Mystery Fault with Electric Underfloor Heating

[Saaj]

I had electric underfloor heating cable and thermostat installed many years ago. It was a loose cable with Heatmiser thermostat. It’s been working fine all these years until September last year when I had issues with it whereby the fuse would trip if I tried to switch-on the the underfloor heating. I recently called an electrician over to take a look and all the cabling from switch to the thermostat was fine and whilst the thermostat was off the wall he tried switching it on and by magic the heating was working fine and it did not cause the fuse the trip. After, about 20 mins of running the heating it was switched off and everything put back together as the electrician had other work to get on with. At the end of the day, he tried switch on the heating using the thermostat and the fuse tripped again. The assumption after some diagnosing was that the thermostat had a fault so it was swapped with the thermostat from another room (this wasn’t a Heatmiser thermostat). This worked for 2 months and now I have exactly the same fault causing the fuse to trip if I switch on the floor heating.

I called the electrician and he is very confused that it worked for 2 months and the same fault re-appeared again. Could it be that this thermostat has also blown?

Are thermostat specific to type of cable or power etc?

I’ve run out of ideas as everything that you can physically see is fine (according to the electrician) and if the heating has worked for two months after replacing the thermostat so it can’t be the cable itself.

Any one come across this or has any pointers or ideas?

Thanks v much

 

[mikeyd]

What sort of fuse is tripping? We like pictures.
Worst case is you need the underfloor heating cable IR tested by someone who knows what they are doing and then you find there is a fault under the floor.
But from what you say maybe it is simpler. "and by magic the heating was working fine" gives some clues but as above, start at the start, pictures please. "Fuse that trips", thermostat, etc.
Did I say we like pictures.

 

[ericmark]

Fuses don't trip, so you must be talking about a MCB, RCD, RCBO. Question one what trips.

As electricians we have test equipment, we do not repair by a series of elimination.

It seems likely the RCD is tripping, and if shared with other items it may not be the UFH at fault.

 

[Murdochcat]

More and accurate information please

 

[ericmark]

If it is the RCD tripping, an electrician should have one of these the top one will measure in 1 mA increments, the bottom one is rather old, so 10 mA increments, so he should have been able to see how much each circuit is leaking.

The main problem is neutral - earth leakage, earth with no load will be the same voltage to neutral, so even with a short circuit it will not trip the RCD, but as the load increases voltage between neutral and earth will increase, so a fault on a toaster may trip the RCD when the kettle is used, which means we need to test with 500 volt earth to neutral clearly with main switch off, with a meter like this again an electrician will have these instruments, years ago we would

have two RCD's coving all circuits, not a very good idea, today however we use RCBO's which are MCB and RCD combined,

these are less likely to erroneously trip, but when they do, we have no idea if due to overload or imbalance.

With an intermittent fault, guess work is the only method, but in the main there are some indications as to what the problem may be, so can't really understand why a thermostat would be changed, it seems very unlikely that would cause a problem.

 

[Saaj]

@ericmark @Murdochcat @mikeyd
Hiya and apologies- have added these pictures and can add more if needed

The arrow shows the switch from where the power cable is going to the thermostat and the other picture is the RCD that is tripping.

Just to add further to my comments in original post.

Yesterday morning the RCD tripped only after trying to put the UFH on using the Thermostat. As soon as I did that the RCD tripped. To reset the RCD I had to switch-off the switch that’s powering up to and beyond the thermostat.

After resetting the RCD, if I flick the switch to on then the RCD trips.

What I am trying to explain is that the RCD tripped for the first time when I used the thermostat to switch the heating on and after resetting it- the RCD trips as soon as I flick the switch back to on position.

The electrician that came over to look had removed the thermostat and tested the cables in there and they all tested fine and soon after he put it all back together it worked fine. Few hours later, he tried to switch it on again and the fault reappeared which is what prompted us to change the thermostat and in doing so it worked for best part of two months only for the same symptoms to reappear.

Does this help?

 

[Murdochcat]

You could be suffering from accumulated leakage. One of the huge down sides of dual RCD or single RCD boards

This happens when more items are switched in that contain electronics

You need somebody with an earth leakage clamp meter

Next time this happens, switch off all the MCBs to the left of the RCD which has tripped, then turn back on the circuit with the underfloor heating

Also and very importantly what test results did the RCD give when it was tested? It could be over sensitive too

 

[EFLImpudence]

What I am trying to explain is that the RCD tripped for the first time when I used the thermostat to switch the heating on and after resetting it- the RCD trips as soon as I flick the switch back to on position.

But is the thermostat still 'on'?

 

[Saaj]

You could be suffering from accumulated leakage. One of the huge down sides of dual RCD or single RCD boards

This happens when more items are switched in that contain electronics

You need somebody with an earth leakage clamp meter

Next time this happens, switch off all the MCBs to the left of the RCD which has tripped, then turn back on the circuit with the underfloor heating

Also and very importantly what test results did the RCD give when it was tested? It could be over sensitive too

Thanks @Murdochcat

I don’t know about the test results for RCD will need to check with the electrician.

I will try out what you have suggested but this has been the setup for best part of 10 years with underfloor heating being in place for as long as that but I will try it out. When I took the pictures to create the fault none of the appliances or switches controlled by those MCBs were in use.

