RCD's in series

[La Poste]

I was having a chat with my electrical boss today about this and that and we discussed wiring a radial in a garden supplying various things like a greenhouse and a garden shed.

He was of the opinion that things like greenhouses should be TT'd with an earth spike and RCD protected.

The thing is if the radial supplying the garden is taken from a ring main in the house then the cable will already be RCD protected at 30ma.

This 30 ma RCD will effectively protect the ring main in the house and the garden radial, the radial will then be split from the main earthing in the house and TT'd.

What my question is is would it be acceptable to run two 30 ma RCD's in series?
I know this would not satisfy discrimination but it would give you a second layer of protection in case the RCD in the house failed.

If both RCD's tripped due to an earth fault then so what, no big deal just reset them both but should the house RCD fail then at least you have a second layer to protect you and also of course should the RCD in the garden fail then you would have the RCD in the house as a second layer of protection.

People talk about 100ma S type RCD's to offer discrimination but what is really wrong with putting two 30 ma RCD's in series, the only thing I can see would be a little inconvenience in that you would have to reset them both but weigh this against having a second layer of protection, especially if the garden is TT'd and disconnection in the event of an earth fault would be 100 % reliant on an RCD then I think it is a reasonable trade off.

Does anyone have any opinions on the matter?

Thanks.

 

[richy3333]

What do the Regs say about RCDs and discrimination?

The 100mA RCD would be on the ring (in your example) and thus n

 

[richy3333]

...and thus not necessarily compliant?

 

[needasparks]

I've had discrimination with two 30ma RDC's where neither tripped on a fault, so i would not take the chance.

 

[La Poste]

I've had discrimination with two 30ma RDC's where neither tripped on a fault, so i would not take the chance.

Wow, why do you think that happened?
Neither of them tripping.

Richy I am talking about two 30ma RCD's in series, forget about the 100 ma RCD.

 

[needasparks]

Wow, why do you think that happened?
Neither of them tripping.

Richy I am talking about two 30ma RCD's in series, forget about the 100 ma RCD.

Exactly, I took one out and worked fine after.

 

[MarkieSparkie]

There will be no discrimination between two 30mA RCDs, it will be a lottery if one or both RCDs trip under fault conditions. "Sods Law" dictates that the most inconvenient RCD will trip.

 

[La Poste]

Exactly, I took one out and worked fine after.

So it ends up being more dangerous.
I'd love to know the reason why, the technical explanation if there is one.
Thanks for the reply.

The idea was to make it safer, if one 30 ma failed you would always have the other to fall back on, if both tripped then not really a big deal.

 

[needasparks]

I think that is the big question, perhaps i just had a flukey one out of the two, but no matter what mA i tested at they would not trip so took one out of line and the problem went. All testing was fine after. I was of the belief also that at least one would work but as proved its not the case.

 

[La Poste]

I think I might try an experiment and see what happens, we have a 30ma RCD in this house protecting a cable that supplies the garden shed, I will stick a 30ma RCD on at the shed end, test it and see what happens.

 

[trev]

Mate your result will be unpredictable and random. With a set up like this it's luck of the draw.

 

[Knobhead]

You have to look at cumulative E/L over the two circuits.
RCD 1 is protecting the house
RCD 2 is protecting the garden Via RCD 1
The house in it’s self will have an inherent leakage current.

Lets assume both RCD’s have an equal threshold trip of 20mA. The inherent leakage in the house is say 10mA. That leaves us with a 10mA headroom. Add 10mA leakage from the garden then the house RCD trips, the garden RCD is quite happy it can only see 10mA.

You or your boss needs to have a rethink.

 

[netblindpaul]

I regularly fit 2 30mA RCD's in series to provide backup protection in certain applications.
Client is aware of the lack of discrimination and is happy.
End user has this explained to them in detail and they are happy once this is done.

If the downstream RCD only allows 10mA of leakage before tripping as per the example from Tony then in my application it can only be a good thing!!!

 

[La Poste]

Thanks for the replies, great debate.

The garden radial will be supplied from the house ring whatever happens, it is impossible to get to the fuse board to lay a new circuit so this leakage current will be the same whatever circuit protection we use.

