Circuit Design & RCCD's

[Pensdown]

I think this topic deserves it's own thread.

Unless your concealed cables have been installed in accordance with one of the five options (very unlikely) you will need to install 30mA earth leakage protection to every circuit that has concealed cables. So a 100mA type S as a mains switch is unsuitable. Also, it will not meet 314.1

Are you suggesting that a front end device is not required.

If the CU is Class II,(531.4.1) the sub-circuits are protected by DP RCBO's and the bus bars are insulated - Yes

The review is taking place because of some problems that have been identified concerning shock protection between the outgoing terminals of a main switch and the supply side terminals of an RCD downstream of a non-RCD main switch.

That should be a short meeting - safe isolation? Shock protection has nothing to do with 100mA type S RCCD's

This is an old chestnut that is similar to the reason for the requirements in 531.4.1.

I'm not sure what 531.4.1 has to do with this because I never mentioned a single RCCD.

314 is, as ever, not really relevant and generally unworkable - it is far more important that you achieve shock protection.

314 is very relevant. It's the minimum requirement for circuit design and it applies to every installation.

I would advise anyone that is considering omitting a front end device on a TT system to confirm with the manufacturer in writing that satisfactory shock protection will be achieved.

I'm missing how a 100mA RCCD provides shock protection?

Note that if a fault occurs between the outgoing circuit from a non-RCD main switch that is 'behind' any RCD the extraneous-conductive-parts and the exposed-conductive-parts in the installation may attain full supply potential and may not be disconnected.

531.4.1 - Class II CU's solve that problem.

 

[Adam_151]

I think that perhaps what nothimagain is alluding to is on a TT system where you have a board of RCBOs and no upfront RCD, you'll have a non-rcd protected lump of copper behind a flipsy plastic cover, in an enclosue that also contains cpcs that have a high impedance to earth, and possibly the only standing between you and an earth fault that won't disconnect, but will bring all the worms out the ground in a 1m circle around the electrode is a bit of green and yellow sleeving.

I do think your interpretation of 314 is a bit strict, I see no reason why every single circuit in a house should need an RCD, certainly I don't see that say loosing the immersion heater until the RCD is reset because the cooker has gone faulty neither presents a hazard, nor an inconvience that is measureable, lighting and sockets are perhaps a different story, but if we are going to install lighting on their own RCBOs, we ought to make sure they arn't the B6 ones, otherwise you arn't really creating much of an inprovement... perhaps a couple of non maintained EM fittings might be a better purchase than RCBOS As for sockets, if we took the inconvience angle to extreme, we would be installing individual RCD socket outlets, and wiring them with singles in galv conduit... perhaps simply having the upstairs and downstairs ones on separate RCDs is sufficient?

 

[NotHimAgain]

Distribution boards are not class II unless specifically indicated by the manufacturer.

The original concern was with metal distribution boards with an RCD as the Main switch - the problem was that a fault between the incoming tails and the metal casing could result in a fault situation similar to the one considered here. A solution was found by providing an insulating kit for the incoming tails which gave a protection similar to class II but it was not actually classified as such.

The problem now identified is somewhat similar and if manufacturers would indicate that their distribution boards were in fact class II it could be solved - BUT - that is not an easy thing to do given the construction of the distribution boards. An alternative approach is to consider 'fault free zones' - areas within the distribution board in which the risk is negligible.

However, none of these things has yet been done by all manufacturers - hence my advice to consult.

With regard to shock protection - are you serious - a 100 mA S type RCD is perfectly adequate to provide the required 1 second disconnection time on this distribution circuit. It is far superior to relying on a cut-out fuse operating with a Zs of several ohms.

The important thing here is electrical safety - if you cannot see the problem - well that is your problem - I can assure you the industry is taking it seriously.

 

[Pensdown]

Hi Adam, good to see you on line

I think that perhaps what nothimagain is alluding to is on a TT system where you have a board of RCBOs and no upfront RCD, you'll have a non-rcd protected lump of copper behind a flipsy plastic cover, in an enclosue that also contains cpcs that have a high impedance to earth, and possibly the only standing between you and an earth fault that won't disconnect, but will bring all the worms out the ground in a 1m circle around the electrode is a bit of green and yellow sleeving.

