Need some ideas - RCD tripping but everything tests fine....

[weevilward]

Hi all,
Customer reports sporadic nuisance tripping of RCD. I cannot reproduce it, or fault anything, except SOMETIMES on EFLI test it trips the RCD (megger mft 1553 'no trip' test).
I'm out of ideas here, can anyone help? This is what I've got:
- TT installation (small domestic);
- 30mA RCCB main switch incomer (Crabtree starbreaker);
- all circuits IR test 8MOhm to L&N-earth;
- RCD testing off load and on load all fine, ramp test 23mA off load 19mA with everything on.
- Ze about 160 ohms down the rod, drops to 12 ohms with all bonding connected.

 

[Deleted member 9648]

RCBOs with a solid neutral are far from ideal with any earthing system.
If there is a neutral to earth fault then the rcbo will react by disconnecting the line and not actually disconnect the fault.
Thus it appears to the user that a fault has occurred but it is now disconnected because the protective device has operated, when in actual fact the fault is still there.
Plus of course you now have a portion of the neutral current from the entire installation diverting through the fault.

Additionally manufacturers have their own requirements for maximum Zs for correct operation of their single pole RCBOs, which may not be achieveable on a TT system.

The point that springs to mind here is that a N-E fault on a system with no RCD element will remain undetected with the same risk of neutral current diverting down the parallel path caused by the fault, at least with a SP RCBO the user will have an indication that there is a problem. I've not come across any manufacturers who give their own max Zs values for standard RCBO's, can you give an example of this?

 

[Octopus]

No, it's tripping at 19mA on ramp test with everything on. (23mA disconnected). So there is nornally 4mA leakage.

Not sure I agree with that...

 

[davesparks]

The point that springs to mind here is that a N-E fault on a system with no RCD element will remain undetected with the same risk of neutral current diverting down the parallel path caused by the fault, at least with a SP RCBO the user will have an indication that there is a problem. I've not come across any manufacturers who give their own max Zs values for standard RCBO's, can you give an example of this?

yes a circuit without RCD protection won't have any means of detecting a N-E fault, however what it does do is disconnect all of the faults it is designed to detect and is capable of protecting against.
The rcbo does not disconnect all of the faults which it is designed to detect or protect against.

The first one that springs to mind is the little company known as Schneider, particularly the part of it formerly known as Merlin Gerin.

 

[static zap]

- can a dirty supply operate RCDs????

Not that I've noticed , but was thinking out of the box ,a neighbour dumping current into nearby earths ! (would need a lot)
( still suspecting a part of install )

 

[tom porter]

TBH if I have these sort of problems with a single RCD install I don't waste a great deal of time if IR tests with everything connected are good. Advise to change to RCBO's. Whatever you do nuisance tripping is almost inevitable with such an install, even if you find a problem chances are something else will trip it before long and then you'll be back again for SFA because "I've already paid you once to fix it". A single RCD don't comply any more anyway!

why does a single rcd not comply

 

[Octopus]

^^ you shouldn't install single up front RCD's any more - a reg about minimizing inconvenience - can't remember the reg number..

 

[Davisonp]

I know I've read somewhere about SP RCBO's not being suitable for TT systems but cant remember where and am still sceptical. In fact on our last NICEIC visit I asked the question and got a blank look.
If anyone can shed any light......

On Site Guide page 20, then 32, 33 and 34.
Says:
Full RCBO CU with main switch suitable for TN and TT.
Twin RCD CU with main switch only suitable for TN
Twin RCD CU with 100mA time delayed RCD suitable for TN and TT.

Many other configurations are also talked about.

 

[Davisonp]

No, it's tripping at 19mA on ramp test with everything on. (23mA disconnected). So there is nornally 4mA leakage.
I realise 8M is a bit low....
Think I'm gonna invest in a leakage clamp meter and have a deeper look - if I get called back that is - still no tripping 48 hrs on......

I tried what you are doing, but I ramp tested a tripping RCD at 19mA with all circuits on, then again at 19mA with all circuits off. Then I purchased a leakage clamp meter and found the leakage is 19mA with all circuits turned on. And soon identified the kitchen ring as the biggest leakage of 10.5mA.

 

[Octopus]

I tried what you are doing, but I ramp tested a tripping RCD at 19mA with all circuits on, then again at 19mA with all circuits off. Then I purchased a leakage clamp meter and found the leakage is 19mA with all circuits turned on. And soon identified the kitchen ring as the biggest leakage of 10.5mA.

Ouch... so its almost certainly accumulated leakage causing the tripping.

You'll find the earth leakage meter VERY handy

and 10/10 for your feedback ..... so many people don't!

 

[weevilward]

No Murdoch, this guy is not the OP , I am!!!!
I haven't fed back yet because I haven't had to go back to the job yet, it's still not tripped since I swapped the RCD (and I still haven't got round to buying me earth leakage tester yet...)
So it's looking more like a faulty RCD now.... (as unlikely as that seems).
I won't sing too loudly yet as I think it will trip again......

 

[weevilward]

I tried what you are doing, but I ramp tested a tripping RCD at 19mA with all circuits on, then again at 19mA with all circuits off. Then I purchased a leakage clamp meter and found the leakage is 19mA with all circuits turned on. And soon identified the kitchen ring as the biggest leakage of 10.5mA.

Ok that kinda proves that my method of ramp testing RCD with & without loads connected is not too reliable. Thanks for the info mate, I will go and buy a clamp meter!

 

[buzzlightyear]

time will tell weevil wobbly!

 

[buzzlightyear]

it is on the end of the line.

 

[static zap]

I won't sing too loudly yet as I think it will trip again......

The cat is out of the bag .
( best get clamping )

 

[Davisonp]

Ok that kinda proves that my method of ramp testing RCD with & without loads connected is not too reliable. Thanks for the info mate, I will go and buy a clamp meter!

Someone was asking me about the model of my clamp meter last week. Below is my reply
=================================
They do not make my clamp meter any more.

The nearest equivalent to mine is the metrel md9272 “Metrel MD9272 AC DC Leakage TRMS Clamp Meter - http://isswww.co.uk/metrel-md9272-ac-dc-leakage-trms-clamp-meter?gclid=Cj0KEQjwg47KBRDk7LSu4LTD8eEBEiQAO4O6r3YH-bGd0Dcc_TU7AGRy_t-V1zjTKssQGjcsBeudrN4aAhcx8P8HAQ” but at £300 it is over £100 more than I paid, and I got the cheapest I could find.

Have a look at megger dcm300e “https://www.test-meter.co.uk/earth-leakage-clamp-meter/megger-dcm300e-earth-leakage-clampmeter-ac/” which is roughly the same price.

The best value at the moment if you want a cheap leakage meter is TM-EL9809 Earth Leakage Clamp Meter “TM-EL9809 High Resolution AC Earth Leakage Clamp Meter - https://www.test-meter.co.uk/earth-leakage-clamp-meter/tm-el9809-earth-leakage-clamp-meter/#product_accordian_product_reviews” which is advertised as a resolution of 0.01mA at a cost of £115 so a good price but I do not know if it is accurate or reliable, although they do offer a calibration certificate so it appears to be accurate.