MCB tripping at random times

[Beelzebub666]

Hi all, hope you can shed some light on an odd problem a friend of mine is having.
The CU is an older one from Hager, main switch and MCBs, no RCD, with an isolator switch before the CU. The CU was replaced a few years back (old one still had fuse wire) on the existing wiring, which must be pretty old but at least isn't tough rubber sheath. The problem circuit does the kitchen, hallway, boiler and a light on a spur protected by a 5A fuse.

Starting recently, the MCB trips after a random time, generally an hour or so, sometimes several hours. I cut all the power, unhooked the neutrals and used a multimeter to check L-E, N-E and L-N, which showed (as far as I know) no cable faults as very high resistance. I replaced the MCB with a spare of equal rating, no difference. I cut off the back end of the circuit which had the boiler, electric towel rail and spur light on it, no difference. The only appliance I didn't unplug fully was a dishwasher, because the morons who did her kitchen have blocked it in pretty good, but it wasn't turned on.

That's as far as my amateur diyer knowledge takes me, what could be causing a randomly timed trip and how would the more knowledgable go about finding out?

 

[martinxxxxxx]

Is your multimeter able to measure in megaohms?

 

[Steve]

Is your multimeter able to measure in megaohms?

I doubt it, unless he has proper test kit costing £100's. Multimeters dont give very accurate readings in megaohms.

 

[martinxxxxxx]

Maybe but you never know

Assume makes an ASS out of U and ME

 

[Beelzebub666]

Nope, K rather than M, but it does have an infinity symbol which I gather is fairly large. It's a pretty old analogue one. Basically i was taking a good result to be when the needle barely twitched past infinity, and a bad result to be when it skipped over to the far end.

 

[Beelzebub666]

I tried swearing at the MCB and threatened it with a neon test screwdriver, but to no avail.

 

[ban-all-sheds]

Even if it can accurately measure resistance in the MΩ range, that's not the same as an insulation resistance test which bangs 500V down the cable...

 

[ericmark]

I would be looking for something on a timer. Could be fridge or freezer with auto de-frost. An immersion heater cutting in. It is very unlikely to be wiring either it is OK or not OK rare will it give fault like you say.

The clamp on ammeter is best tool to see if drain is always there.

From time to time one gets really odd fault. I had one where next door had tapped onto ring main. But these faults are rare. With no clamp on speed of electric meter can give some pointers.

Again I have had it where a old storage radiator was still connected but not used for years and someone had clicked the switch.

To have an earth fault where it trips a RCD is common but to trip a MCB it has to be a really big fault and very unlikely to be on fixed wiring.

Even in my own house I found I had an old time switch that had worked central heating that I had forgot about. I could hear faint noise in dead of night took ages to find as in garage.

I hope that gives some pointers?

 

[Beelzebub666]

As far as I could tell I had everything on the circuit unplugged or disconnected except this dishwasher I can't get to, which was not turned on and doesn't trip the fuse when it is. The wiring seems a little unorthodox as there's only one cable coming out of the box for what should be a ring main.
I've advised the calling in of a professional electrician, which i'm guessing is going to be pretty expensive for a Dr House MD kind of diagnosis. It doesn't sound like i'm going to find whatever it is with a multimeter and a neon driver.
Thanks anyway eric, once professional judgement has been rendered i'll post it here.

 

[plugwash]

Nope, K rather than M, but it does have an infinity symbol which I gather is fairly large. It's a pretty old analogue one. Basically i was taking a good result to be when the needle barely twitched past infinity, and a bad result to be when it skipped over to the far end.

What exactly is the range your meter is set to? 1K? 10K? 100K? something else?

Depending on what the range of your multimeter is even a "barely twitched past infinity" could be an indication that things are bad.

Also there is the voltage issue others have mentioned, insulation faults are very often nonlinear with a much lower effective resistance at mains voltage than at what a multimeter uses.

 

[ban-all-sheds]

Thinking about it, Eric's right - it's so unlikely to be a cable fault. Even if it is due to insulation breakdown that doesn't show up at multimeter voltages you're looking at something dissipating 10's of kW to trip the breaker - by now the heat from that would be having obvious effects.

 

[Beelzebub666]

Nope, K rather than M, but it does have an infinity symbol which I gather is fairly large. It's a pretty old analogue one. Basically i was taking a good result to be when the needle barely twitched past infinity, and a bad result to be when it skipped over to the far end.

What exactly is the range your meter is set to? 1K? 10K? 100K? something else?

Depending on what the range of your multimeter is even a "barely twitched past infinity" could be an indication that things are bad.

Also there is the voltage issue others have mentioned, insulation faults are very often nonlinear with a much lower effective resistance at mains voltage than at what a multimeter uses.

1K, but I think I may have found the problem, i'll explain below.

 

[Beelzebub666]

OK, there's a tiny little back extension which leaks down the side in one spot, where there are no electric outlets. Following the circuit round though, it does come through the wall at that point and duck under the tiled floor of the extension. Using my little multimeter set to 10V ac, I got a positive reading from the damp tile adhesive right next to the wall to earth, just a couple of volts but varying over time and occasionally shooting right off the 10v scale. When the fuse pops there's no voltage there, and it's just that one little patch that gives a reading.

Now, i'm no electrician, but I don't think floors should register any voltage at all. My theory is the cable was damaged when the cowboys who did the extension stuff put it in, and they figured covering it with tiles anyway would hide the problem and save them replacing it. Now the area is damp the moisture builds to form an arc, creating a short so bizarre that i'm not actually a DIY dimwit for not finding it sooner after all, given that I don't have an insulation tester.

Does that sound reasonable? Because smashing out a tile and taking off some plaster to find the thing is a bit more than I was planning to take on, just doing my friend a favour.

 

[ban-all-sheds]

Put on a pair of rubber-soled shoes, and see if you can provoke it by treading on that tile.

Or pour a jug of water onto it.

Then let your friend decide whether to start removing tiles.

It still seems puzzling though. You need 160A to fast-trip a B32 breaker, and even at 80A it'll take a minute to operate. Water is not actually that good a conductor, and a small amount getting into a damaged cable would surely be boiled away by 19kW of heat before the breaker would go? An RCD tripping yes, that only needs 30mA, but it takes so much current to trip an MCB that I have my doubts that a bit of water could do it.

 

[Beelzebub666]

Put on a pair of rubber-soled shoes, and see if you can provoke it by treading on that tile.

Or pour a jug of water onto it.

Then let your friend decide whether to start removing tiles.

It still seems puzzling though. You need 160A to fast-trip a B32 breaker, and even at 80A it'll take a minute to operate. Water is not actually that good a conductor, and a small amount getting into a damaged cable would surely be boiled away by 19kW of heat before the breaker would go? An RCD tripping yes, that only needs 30mA, but it takes so much current to trip an MCB that I have my doubts that a bit of water could do it.

It generally takes an hour or so to trip. You think this might be an unrelated fault then? And I am at least right in telling her that a voltage reading proves a fault, if not the fault, and she should have that fixed?

160A to trip one? Wouldn't that take out the company fuse?