrals from the SplitCon cable to the smaller CU are a right mess and possibly dangerous as it looks like the screw is biting on insulation. They need improving.)

@pc1966 - a very random thought - I don't suppose a broken supply neutral and very unbalanced phases between neighbouring properties could cause enough of a current hike to have this scatter gun effect?!

 

[davesparks]

I tested the cables on all the circuits that were tripping and they have all passed a 500v insulation resistance test.

When you say passed an IR test do you mean they had a perfect >999megohm or a just scraped through at 1megohm?

ut to trip a 40A MCB instead of a 16A is the most baffeling part to me. It’s also not consistent at all when it decides to trip

A 40A MCB and a 16A MCB will not discriminate on a fault, it is pretty much pot luck which one will trip first.

 

[pc1966]

@pc1966 - a very random thought - I don't suppose a broken supply neutral and very unbalanced phases between neighbouring properties could cause enough of a current hike to have this scatter gun effect?!

I would be surprised if you would see that happen a couple of times and not have blown some equipment up in the process!

To throw enough current to trip a 40A MCB is not going to go well with a PC, etc.

 

[pc1966]

I tested the cables on all the circuits that were tripping and they have all passed a 500v insulation resistance test.

When you did the IR testing of the cables did you do it more or less in the following way:

• The CPC (earth) of the cable remains connected to the installations main earth terminal at all times (basically you leave them in place as normally connected).
• After isolating both ends of the cable L and N (maybe just open the garage CU main switch as it should be double-pole to isolate those) connect N to E and then IR test L to N+E
• Then connect N to L and IR test L+N to E

I'm asking as it is common to do a quick and safe test of L+N to E since it avoids the risk of over-voltage on low power electronics like LED lights, dimmer switches, etc, and it usually picks up faults on T&E cable as that has the E (CPC) between L & N. However, in the case of SWA or split-concentric cable that has been crushed you might get a L-N fault and that would trip the over-current MCB protection without tripping the RCD protection for faults to E.

The reason for keeping the CPC connected is in case there is damage like a nail, etc, to true Earth but not to the CPC. With the CPC earthed then it shows up on the same test.

 

[Lewis Curle]

When you did the IR testing of the cables did you do it more or less in the following way:

• The CPC (earth) of the cable remains connected to the installations main earth terminal at all times (basically you leave them in place as normally connected).
• After isolating both ends of the cable L and N (maybe just open the garage CU main switch as it should be double-pole to isolate those) connect N to E and then IR test L to N+E
• Then connect N to L and IR test L+N to E

I'm asking as it is common to do a quick and safe test of L+N to E since it avoids the risk of over-voltage on low power electronics like LED lights, dimmer switches, etc, and it usually picks up faults on T&E cable as that has the E (CPC) between L & N. However, in the case of SWA or split-concentric cable that has been crushed you might get a L-N fault and that would trip the over-current MCB protection without tripping the RCD protection for faults to E.

The reason for keeping the CPC connected is in case there is damage like a nail, etc, to true Earth but not to the CPC. With the CPC earthed then it shows up on the same test.

I have more or less already done what you have described sir, but I’ll double check it again tonight just to be sure.

Something I noticed last night was when my partner started using the downstairs ring circuit to watch TV, turn lamps on and turn in the dishwasher. The downstairs ring tripped then the garage a few minutes later.

I’m unsure on how these circuits could be effecting each other but when my partner went to bed last night I have about an hour and half of no tripping on the pc. So I’m pretty sure there’s not an L-E fault anywhere like a dead short as it would trip instantly.

But I will double check the insulation resistance tonight,

Thanks all for your replies so far.

 

[Lewis Curle]

When you say passed an IR test do you mean they had a perfect >999megohm or a just scraped through at 1megohm?



A 40A MCB and a 16A MCB will not discriminate on a fault, it is pretty much pot luck which one will trip first.

They passed with the max reading off the tester. So when you say it will not discriminate on a fault what do you mean? The MCBs in the garage board have never tripped it’s always the 40A MCB one. So if there was an overload somehow being caused it would trip the 16A one first no?

Thanks for your reply though mate

 

[pc1966]

So when you say it will not discriminate on a fault what do you mean? The MCBs in the garage board have never tripped it’s always the 40A MCB one. So if there was an overload somehow being caused it would trip the 16A one first no?

Fault discrimination, now known as selectivity, is the ability of a cascade of protective devices to only trip the closest to the fault.

In the case of MCB they are selective on overloads that take time on the thermal trip side (so up to x3 or so from the rated value of the upstream device) but if you get a big fault, such as a short-circuit, then both devices would have the "instant" magnetic trip underway before the downstream device has actually opened and isolated the fault. Basically in many case both devices trip.

When cascading breakers when it matters, you need to add a short delay on the upstream device to give the downstream one time to act, but if it does not (or the fault is before it) then the upstream will go. Usually that means using a MCCB (at high cost) or a fuse (cheap, effective, but a pain to replace and typically means electrically-skilled staff).

If you look at commercial/industrial catalogues of electrical devices they often will have tables in the back giving you the selectivity limits for various combinations of fuse/MCB, MCCB/MCB, MCB/MCB and so on.