commonly encountered when your power is supplied by overhead wires and does not include the provision of an earth connection by the supplier.
These are unusual in that the resistance to earth is usually much higher than with other earthing systems (since it has to rely on a local earth spike). Hence under fault conditions that result in (say) a short to earth you won't get the massive fault current flow that you would with another type of earth. This has a knock on implication that the fault current may not operate an over current device (fuse / MCB etc) fast enough (or at all) to clear the fault.
So TT setups require RCD protection on all circuits - even lighting ones. To mitigate against the problems that can result from this, it is usual to use more than one RCD. Often a primary one with 100mA trip that feeds all circuits, and another 30mA trip device for the socket circuits.
John, This is really interesting. Yes we do have a TT system in that case. Our mains power is supplied to a farm building next door and into our own meter which then has a 100 ma RCD (all within the building) which then feeds to overhead wires to our house (we bought the farmhouse from the farmer who has retained the building). When we first moved in this supply went straight from teh overhead wires to an old style circuit board which never gave a problem. Following an electrical inspection, it was recommended to put in a new CU with a 30Ma RCD protecting the 13 amp sockets. Immediately this was done then the fault stated to happen. i.e. turning a light switch off sporadically trips the 30 Ma RCD protecting the sockets. The 100 MA in the farm building never trips. So in summary we have the following: mains to our meter, then to a single 100 Ma RCD then into overhead wires for about a 25 meter span, into a new consumer unit with one 30 Ma RCD protecting the 13 amp sockets and no RCD protection on the lights. We have had the earth spike upgraded as part of the recent work. I would be grateful for your thoughts.
OK your description does indeed sound like a TT system[1]. The 100mA RCD ought to be a time delayed type (something like a MK 6400S). This will ensure that in the event of a big earth leakage on the socket circuit it will always be the sockets RCD that trips first[2] before the main one gets a chance to trip and take out the lights.
Back to your tripping issue:
It is still possible that you don't actually have a wiring fault at all, and there is an appliance that is causing a high leakage. This may be leaving the socket RCD right on the brink of tripping, such that the smallest disturbance pushes it over the edge. Typical candidates are items on the socket circuit that involve water and heaters (washing machine, TD, DW etc). Outside power sockets or feeds to outbuildings that may get damp are also prime candidates. Tracking down these types of faults can be time consuming. You can still eliminate the fixed wiring from the equation using methods described previously. You could then set about trying to identify where a dodgy appliance may be. The "binary chop" solution can help - unplug half the appliances on the ring and see if it trips. That may help isolate the problem device to half of them etc. By adding in or removing half of the remaining devices each time you should be able to home in on the problem one quite quickly.
Getting a sparks with a RCD tester can help. These inject a known (user selectable) fault leakage current into the circuit and time how long it takes the RCD to operate. Seeing how much fault current you need to add to empty circuit and then a circuit powering its usual devices can give an indication of how far into your "trip budget" they have eaten.
[1] Note for completeness: it is also possible to have a PME setup with TT. The clue is looking at the head end where the cable terminates before the meter. A PME head end would have a substantial earth wire connected into the same block that the neutral tail emerges from before it goes to the meter. However from your description this is probably not your setup.
[2] Cascaded RCDs need the time delayed type at the head of the chain because you can't otherwise guarentee that a large leakage current will not cause the upstream one to trip as well as, or in place of, the downstream one.
Yup. checking all appliances with a multimeter on highest R setting from L&N to E is probably the simplest move forward. While this cant pick up 100% of faults, it will almost always pick up appliance leakage faults in practice. I cant recommend splitting applinces between different circuits to see what trips as its just too unreliable to gain satisfactory narrowing down of the fault, and very time consuming.
If its not an appliance fault, I come back to standard poor design as the other likely cause, with rcbos as the compliant solution.
The inspection report did say for the single 13 amp circuit (with rcd protection) that keeps tripping:
" investigate high R1 + R2/Zs for circuit" could this imply the earth leakage problem that you are possibly describing.
