In the last few weeks we've had a couple of power cuts caused by the RCD in the CU tripping but it's not obvious what it causing it. None of the MCBs have tripped.
It seems that each time the washing machine is on its spin cycle and I am wondering whether that is responsible. However, we have had the washing machine on plenty of other times and nothing has happened.
Is there anyway to pin point what is causing these trips? Is a washing machine a likely culprit?
I suppose a board full of rcbos would make it easy to identify the problem but I haven't got one of those ;) But didn't I read somewhere here that rcbos only switch the live not the neutral (makes sense since the neutrals are connected to the bus bar not the rcbo), so is an advantage of using rcds that it gives you protection on both poles?
No, RCBO's have a live and neutral connection. The neutral from the circuit is connected to the RCBO, and there is a flying neutral connection on the RCBO which is connected to neutral busbar.
I fitted out the new CU with RCBO's. More expensive yes, but in the scheme of things the extra cost doesn't seem something to worry about to much.
It could be caused by anything. Best bet is to begin by insulation testing all your appliances - its very simple. I'd start by doing it with a multimeter, which will catch a lot of such faults but not all. Test resistance on the highest range from L&N to E, you're looking for any appliance with a low reading.
If so, it worth remembering that while these can work fine for some folks for years, they can also cause no end of trouble for others!
I would suggest you start reading here - since this is quite a "deep" question:
formatting link
I suppose a board full of rcbos would make it easy to identify the
Not only makes it easier to isolate a fault, also makes it less likly to happen in the first place (since each circuit has its "own" 30mA leakage budget, rather than one share between them all).
The RCBOs are connected to both poles (which when you think about it, is the only way they can detect an imbalance between current flow between them). However some them will only open the line when tripped. (the double module width ones usually do both poles, and the single module ones, just the one. However you need to check the manufacturer/model specific details).
Of course yes, I wasn't really engaging brain in my answer....
checking on a spare single pole MK Sentry one I have here, the little diagram on it certainly seem to say it switches just the live (as you'd expect for a single pole one...)
In the last few weeks we've had a couple of power cuts caused by the RCD in the CU tripping but it's not obvious what it causing it. None of the MCBs have tripped.
It seems that each time the washing machine is on its spin cycle and I am wondering whether that is responsible. However, we have had the washing machine on plenty of other times and nothing has happened.
Is there anyway to pin point what is causing these trips? Is a washing machine a likely culprit?
I suppose a board full of rcbos would make it easy to identify the problem but I haven't got one of those ;) But didn't I read somewhere here that rcbos only switch the live not the neutral (makes sense since the neutrals are connected to the bus bar not the rcbo), so is an advantage of using rcds that it gives you protection on both poles?
Sorry for the delay, I've been away from the computer for a couple of days.
Sorry for the duplicate post. The OP and all the replies never showed up on my computer but other posts I made at the same time did, so I assumed it had not got through. I almost reposted with a "sorry if this gets duplicated" line appended but I was so sure it hadn't worked first time that I sent the post a second time without modification.
I used to read the group in agent, sorted by subject. I read many interesting posts and learnt a lot but could not join in because the posts I was reading were months or even years old! So I've changed to sorting by date.
It turns out that someone posted a few years ago that when they replaced white plastic sockets with metal ones, their rcd tripped and they used the same subject line. Agent thought my post was a continuation of that so all the replies were filed under this old post which was so far down the list I never would have found it.
Thanks for the replies, I haven't had time to read them all and act upon them but will report back shortly.
As you can tell, I've never seen or used an RCBO before. Of course, when I think about it, it would need live and neutral to work as an RCD. I'm sure there's some combination of punctuation to make a blushing face that I should use here!
From what's been said, it sounds as though single width ones only switch the live. If you have to use double width ones to switch both poles, you'll quickly use up CU space. Is there any disadvantage to only switching one pole?
Yes it is. Until now we've never had any nuisance trips in the years we've lived here, so I've not seen the need to update it. My worry would be would the wires be long enough to reach if we replaced with a new CU or would I have to run new cables to the first and last socket of each ring, each light, etc? Also, in order to do it myself, I'd need to buy an expensive meter to measure the resistances; at the moment I've only got a "basic" multimeter.
Is a split CU any better though? If you have a nuisance trip rather than the whole house going off, half the house does and surely that's just as annoying/inconvenient?
We had another trip yesterday, again when the washing machine was on its spin cycle, so I am sure that is the culprit but we had three washes and it only tripped on the second wash.
I read the faq you linked to and it begins by saying something along the lines that nuisance trips are very difficult to pin point and that is proving to be very true.
You need to switch both poles to isolate a circuit. Hence a single pole disconnection would not resolve a tripping issue caused by a neutral to earth short. On TT supplies its a good idea to isolate both poles when working on a circuit, since there may be a significant potential difference between N an local E. On the other hand, disconnecting N has more server implications in the presence of other wiring faults like a borrowed neutral.