Valid point but as long as the CU is selected correctly I really can't see a problem. Better makes have fully insulated bus bars.

I do think your interpretation of 314 is a bit strict, I see no reason why every single circuit in a house should need an RCD, certainly I don't see that say loosing the immersion heater until the RCD is reset because the cooker has gone faulty neither presents a hazard, nor an inconvience that is measureable, lighting and sockets are perhaps a different story, but if we are going to install lighting on their own RCBOs, we ought to make sure they arn't the B6 ones, otherwise you arn't really creating much of an inprovement...

As you know, 314 applies to circuit arrangements in general and as we move forward the number of circuits in the average house has grown even though the connected loads have not really changed that much.

Designers have decided it would be better to separate the up and down lighting even though in most cases one 6amp or in extreme cases a 10amp circuit would do the trick.

They have also decided to separate the up and down RF when in most properties there could be one circuit with a separate RF or radial to the Kitchen.

And then the same designers start grouping circuits together so they trip under earth fault conditions. Why? In terms of circuit design it’s going backwards?

perhaps a couple of non maintained EM fittings might be a better purchase than RCBOS

I've got a few EM's but they're not cheap

As for sockets, if we took the inconvience angle to extreme, we would be installing individual RCD socket outlets, and wiring them with singles in galv conduit...

Now you're taking the ****

perhaps simply having the upstairs and downstairs ones on separate RCDs is sufficient?

I can't get my head around any design where a fault on one circuit is guaranteed to take out another. It's poor. Should we ignore discrimination as well?

 

[Pensdown]

I think we are in agreement that as long as the CU meets the requirements of class II then the 100mA TD RCCD is not required?

With regard to shock protection - are you serious - a 100 mA S type RCD is perfectly adequate to provide the required 1 second disconnection time on this distribution circuit. It is far superior to relying on a cut-out fuse operating with a Zs of several ohms..

A 30mA RCCD provides "shock" protection so we're talking at cross purposes.

 

[NotHimAgain]

Pensdown wrote

I can't get my head around any design where a fault on one circuit is guaranteed to take out another. It's poor. Should we ignore discrimination as well?

I don't want to open a full discussion on 314 - you have your views and I have mine - and our EU partners have theirs, and they don't see any problem with 'front end' RCDs.

However, on the above quote - discrimination is a very complex area. It includes discrimination for overcurrent protection - both fault and overload and these often have to be assessed separately. It also includes series connected RCDs, and we have issues here concerning operating time envelopes and disconnection of all live conductors (including the neutral).

Discrimination cannot always be achieved for all aspects of circuit protection.

Next shock protection - we now have basic protection (direct contact) and fault protection (indirect contact). We also have Additional Protection (415), as a sub group within fault protection.

Basic protection does not concern us here - so looking at fault protection, we have section 411. Now without labouring the point too much, just look at: 411.4.4; 411.5.2 and (if you must ) 411.6.3. All of these list RCDs as a means of fault protection and none of them specify that the operating current (I delta n) has to be 30mA.

In fact the only time this is specifically specified is in relation to Additional Protection.

Now that is interesting because the 16th Edition requirement to provide a 30 mA RCD for socket outlets that might be used for equipment outside (471-16-01) was for direct contact protection (basic protection) - but the new requirement for socket outlets, that effectively does some of the same job, is for fault protection (411.3.3) - what tricky lives we lead .

So the point is (finally) any RCD may be considered as a device for fault protection (i.e. shock) - it just depends on the required disconnection time.

Next the Distribution Board - Class II issue. This is really at the crux of it. The Electrical Safety Council recently published some guidance that appear to suggest that a 'front end' RCD is not required in some domestic TT installations - in other words largely in agreement with what you have stated. Now some public spirited chap challenged this and part of that challenge was - have all manufacturers confirmed that their distribution boards meet the requirements for Class II in the area of concern (you see - we are way ahead of you ) - answer NO, in fact none have.

This prompted the review - there are other concerns but that is enough for now.

 

[Pensdown]

I don't want to open a full discussion on 314 - you have your views and I have mine - and our EU partners have theirs, and they don't see any problem with 'front end' RCDs. .

I do have a problem with the way front end RCD's are used. Take the discussion that GaryMo reffered to in his post Thread on IET forums

The poster was advised to fit a 100mA type S at the origin and single pole RCBO's.

If an N-E fault developed on one of the circuits, with single pole RCDO's there is no way the customer could isolate the fault. They would be left without any power whatsoever until the fault had been found and cleared. Depending on when it happened that could be quite some time.

No lights, probably no heating, no fridge or freezer and in some cases no land line phone. That’s more than just inconvenient and it’s totally unnecessary because it can be designed out.

Most of the internet discussions that I’ve read on this subject assume that if an RCCD trips, it will just simply re-set. But we all know that quite often that’s not the case.

 

[NotHimAgain]

First let us look at the anecdotal evidence - there are 10s of 1000s of installations in the UK that have 30 mA RCDs (not S type) at the front end, and yet there are no statistics (that I know of) that suggest 1000s of Grannies are falling down the stairs in their bungalows and flats .

Now I can't prove any of that - the figures don't exist - but I am reasonably confident that my assertion is accurate.

Add to this the even greater number of installations in Continental Europe - particularly in the southern countries - that use 30 mA front end RCD protection.

Now no one is advocating the continued use of 30 mA front end RCDs. The 100 mA S Type or even the 300 mA S type is fine for most situations. I contend that these coupled with appropriate downstream protection provide a reasonably reliable supply protection scheme for a domestic TT systems - I will address neutral to earth faults later.

Note that the electrical installation for the JET project (fusion reactor test bed) in Oxford was designed and installed by the Italians using their standards - it has given over twenty years of good service and it contains many RCDs in series arrangements (NOT SINGLE POLE RCBOs). This installation has such a high demand that it takes its supply direct from the super grid at 400kV - it is the only one in the UK that does that.

So IMO front end RCDs do not present a significant problem.

If we omit the front end device we create a need to provide greater security within the distribution board to provide shock (fault) protection. We have both mentioned class II as one method that might be employed - but that is far from a simple matter. It could be solved by specific design, but this may restrict the flexibility that many seem to want in board layout. Other solutions may exist, for example, 'fault free zones' within the distribution board. IMO These solutions should be design, implemented and verified by the manufacturer.

It may well be that the Electrical Safety Council discussion will trigger action by the manufacturers to implement one of these solutions. If they do one of the problems is solved.

Neutral to Earth faults - the most common defect in electrical installations - why? - because they are not detected by the bang test .

Do they matter - yes they cause load currents to flow in any parallel earth path, and this could have serious consequences. They effectively turn TN-C-S systems into TN-C systems. In most cases the use of a TN-C system is actually illegal. In any event they are faults and they should not be allowed to persist.

Given all of that we should seek to find them - not hide them. So the fact that a front end device might respond to a neutral to earth fault is not, in itself, a problem. However, we should seek to minimise power supply interruptions whenever possible. You have indicated that a single pole RCBO will detect but not isolate such a fault, and this will result in the front end device operating.

I could take you to a relatively large Sports Centre in the South West that was wired by one of the largest contractors in the country. They have produced exactly this effect on a grand scale. A neutral to earth fault on any final circuit brings out the main MCCB via its residual current sensing system. The whole site comes to a crash stop and has to be evacuated - nice .

This problem has an easy solution - you simply have to ensure that all RCDs open all live conductors (including the neutral). Now we come to single pole RCBOs - I cannot find any trace of them on suppliers sites that serve our EU buddies so this could well be yet another UK only self inflicted problem.

Note that for TT, any device you call an isolator must open all live conductors - so if any of your RCBOs is providing an isolation function it must open all live conductors. Now I can find plenty of double pole RCBOs (generally SP + neutral switching) but they all require two slots. I don't know of a single slot, double pole device - do you? I have found some MCBs but no RCBOs.

And finally - if your TT system has been form from a TN-C-S supply (as you might do for a shed or outside hot tub) a neutral to earth fault that is not disconnected from the supply will connect your TT system to the supply neutral via the fault.

Some manufacturers (MK etc) make consumer units that can have as many as three split bus-bars, each controlled by a double pole RCCB and you can also use a front end S Type RCCB - surely this could provide sufficient supply security for your Granny .

 

[sm1thson]

The poster was advised to fit a 100mA type S at the origin and single pole RCBO's.

If an N-E fault developed on one of the circuits, with single pole RCDO's there is no way the customer could isolate the fault. They would be left without any power whatsoever until the fault had been found and cleared. Depending on when it happened that could be quite some time.

No lights, probably no heating, no fridge or freezer and in some cases no land line phone. That’s more than just inconvenient and it’s totally unnecessary because it can be designed out.

wasnt this scenario the same under typical 16th edition TT setups? a (single pole) MCB wouldnt isolate the neutral either, so an NE fault on a lighting circuit protected by an MCB and the incomer RCD would leave the house with no power!?! did people use to design to over come this (i ask in seriousness I am not a spark, what i have seen of TT instalations suggests as i have written (but i am willing to be wrong))

 

[Pensdown]

First let us look at the anecdotal evidence - there are 10s of 1000s of installations in the UK that have 30 mA RCDs (not S type) at the front end, and yet there are no statistics (that I know of) that suggest 1000s of Grannies are falling down the stairs in their bungalows and flats .

But if one did I wouldn't like to be the designer, given all the other options that are available.

Add to this the even greater number of installations in Continental Europe - particularly in the southern countries - that use 30 mA front end RCD protection.

I hope that your not suggesting we adopt their practices

Now no one is advocating the continued use of 30 mA front end RCDs. The 100 mA S Type or even the 300 mA S type is fine for most situations. I contend that these coupled with appropriate downstream protection provide a reasonably reliable supply protection scheme for a domestic TT systems - I will address neutral to earth faults later.

We are both singing from the same sheet. It's the down stream protection that is being ignored which then causes the up stream protection to trip unnecessarily.

I could take you to a relatively large Sports Centre in the South West that was wired by one of the largest contractors in the country. They have produced exactly this effect on a grand scale. A neutral to earth fault on any final circuit brings out the main MCCB via its residual current sensing system. The whole site comes to a crash stop and has to be evacuated - nice .

I hope the installation is still in the defects period

This problem has an easy solution - you simply have to ensure that all RCDs open all live conductors (including the neutral). Now we come to single pole RCBOs - I cannot find any trace of them on suppliers sites that serve our EU buddies so this could well be yet another UK only self inflicted problem.

I couldn't agree more. Single pole RCBO's are OK for TN installation but IMO they are totally unsuitable for TT installations. I can see why designers want to use up stream RCCD's but they should then install down stream protection that fully discriminates with it, which must include N-E faults.

Note that for TT, any device you call an isolator must open all live conductors - so if any of your RCBOs is providing an isolation function it must open all live conductors. Now I can find plenty of double pole RCBOs (generally SP + neutral switching) but they all require two slots. I don't know of a single slot, double pole device - do you? I have found some MCBs but no RCBOs.

No I don't, but why is a twin module a problem?

Some manufacturers (MK etc) make consumer units that can have as many as three split bus-bars, each controlled by a double pole RCCB and you can also use a front end S Type RCCB - surely this could provide sufficient supply security for your Granny .

That's a step in the right direction and it would be better (given the current availability of other products) than most TT installations that I've seen.

PS..I don't normally like replying in this way but you do raise a lot of points in one post.

sm1thson, Yes your right. It's an age old problem that started in the good old days of Chilton ELCB'S.....or maybe before...