I am stuck with the 30 ma RCD in the house as per regulations, it is a ring main and supplies sockets for general purpose, however if the garden is TT'd with let's say an EFLI of 150 Ohms then if the house RCD were to fail and an earth fault in the garden were to happen I will be left with a live frame to a greenhouse if everything is bonded as it should be.

I am trying to eliminate this risk by adding another layer of protection and this is one idea.
The other being to rod the ground so well that I get an EFLI low enough to trip a 13 Amp fuse not an easy thing to do.

 

[Engineer54]

Might be an idea to do the installations to these out buildings and greenhouse correctly, instead of spending time and money trying to get around things.

With 2 RCDs in series you will never know what is going to happen, both tripping together, not tripping at all, upstream tripping, down stream tripping. All depends on several things coming together at that particular point in time.

 

[netblindpaul]

One should always trip, there are even reasons for redundant protection.
However, in this application as E54 says, just do it right.
If you don't need redundant protection don't bother.

 

[1shortcircuit]

Just a question but why can't you configure your circuits so that the greenhouse/shed/outbuiling or whatever it is runs from an rcbo on an unprotected way?

RCD's will remain happy and so will outbuilding and discrimination will not exist

Sometimes the perfect solution is not always the easiest to achieve but it does exist one way or another Worst case, stick a two way consumer unit (main switch and rcbo) in before the consumer unit) :thumbsup

 

[La Poste]

It's just impossible to get a new cable from the CU.

The guy has had all new floors put down and the CU is in a sort of concrete bunker (and so is the main incomer) in the middle of the house.
He's had all sorts of extensions added in an outward direction from this concrete bunker and sealed all the floors with wood laminate.

I'm going to really take my time on this one but as it looks at the moment a new cable is nigh on impossible.

Supply is underground TN-S.

 

[1shortcircuit]

Surface mount trunking right the way through the property " Gert Lush" lmao

No such thing as impossible, just means that a sacrifice or two may be required

:thumbsup

 

[La Poste]

I will go and do some more testing tomorrow but right now I gotta crash.

Thanks for your help everyone.

I won't be starting this job until after my assessment that's assuming I pass but I will test as much as I can in the meantime.

 

[SPARTYKUS]

Always a way La Poste!

 

[richy3333]

Always a way. If you can't do he job pro

 

[richy3333]

There's always a way. If you can't do the job properly don't do it. A good reputation is hard earned and easily lost.

Do you have any photos of this concrete bunker?

 

[mydigitalhome]

I have been doing some testing work on a Nursing Home (new client) which had two 30ma RCDs in series (1 x DB supplying 5 x DB's approx 40m appart) and while testing it was tripping at both RCDs 1 x on one and 5 times on the other really annoying as made testing awkward ! Not ideal.

Back to original scenario of outside supply I always put the outside supply on the non RCD side of the board supplied via a RCBO 100ma and 30ma at shed CU this gives full discrimination for the outside circuit. How do the rest of you do it?

 

[1shortcircuit]

I always put the outside supply on the non RCD side of the board supplied via a RCBO 100ma and 30ma at shed CU this gives full discrimination for the outside circuit.

Are you sure about that?

What are you doing when you're testing at x5 on RCD test?

You would need a time delayed RCD as I understand it from reading posts on the forum?

 

[mydigitalhome]

The RCD testing was not consistant sometime tripped only on one rcd other times took both out. Yes you are correct time delayed 100ma sorry took that for granted.

 

[Octopus]

Back to the OP, if the circuit serving the outbuilding is protected by a 30mA RCD, why bother with a 2nd RCD?

 

[Edd]

I've had discrimination with two 30ma RDC's where neither tripped on a fault, so i would not take the chance.

Im really intrested in this, anyone got a technical reason or idea why this could happen?

 

[SPARTYKUS]

Im really intrested in this, anyone got a technical reason or idea why this could happen?

Faulty RCDs? Cannot see any other reason theres no magic secret with them as far as Im aware. Needasparks what do you mean you had discrim. with two 30mA RCD's?