One thing I did struggle with when I took the pictures was that I could not put the RCD back on - it just wouldn’t stay put and after 10 or 15 attempts it decided to stay put. The heating switch that caused it trip was off so it had no reason to not stay put.

Thanks

 

[Saaj]

But is the thermostat still 'on'?

@EFLImpudence The thermostat comes on when I flick the switch but it’s only for about a second. The RCD trips in about a second or so when I put the switch back to on position

 

[ebee]

You need somebody with an earth leakage clamp meter

Earth Leakage Clamp Meter , what defines the "earth leakage" part of the name compared to a Clamp Meter ?
I just thought that a clamp meter is a clamp meter and depending upon its range, resolution and accuracy then it can be used as the user decides

 

[EFLImpudence]

Earth Leakage Clamp Meter , what defines the "earth leakage" part of the name compared to a Clamp Meter ?

By definition a clamp meter which measures accurately a low enough current so that it may be used to measure what we call earth leakage in connection with RCDs.

 

[ericmark]

Both are clamp on ammeters, neither has earth leakage in its name, one on the left will also measure frequency, but AC amps only in 10 mA increments, one on the right will measure both AC and DC amps, in 1 mA increments, and also non-contact AC volts. In theory the limit for background leakage is 30% of the RCD rating, so with a 30 mA RCD that's 9 mA, so would need to use one on the right to show if within limits. However, 10 mA is not far off 9 mA, but it can't show the 3.5 mA limit before special plugs and sockets are required. So the yellow one would not show the 8 mA shown here. But it seems many electricians think the insulation tester is good enough, but this meter only measures DC and our mains is AC at 50 Hz, so leakage due to capacitive or inductive linking will not show up. When the RCD will not hold in, then no option but use the insulation tester with 250, 500, or 1000 volt DC, but when at a stage when it only trips every so often, what we are looking for is the cumulative leakage from many items.

I have never worked out why when completing a minor works or installation certificate why the leakage is not recorded. However, as it stands, we are looking at what @Saaj can do. So step one is unplug items, not just switch off, as a neutral - earth fault is often still there even when the appliance is switched off, as many switches only switch the line.

I would say unlikely a thermostat will cause earth leakage tripping, as most I have used, the earth terminal is only to allow the earth to be terminated as required by 411.3.1 and does not connect to anything within the thermostat.

The UFH mat may well have developed a fault, however it is also a high load, so any other items with an earth-neutral fault could cause the trip when the heating load is applied. We, I think, can assume the electrician used is not completely incompetent, and if the heating mat was faulty he would have found it? So seems likely it is an earth-neutral fault with some item not in use when he tested.

The other option is walking on the floor is flexing some part, so the fault is intermittent. Can also be fixing screws, had it where a socket fixing screw sometimes touched a neutral wire which had been nicked with a knife when first installed.

In the main we get a feeling for likely results, having measured the earth leakage many times, we can look at a home, and say this home will be near the limit, or this home a single RCD on many circuits should be no problem. To my mind, 8 circuits on one RCD is over egging it a bit, back when fitted likely not an option to do anything else, single width RCBO have not been around that long, and 8 x £20 to swap to all RCBO for at least that side of the board, is clearly expensive. Some Wylex up to £35 each, if the consumer unit is high integrity type then can have just one or two RCBO's, but I do not really know what fits that board.

However, my first action would be to swap the UFH to the other side of the board, and see if fault also swaps sides. That does seem a cheap option to prove if UFH or something else.

As electricians, we are supposed to advise the client or the problems with only two RCD's or lack of SPD, not sure if I really know the problems with a lack of the latter, and to try and explain what saving a few £100 on a twin RCD to an all RCBO consumer unit will mean is hard, some people can get away with twin RCD's for years, but you as customer has to do the risk assessment, and having lost two freezers full of food just before I moved, when I came to change my consumer unit, I paid the extra to protect the freezers, not that I really needed to, as now the freezers are on RCD sockets fed from my UPS (uninterruptible power supply) so only the back kitchen has a freezer on same circuit as other items. Front kitchen freezers have their own RCD sockets.

But you have to decide what risk you want to take. The electrician should not do that for you.

 

[ebee]

Both are clamp on ammeters, neither has earth leakage in its name,

Thanks Eric, that was my point - a misnomer can be a misleader , just plain wrong.
A few years back the Wholesalers invented a term and then folk followed it, so it became common, but still wrong "17th Edition Consumer Unit" there is no such animal exists.
Any consumer unit can be made to comply or made not to comply with the 17th Edition.

"High Integrity Consumer Unit " is another daft term too.
How can any one claim that a consumer unit with two RCDs and a few ways for MCBs or RCBOs be high integrity? Surely a consumer unit with a plain DP Mainswitch then all RCBOs circuits is better classified as High Integrity? If you need to use that name.

Twin RCD units better named a bit better than low integrity (Front End Single RCD rather than no RCD at al)l?

You can test current, leakage or not, high level or low level, just how you like, it does not need nor deserve the term "Earth Leakage"

 

[EFLImpudence]

You can test current, leakage or not, high level or low level, just how you like, it does not need nor deserve the term "Earth Leakage"

So - what do you call current that is leaking to earth?