Suggestions for better RCD discrimination

[Rockingit]

Have an existing old client with what's best described as a sprawling small-holding of a property.

There's an armoured external supply which branches into a submain for various sheds/workshops/stables (some of which are 150-200m) etc all of which have standard RCD final circuit protection and all the feeders are SWA, so there's not really a design need for an upfront RCD coming out of the house, except.... naturally the outgoing SWA only starts on the opposite side of the building to the CU, when the original builder put it all in there's standard 6242Y going from the CU through the fabric of the building to feed it - so that does require 30ma protection. Needless to say, this is leading to nuisance tripping of the feed RCBO in the CU.

Board is an all RCBO Easy9 that I updated to a couple of years ago, and Schneider naturally don't make anything so sensible as an S type RCBO for them (have asked technical). If I fit an external unit to the CU it will have to be a mess of replace RCBO for MCB > OUT > RCBO > Back in again and joint to the 6242Y. Equally, running a new length of SWA out and jointing on to the external feeder isn't really an option either as it would involve significant ground works.

Suggestions welcome! R.

 

[Rockingit]

I know not compliant, but I would be debating as to fitting 100ma s type at the source depending on if the loss of supply could result in an accident.

So here’s a thing…. I need to get my head into the brown book and scour the regs but off the top of my head my recollection is that the requirement for cables in walls is simply 30mA, nothing said about S types at all??

 

[mainline]

So here’s a thing…. I need to get my head into the brown book and scour the regs but off the top of my head my recollection is that the requirement for cables in walls is simply 30mA, nothing said about S types at all??

It would not meet the disconnect time.

 

[pc1966]

Nobody makes a type S 30mA RCD as @mainline points out it would not meet disconnection times.

You have two issues for selectivity:

• One is when there is a 'hard fault' and so you need some delay to allow the downstream RCD to clear irrespective of the magnitude of the fault current.
• The other is for a 'soft fault' where the imbalance is small and rises slowly, or indeed is simply there all of the time due to accumulation of appliance leak currents, etc. For that you need a difference in trip threshold so again the downstream one always goes first.

So if designed you have no other choice then to go 100mA selective for upstream and then, to meet regulations, it has to be installed so accidental cable penetration is very unlikely. Yes, I know the BBC have some guidance limiting temporary supply RCDs to 100mA / 200ms max as it gives you some protection against a fatal shock, but that is not meeting the BS regs.

If you can clamp the sub-main and check the leakage you might find it is hovering in the 10-15mA region in which case it is not going to be cured by anything other then a change of upstream RCD to a higher current and all of the other implications of that. On the other hand, if it is normally below 10mA and the issue is the odd spike incident (like my oven that trips the RCBO a couple of times a year on switching off) then it might just be worth looking at ones of these:

A9R61263 - Residual current circuit breaker (RCCB), Acti9 iID, 2P, 63A, A-SI type, 30mA | Schneider Electric Global

Schneider Electric Global. A9R61263 - Residual current circuit breaker (RCCB), Acti9 iID, 2P, 63A, A-SI type, 30mA.

www.se.com

They claim to be less prone to false triggering while still meeting the overall low frequency trip current and times. Here is the FAQ comparing the A-SI type they call it with the normal type A:

What is the difference between A9D32620 and A9D33620 RCBO Acti 9 devices? | Schneider Electric Australia

 

[Rockingit]

Nobody makes a type S 30mA RCD as @mainline points out it would not meet disconnection times.

You have two issues for selectivity:

• One is when there is a 'hard fault' and so you need some delay to allow the downstream RCD to clear irrespective of the magnitude of the fault current.
• The other is for a 'soft fault' where the imbalance is small and rises slowly, or indeed is simply there all of the time due to accumulation of appliance leak currents, etc. For that you need a difference in trip threshold so again the downstream one always goes first.

So if designed you have no other choice then to go 100mA selective for upstream and then, to meet regulations, it has to be installed so accidental cable penetration is very unlikely. Yes, I know the BBC have some guidance limiting temporary supply RCDs to 100mA / 200ms max as it gives you some protection against a fatal shock, but that is not meeting the BS regs.