 

[Deleted member 26818]

The fault didn't produce a high enough fault current.

 

[SPARTYKUS]

Ps do you mean a time delayed 30mA device? You can get them, the problem them being not conforming on x5 test....

 

[Guest123]

You would need a time delayed RCD as I understand it from reading posts on the forum?

Correct.

There would be very little discrimination between a 100mA RCD & a 30mA one unless the 100mA RCD was an 'S' type one.

 

[needasparks]

Faulty RCDs? Cannot see any other reason theres no magic secret with them as far as Im aware. Needasparks what do you mean you had discrim. with two 30mA RCD's?

Could possibly have been a faulty rcd and i by chance took the bad one out of circuit of the two but neither would trip at any x1 or x5, at the time i was doing a job and i did not know there was an upstream one fitted, did job then carried out test and it came to light. Have had this twice before, one was an old farm years ago and another was when someone had an rcd socket on an rcd protected ring, In all cases there was problems with as far as i see it discrimination between the rcd's of the same size. Perhaps it was just my luck or in all cases one rcd was faulty. don't really know and never had a reason to test the one i took out as all worked fine when only one in line. I know about the size/delay difference, same size lottery and assumed this was another thing to add.
either way i never place two of the same size in series from problems i've experienced in the past. obviously thats just my opinion but in this game we have to stand by what we know and trust, I'm sure we all do and that's what makes the customer call us back to do more work when it does what you say it will.

 

[Guest123]

Ps do you mean a time delayed 30mA device? You can get them

That's a new one on me, why would any manufacturer install a time delay to a device whose primary function is to save you from a shock, or worse. Do you have any links to these devices????

 

[SPARTYKUS]

Lenny, in answer to your question,



I know not!

Will find link to it!

Please hold caller....

 

[SPARTYKUS]

http://www.protekuk.co.uk/rcd-data-sheet.pdf


Second page

 

[SPARTYKUS]

Must be an application for it somewhere, not in domestics though I doubt.

 

[La Poste]

Back to the OP, if the circuit serving the outbuilding is protected by a 30mA RCD, why bother with a 2nd RCD?

Ever since I went to my sisters place and did some tests and found her 30 ma rcd welded shut I have always had doubts
about relying solely on 1 rcd, especially in a garden, at least with my sister she had a low enough EFLI to trip an mcb in the event of an earth fault, in a garden
that is TT'd you would not have that luxury, the circuit would simply not switch off in the event of an earth fault and would remain live until someone touched something metal.

This garden already has an existing underground cable but somewhere along the line someone has buried an underground joint to extend this cable and the readings I get from testing are very erratic.

The Line to neutral IR resistance jumps from being at 0.2 M Ohms up to about 2.5 M Ohms at 500 V and sort of flitters about between these two extremes and won't settle.
The existing underground cable is a two core roughly 6mm but in reality is an older multistrand cable with the armour used as the cpc.
The armour has broken down and I get an end to end resistance reading of 12 Ohms on the armour, the cable is only 15 metres long so it really does look like this underground joint has got water contamination and the cable should be abandoned IMHO.

There are three small buildings in the garden, one greenhouse, one shed and a sort of chalet, he wants power to all three.

The greenhouse will be damp with a floor that is basically real earth IE the garden soil.
The shed will also be damp but has a wooden floor leading out onto decking.
The chalet seems quite dry with a wooden floor also.

I'm thinking the best way to do this would be to run a new cable which is going to be really hard but I never thought of using trunking to conceal this cable from the fuse board to the outside of the house. This is a good idea as long as I can get some nice skirting trunking maybe, I will have a look today so thanks for the suggestion.
Once the cable is outside the house I can then surface clip or bury it around the perimeter of the garden down to the end.

At the moment I think the best method would be to run a new supply in trunking (not easy) protected by a 100 ma S type.
This would go to a central distribution point in the garden somewhere.
From there I could run a supply to the greenhouse and rod it there.
From the central point I could also run a supply to the shed and rod there also.
Also run a supply to the chalet and again rod it.
This would give me three rods in the garden at different places which should give me a better overall Ra reading.
The main house supply cable is under ground TN-S with a lead sheath that is terminated at the MET in the house so this should also help lower my Ra readings (the lead sheath of the supply cable runs underground through the garden)
Each garden building would have its own small CU and each small CU would be protected by a 30 ma RCD.