Also, I am trying to work out how an earth leakage on one of the appliances on this 13 amp circuit is then tripped by switching the light off on the lighting circuit? Where is the connection, is it through the earth system (TT, we now know)
Sorry, you can see that my knowledge is low with this matter. Regards, Dick
Since we now know that the earthing is to a local earth electrode (TT system) it is to be expected that Zs will be high, but it should be similar to within an ohm of so on all circuits, since the earth electrode resistance dominates over the resistance of the copper wires. Were you given a "schedule of test results" - this should give values for either or both of R1+R2 and Zs for each circuit. Any values you can quote for both relevant circuits, lighting and power, might be of interest.
Well answer>> My thoughts are turning to some sort of earth fault here. Perhaps a
which might be getting close. I'm wondering whether you have a major appliance, such as a washing machine or dishwasher (which will contain a mains filter (suppressor)) plugged into the socket with the high R1 + R2 reading. If so, does unplugging it make the problem go away?
Also do your test results indicate that the insulation resistance (L+N to earth) is OK on all circuits?
Also bear in mind that the RCD doesn't care about how the leakage current gets to earth - it doesn't necessarily have to flow in the CPC(s) (earth wire(s)) of the circuit concerned, or any other circuit. The RCD simply trips when the L & N currents are significantly different.
Have you got anything plugged into the ring circuit that involves an extension lead (or a circuit spur) going outdoors?
What, exactly, was going on when the company which installed the new CU were prompted to say:
and is this the same company that did the original test report?
I thought it more likely that the rcd would be feeding 2 rings, with the OP not differentiating. One rcd tripping all sockets is consistent with 2 or more rings, and 2 or more is a lot more likely I'd say. But as usual I/we dont really know.
_If_ thats the case, as is statistically most likely, fitting RCBOs would:
- probably solve the immediate nuisance tripping problem
- avoid most future nuisance trips in an apparently issueful installation
- mean that power to much equipment would stay on in the event of an earth leakage fault
- and finally prevent any n/e fault on one ring from tripping the lot, since RCBOs break N as well as L.
Ie it would most likely solve both immediate and longer term design problems.
But without harder details we're still playing guessing games.
I thought it more likely that the rcd would be feeding 2 rings, with the OP not differentiating. One rcd tripping all sockets is consistent with 2 or more rings, and 2 or more is a lot more likely I'd say. But as usual I/we dont really know.
_If_ thats the case, as is statistically most likely, fitting RCBOs would:
- probably solve the immediate nuisance tripping problem
- avoid most future nuisance trips in an apparently issueful installation
- mean that power to much equipment would stay on in the event of an earth leakage fault
- and finally prevent any n/e fault on one ring from tripping the lot, since RCBOs break N as well as L.
Ie it would most likely solve both immediate and longer term design problems.
But without harder details we're still playing guessing games.
I thought it more likely that the rcd would be feeding 2 rings, with the OP not differentiating. One rcd tripping all sockets is consistent with 2 or more rings, and 2 or more is a lot more likely I'd say. But as usual I/we dont really know.
_If_ thats the case, as is statistically most likely, fitting RCBOs would:
- probably solve the immediate nuisance tripping problem
- avoid most future nuisance trips in an apparently issueful installation
- mean that power to much equipment would stay on in the event of an earth leakage fault
- and finally prevent any n/e fault on one ring from tripping the lot, since RCBOs break N as well as L.
Ie it would most likely solve both immediate and longer term design problems.
But without harder details we're still playing guessing games.
In reply to the points raised by Andy, I will post the R1 + R2 test result values this weekend when I am home. To answer the rest of your points, this lighting circuit does consist of old style flourecent lights. There is also a connection running to an outbuilding from the consumer Unit (protected by the 30 Ma rcd) which you could describe as "very old and dangerous" (and damp) which will be replaced. This circuit to the outbuilding has 13 amp (2 freezers and dryer) and lighting on the same circuit (tripping one), I think, with the outbuilding lighting as a spur. On the 13 amp ringmain (tripping one) in question (with high Rs) there is a washing machine, diswasher, etc. This same ringmain goes to the sheds so perhaps this is where the lighting and 13 amp ringmain come into contact (through the spur)! In another part of the house, but on the same lighting circuit, that causes the trip, I know there are some wall lights (metal) without an earth. These will clearly need replacing which we will do soon. The first company that did the initial test (and found the high Rs) was different from the company that did the work and put the new RCD protected consumer unit in. The fault did not exist, or at least the tripping did not happen when the first testing was done with an old style fuse board, and single 100 amp RCD protecting the whole system at supply source. Thanks for your help with this. I am certainly learning a great deal about electricity! Regards, Dick
My last response is at number 28. Also, the following is on respective sides of the CU. As the lables are written under each switch on the CU.
Non RCD side : aga cooker, immersion, lights for wc larder office, doorbell timeclock, lights kitchen dining room hall, (THIS IS THE CIRCUIT THAT CAUSES THE TRIP) lights upstairs lounge,
RCD side: sockets larder, sockets single kitchen, sockets utility, sockets outbuilding, sockets heating kitchen dining room, sockets study lounge bedrooms, shower Dick
Oh right, well this, and your other post about the outbuilding circuits, changes things rather a lot! This now starts to sound like an old and rambling installation and I doubt that we're going to diagnose the problem by remote control, as it were. It would make things clearer though if you could repeat gist of the other post, listing what's on each of the circuits, but using the circuits list above as headings, IYSWIM.
Mention of lights with no earth is interesting, and dates the lighting circuit(s) to pre-mid-1960s. I'm wondering whether some lights, especially the old fluorescents, have been earthed to nearby power sockets, "borrowing" the earthing of a ring circuit? This might be a significant factor in the nuisance tripping - as is the fact that a "very old and dangerous" (and damp)" outbuilding is also fed via the same 30 mA RCD.
You mention that there is still work to be done, re-wiring the outbuilding and sheds, and house lighting circuits(?). When this is done the problem may well disappear, and the specific cause might even be found.
I suggest you find another electrical firm to do the work though, or use the one which did the original PIR. I'm quite unimpressed with Firm No. 2, who appear to have walked away, knowing there was a problem, have put a circuit feeding an outbuilding (and then sheds?) on the same RCD that's protecting sockets in the house, and also quite unnecessarily put the shower on the same RCD. They also seem to have connected old lighting circuits with no CPCs (earths) to a new consumer unit. Were you given an Electrical Installation Certificate, or any other documentation for that matter, by this firm?
To have that lot on one RCD is just foolish. I would not re-employ whoever did that.
Andy said remote diagnosis is unlikely, I think so too. The contradictory information means too many uncertainties as well as complication, plus no ability to megger circuits from afar, or look at them to see what sort of state theyre in.
FWLIW unearthed metal lights could either be 60s or earlier, or they could be brand new. Unearthed metal lighting is right back in fashion.
Whatever you do you'll need to redo the CU, and doing so is likely, but not certain, to solve your problems. So thats what I'd do first. 6 RCBOs would set you back =A3200, you could halve that by combining socket circuits and using 3 RCBOs. RCBOs should be put in the non-rcd side. Its quite likely that some of those circuits on the RCD dont need to be connected to any RCD, so you may end up with 2 or 3 RCBOs needed, and probably the problem solved for =A3100 or less.
And if a problem does persist, it will then be on one circuit only, and not take out most of the house, fridge freezer included. You would thus have a fraction of the tracing work to do.
Could I thank you all for your time in trying to determine this electrical fault. It is clear that I am unlikley to be able to provide you with all the detail to allow you to identify the exact cause by email. I have therefore decided, as suggested, to have the wiring done in the outbuilding and earthing connected to the sidelights in the lounge. If that doesn't fix the problem (happened once in the last 2 weeks now) then I will look at getting a further independent inspection.
Thank you once again for your time and advice. If I sort the problem then I wil let you know. Your comments have been most useful and educational to me.
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.