If you can clamp the sub-main and check the leakage you might find it is hovering in the 10-15mA region in which case it is not going to be cured by anything other then a change of upstream RCD to a higher current and all of the other implications of that. On the other hand, if it is normally below 10mA and the issue is the odd spike incident (like my oven that trips the RCBO a couple of times a year on switching off) then it might just be worth looking at ones of these:

A9R61263 - Residual current circuit breaker (RCCB), Acti9 iID, 2P, 63A, A-SI type, 30mA | Schneider Electric Global

Schneider Electric Global. A9R61263 - Residual current circuit breaker (RCCB), Acti9 iID, 2P, 63A, A-SI type, 30mA.

www.se.com

They claim to be less prone to false triggering while still meeting the overall low frequency trip current and times. Here is the FAQ comparing the A-SI type they call it with the normal type A:

What is the difference between A9D32620 and A9D33620 RCBO Acti 9 devices? | Schneider Electric Australia

So, here's a thing. Not conventional for a domestic, and not cheap, but an adjustable RCD could in theory be set for bang-on 300mS @ 30mA / 40mS @ 150mA and be compliant, basically relying on the downstream devices being inherently faster (which we all know they typically are, by a significant margin).

 

[pc1966]

So, here's a thing. Not conventional for a domestic, and not cheap, but an adjustable RCD could in theory be set for bang-on 300mS @ 30mA / 40mS @ 150mA and be compliant, basically relying on the downstream devices being inherently faster (which we all know they typically are, by a significant margin).

Yes - if the "adjustable" bit is continuous and not a set of pre-determined steps.

Last time I priced an adjustable RCD (actually separate Hager HR520 EL relay, sense toroid, and actuator to trip the Hager 125A 3P MCBs) it all totalled around £840 inc VAT at "trade" price.

 

[Pretty Mouth]

So, here's a thing. Not conventional for a domestic, and not cheap, but an adjustable RCD could in theory be set for bang-on 300mS @ 30mA / 40mS @ 150mA and be compliant, basically relying on the downstream devices being inherently faster (which we all know they typically are, by a significant margin).

The upstream RCD is probably seeing a lot more leakage than individual downstream RCBOs, because it's protecting the whole installation. It's likely to be closer to it's tripping threshold, so a small amount of leakage on a downstream circuit might push it over the edge, while failing to trip the downstream RCBO.
So the above might work, for some faults, might not.

 

[pc1966]

The upstream RCD is probably seeing a lot more leakage than individual downstream RCBOs, because it's protecting the whole installation. It's likely to be closer to it's tripping threshold, so a small amount of leakage on a downstream circuit might push it over the edge, while failing to trip the downstream RCBO.
So the above might work, for some faults, might not.

Indeed, I would clamp the sub-main to begin with to get an idea of the margin for whatever approach is needed.

But if its is close to 15mA already I don't see any real solution without fixing the initial cable run's need for 30mA additional protection. By time you look at an adjustable solution, etc, the cost is getting seriously large and it might still be a bit flaky!

 

[Rockingit]

Indeed, I would clamp the sub-main to begin with to get an idea of the margin for whatever approach is needed.

But if its is close to 15mA already I don't see any real solution without fixing the initial cable run's need for 30mA additional protection. By time you look at an adjustable solution, etc, the cost is getting seriously large and it might still be a bit flaky!

You're of course absolutely correct and we've had this entire thread without me being on site again yet, however my suspicion is that there's probably no accumulative leakage to speak of at all, what we're dealing with is purely fault current. Downstream of all of this is basically just lighting and typical resistive/inductive stuff in a workshop or two.

As for costs.... the last one I did a couple of years back was around the same £7-800 mark, was a fancy all-in-one Schneider ELCB jobby.

 

[pc1966]

So, here's a thing. Not conventional for a domestic, and not cheap, but an adjustable RCD could in theory be set for bang-on 300mS @ 30mA / 40mS @ 150mA and be compliant, basically relying on the downstream devices being inherently faster (which we all know they typically are, by a significant margin).

I wonder if that is what Scneider's "A-SI" really is? 30mA with a very short delay but still meeting times?

At £300 a bit much to buy one just to play with...

 

[pc1966]

You're of course absolutely correct and we've had this entire thread without me being on site again yet, however my suspicion is that there's probably no accumulative leakage to speak of at all, what we're dealing with is purely fault current. Downstream of all of this is basically just lighting and typical resistive/inductive stuff in a workshop or two.

As for costs.... the last one I did a couple of years back was around the same £7-800 mark, was a fancy all-in-one Schneider ELCB jobby.

Yes, not cheap any how.

I suppose the irony is if this had been found with a 100mA delay RCD and so no "additional protection" it would only get a C3 inspection code!

Assuming no muppet ran the cable through the bathroom of course...

 

[mainline]

The chance of someone putting a nail or something in the cable just at the right spot to receive a shock, and a 100ma S rcd not tripping is highly unlikely, i would say they have more of a chance of being injured with the sudden loss of power.

The chance of a customer paying £1000s to sort out a random tripping problem is even more unlikely.