The other way I was thinking would be to bond everything and export the main bonding from the house but I would have to put a ground mat down in the greenhouse
to bond the floor to the frame but if I managed to extend the bonding from the house at least I could use the low EFLI of the supply circuit IE not TT anything so that if an rcd did fail in the event of an earth fault I would have the added protection of an MCB tripping at the main fuse board in the house.
Bonding everything in this damp garden enviroment seems a bit hit and miss to me especially with the damp and possibly kids running around in bare feet so this idea gives me the willies a bit, I mean how can I make the whole garden an extension of the house equipotential zone especially when he wants a couple of outside sockets installed?

I am going to do a bit more testing today.

As far as part P goes I would like to wait until after my assessment but if the guy really wants me to get a shift on I will have to notify LABC before I start and pay the fees.

Any thoughts on this one?

Thanks for taking the time to read this lengthy post.

 

[Engineer54]

Could possibly have been a faulty rcd and i by chance took the bad one out of circuit of the two but neither would trip at any x1 or x5, at the time i was doing a job and i did not know there was an upstream one fitted, did job then carried out test and it came to light. Have had this twice before, one was an old farm years ago and another was when someone had an rcd socket on an rcd protected ring, In all cases there was problems with as far as i see it discrimination between the rcd's of the same size. Perhaps it was just my luck or in all cases one rcd was faulty. don't really know and never had a reason to test the one i took out as all worked fine when only one in line. I know about the size/delay difference, same size lottery and assumed this was another thing to add.
either way i never place two of the same size in series from problems i've experienced in the past. obviously thats just my opinion but in this game we have to stand by what we know and trust, I'm sure we all do and that's what makes the customer call us back to do more work when it does what you say it will.

It certainly can and does happen NS... RCD devices are not exactly fail proof or fool proof, and none can be matched, so basically anything can happen when you have 2 RCD devices of the same protection value in series.

It's one thing to say what ''Should'' happen, however it's another when it doesn't happen!! with serried RCD devices !! One or both will trip maybe 99 times out of a hundred, then you get that one time it doesn't!!...lol!! Seems as if you have seen this happen, on more than one occasion too!!!

I keep saying this, ...RCD devices are not as reliable as most will have you believe, and should never be used as a cure all solution, to get you out of trouble. A good reliable earthing system to start with, should always be the basis of any installations ADS...

 

[EastAnglian]

Ever since I went to my sisters place and did some tests and found her 30 ma rcd welded shut I have always had doubts
about relying solely on 1 rcd, especially in a garden, at least with my sister she had a low enough EFLI to trip an mcb in the event of an earth fault, in a garden
that is TT'd you would not have that luxury, the circuit would simply not switch off in the event of an earth fault and would remain live until someone touched something metal..

(1) Lets get a few things right ......

The statutory notice requiring RCDs to be tested frequently (quarterly at max) is there for a reason, its not just wall paper. The RCD welded shut for a reason I assume, most likely a big pop one day! At that point if the person who "reset" it didn't test it with the test button, or get somebody to look at it they were at fault.

(2) A human (about 500 ohms per limb) touching a live conductive part will NEVER EVER draw enough current to break a 3A fuse, let alone a 13A fuse or 20A circuit breaker. At around 100mA current through the body your muscles will start to fibrilate, and if its your heart muscles which are thus affected then its "Goodbye England, look for me on the other side".

This thread has loads of good advice about how to comply with the Regulations for this installation and I really do recommend you find a way to implement them. 100mA time-delayed RCD upstream and 30mA RCD downstream is the RIGHT way to do it if you are not content with the 30mA RCD on the ring final protecting your (fused?) spur to the garden.

Sure thing, the Regs are not statutory, so you could do lots of tests & calculations and come up with a design which you as designer believe is as safe or better than the "standard" way - but be prepared to stand in a Coroners Court and justify your calculations to a jury because its BS7671:2008 (2011) which the expert witness will measure you against!