Alas past performance is not a good indicator to the future here. Things can and do change.
The wires may or may not be long enough. If replacing a relatively modern CU (i.e. a DIN rail style one) with a modern one, then chances are they may be. Perhaps if replacing an old wylex rewireable fuse one then no. However the solution in most cases is a simple as crimping extensions onto the end of the wires which are too short. If they are all to short then using a set of DIN terminals in the position of the original CU is an elegant solution. Rather like in Andy's picture here:
formatting link
need to buy an expensive meter to measure the resistances; at the
You can hire multifunction testers if needs be, or there may be someone here local enough to lend you one.
A split is better in the sense that the split makes the probability of a trip smaller as well. You have fewer circuit protected by each RCD - and so more overall leakage budget to play with. (you could argue in the presence of a real but minor fault this is not such a good thing!)
Splits these days are not necessarily two way. The extreme version being an all RCBO board with a RCD per circuit. That way if it is the washing machine, you only trip its socket circuit. If its the washing machine and its just because the leakage from it is slightly high, then you don't get a trip at all.
When on the borderline of the trip threshold,you can get hard to predict behaviour such as you are seeing.
Yup they can be. You have to be quite methodical in tracking down the cause, and even then it can be hard because you are often dealing with the interaction of multiple factors - take any one of them away and the problem vanishes, and yet each factor in its own right is not conclusively a problem.
What sort of CU have you got at the moment? Are RCBOs available for it?
Thanks. Are all the required tests in the latest edition of the OSG and/or wiki? I'd like to see what I would have to do to determine whether they are within my capability first. Is it just measurements of L, N, and E resistances and a test of the RCD?
It's a crabtree one with one rcd covering 8-10 (will check later how many) MCBs. I guess that makes it 16th edition? It uses the starbreaker mcbs available from edmundsons, screwfix, etc. There are RCBOS. Are you suggesting I fit one to the washing machine ring and see what happens?
We've done some more washing since and no trips, so like you say perhaps lots of devices are contributing and the washing machine pushes it over the edge on certain days?
Yup basically. Plus its worth doing an insulation resistance check on each circuit, checking the earth loop impedance. The OSG gives a reasonably detailed account of the tests - some in more detail than you require just for a CU swap (for example they explain how to establish the true topology of ring circuit - whereas for this level of confidence you basically want to know you have continuity of the conductors, and the round trip resistances look plausible).
If 8-10 MCBs is a subsection of the CU and there are other non RCD protected ways, then that sounds like a 16th edition style split load CU. If 8-10 MCBs are all the ways and the RCD protects all the circuits, then that is more of an early 15th edition style arrangement.
Yup, if you have a reasonable hunch that its that circuit causing the problem, and you can move it to a non RCD protected section of the CU, but then add a dedicated RCBO for it instead you would kill several birds with one stone. Firstly you may prevent nuisance tripping of all the other circuits. You may also prevent tripping on the problem circuit
- depending on how bad the leakage problem is.
Its possible certainly. I have had a similar problem with a tumble drier in the past.
Things like heater elements in washing machines are known for these problems. They can get tiny cracks or pinholes in the metal case of the element. Water seeps into the hygroscopic mineral insulation and they start leaking to earth. Use them a few times and you drive some of the moisture out which reverses the problem slightly for a while. Leave it unused a few days and the problem returns. (You can never really fix one once its got wet since the mineral insulation once combined with water undergoes a chemical change that lowers its resistance, and heating alone to normal operating temperatures can't reverse it)
It's 10 mcbs and they are as you describe, so it must be a 15th ed install.
I don't have a split section so unfortunately I am stumped.
Thanks. I had read about washing machine elements but didn't quite understand what it all meant until you explained it. How would a tumble dryer leak since it shouldn;t contain water? I thought their heaters were just wires like a fan heater/hair dryer.
Yup, you are in a classic situation of not wanting to start from where you are!
10+ circuits on one RCD is asking for trouble anyway in a modern household. There will be folks who get away with it, but its not something you can count on.
However, depending on the CU, it may be possible to introduce a split. Depends on what bits you can get for it, and if it has the option of a split neutral bus bar.
(I had the same problem here when I moved in - all ways on a single RCD, and an obsolete CU for which one could not get spares, so the only sensible solution was to rip and replace the lot, and do some sensible partitioning along the way).
Well indeed they are - however they are in an environment with lots of moist washing and air being blown about the place. It does not take much accumulated moisture to leak a mA or two. When added to that from all the other circuits, that may be enough.
Thanks. So possibly each circuit has a "leak" of 3mA and together they are adding up to 30mA, whereas if I had two groups of 5 MCBs, each split would only have a leak of 15mA and not trip? I think I'll post asking about split CUs in a couple of days time ;)
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.