I also humbly suggest that you don't offer this install to your Part P assessor unless you can back-off the design to the regs!

 

[SPARTYKUS]

Ever since I went to my sisters place and did some tests and found her 30 ma rcd welded shut I have always had doubts
about relying solely on 1 rcd, especially in a garden, at least with my sister she had a low enough EFLI to trip an mcb in the event of an earth fault, in a garden
that is TT'd you would not have that luxury, the circuit would simply not switch off in the event of an earth fault and would remain live until someone touched something metal.

This garden already has an existing underground cable but somewhere along the line someone has buried an underground joint to extend this cable and the readings I get from testing are very erratic.

The Line to neutral IR resistance jumps from being at 0.2 M Ohms up to about 2.5 M Ohms at 500 V and sort of flitters about between these two extremes and won't settle.
The existing underground cable is a two core roughly 6mm but in reality is an older multistrand cable with the armour used as the cpc.
The armour has broken down and I get an end to end resistance reading of 12 Ohms on the armour, the cable is only 15 metres long so it really does look like this underground joint has got water contamination and the cable should be abandoned IMHO.

There are three small buildings in the garden, one greenhouse, one shed and a sort of chalet, he wants power to all three.

The greenhouse will be damp with a floor that is basically real earth IE the garden soil.
The shed will also be damp but has a wooden floor leading out onto decking.
The chalet seems quite dry with a wooden floor also.

I'm thinking the best way to do this would be to run a new cable which is going to be really hard but I never thought of using trunking to conceal this cable from the fuse board to the outside of the house. This is a good idea as long as I can get some nice skirting trunking maybe, I will have a look today so thanks for the suggestion.
Once the cable is outside the house I can then surface clip or bury it around the perimeter of the garden down to the end.

At the moment I think the best method would be to run a new supply in trunking (not easy) protected by a 100 ma S type.
This would go to a central distribution point in the garden somewhere.
From there I could run a supply to the greenhouse and rod it there.
From the central point I could also run a supply to the shed and rod there also.
Also run a supply to the chalet and again rod it.
This would give me three rods in the garden at different places which should give me a better overall Ra reading.
The main house supply cable is under ground TN-S with a lead sheath that is terminated at the MET in the house so this should also help lower my Ra readings (the lead sheath of the supply cable runs underground through the garden)
Each garden building would have its own small CU and each small CU would be protected by a 30 ma RCD.

The other way I was thinking would be to bond everything and export the main bonding from the house but I would have to put a ground mat down in the greenhouse
to bond the floor to the frame but if I managed to extend the bonding from the house at least I could use the low EFLI of the supply circuit IE not TT anything so that if an rcd did fail in the event of an earth fault I would have the added protection of an MCB tripping at the main fuse board in the house.
Bonding everything in this damp garden enviroment seems a bit hit and miss to me especially with the damp and possibly kids running around in bare feet so this idea gives me the willies a bit, I mean how can I make the whole garden an extension of the house equipotential zone especially when he wants a couple of outside sockets installed?

I am going to do a bit more testing today.

As far as part P goes I would like to wait until after my assessment but if the guy really wants me to get a shift on I will have to notify LABC before I start and pay the fees.

Any thoughts on this one?

Thanks for taking the time to read this lengthy post.

Hi La Poste, great post, great to read someones thought processes, very interesting problem. Chewy I'd say. And they're the best aren't they! (chewy ones)

Can you not preserve the TNS earthing system (you know the one, the one that was used until the 70's when all of sudden the laws of physics changed and it WAS after all, safe to use your Neutral as an earth!) and simply use that, run a length of 4 10-or-more mm2 core armoured? Then you're best practise not relying on the armour for your earth (but will be using it as well anyway), you have a separate core should you wish to run an extra bond back to the MET in the house, and ...no type S RCD up front needed. Assuming a good Ze of course.


Just an idea. And if armoured probably no need to trunk (except aesthetics), as long as you use cleats and not tape :drool5: