What type RCD for energy saving lights?

[Risteard]

You won't find a Type B RCBO in terms of RCD characteristics. I've only ever been able to get Type B RCCBs.

If Type AC were banned for new installations then you wouldn't have had the issue whereby they only manufacturer Type AC.

The exceedingly poor brand MK only made Type ACs and thankfully have been put out of production.

 

[JohnW2]

Thank you @flameport that really did help, so next question is if for example I have a faulty USB socket which stops the RCD tripping ...

What sort of 'fault' are you envisaging. I though we were talking about a potential issue when some equipment result in appreciable dc current comments when 'in normal use'.

... is only that RCD affected or does it stop all 14 RCBO's from working?

I can't see how anything happening on a totally different circuit could affect how an RCBO behaved, can you?

Fact that test button is affected means pushing test button should highlight fault, although as pointed out test buttons often have higher than a 30 mA load so not really the best test, but one hopes one will find fault before the RCD is required.

I'm not sure what you are saying here. As I said before, I suspect (but do not know for sure) that, even for a non-AC-type RCD/RCBO, the test button does more than create an AC current imbalance (with no DC component).

Kind Regards, John

 

[DetlefSchmitz]

I'm not sure what you are saying here. As I said before, I suspect (but do not know for sure) that, even for a non-AC-type RCD/RCBO, the test button does more than create an AC current imbalance (with no DC component).

I think he's saying that if there is a dc content (desensitising the RCD) caused by specific equipment at the time the test button is pressed, then the test button may also be prevented from operating the desensitised RCD.

 

[JohnW2]

I think he's saying that if there is a dc content (desensitising the RCD) caused by specific equipment at the time the test button is pressed, then the test button may also be prevented from operating the desensitised RCD.

Yes, maybe - and if that's the case (which may well be what he way getting at), at least it is then giving the 'right answer' (that the RCD is not tripping when it should, maybe because of a DC component in what is going through the RCD).

I must say that it has always been my practice to minimise as far as I can the number of active loads on circuits it's protecting when I press the test button of an RCD - so maybe I 'miss out' on discovering that some connected equipment may (normally) be impairing the ability of an RCD to work as intended.

As I've said, I really don't yet understand much of this. I originally though that, as discussed by flameport, we were talking about DC components saturating the core of the sense coil and thereby impairing the RCD's ability to respond to an imbalance of AC components of the current. However, as I've said, BS7671 seems to now major on talking about the ability of RCDs other than Type A being able to trip in response to imbalances of pulsatile etc. currents, with or without DC components being present - so I really don't know.

As I've also said, nor have I get found out what the RCDs other than Type AC do to get around these apparent issues - do you have any knowledge of that?

Kind Regards, John

 

[DetlefSchmitz]

I must say that it has always been my practice to minimise as far as I can the number of active loads on circuits it's protecting when I press the test button of an RCD -

Probably a sensible precaution, that the reason I rarely test them, too many critical loads connected.

do you have any knowledge of that?

Sorry, never looked into it. What also seems to be lacking is quantifiable facts. If this is as big a problem as people are making out, you might wonder how we got into this incredible mess.

 

[JohnW2]

Probably a sensible precaution, that the reason I rarely test them, too many critical loads connected.

Maybe sensible but, as I said, perhaps with the consequence of failing to detect situations in which, under normal circumastances, connected equipment was impairing the ability of RCDs to function correctly. I used to think similarly to you, but it then occurred to me that '5 second (or whatever) power cuts' are so relatively common around here that it was rather silly of me to worry about the effect on 'critical loads' of my pressing test buttons one every few months, or whatever!

Sorry, never looked into it. What also seems to be lacking is quantifiable facts.

Very much so. What I have certainly never seen is any indication of the amount of DC component and/or the degree of 'non-sinusoidality' (good word ) that a Type AC can tolerate whilst still working within spec and/or the amount of DC component and/or the degree of 'non-sinusoidality' that a non-Type-AC can tolerate and still work 'as intended' (whatever that may be).

If this is as big a problem as people are making out, you might wonder how we got into this incredible mess.

Indeed. If it really is "as big a problem as people are making out", then I presume that Type AC ones would not have come into being in the first place had it been an issue at the time and we had had the means of circumventing the problems. However, those decades ago when RCDs first came into being, the 'connected equipment' was mainly 'much more straightforward' (hence much less of the messy waveforms and/or potential DC components), and we may not even have had the technological means at the time of addressing the issue, even if we had wanted/needed to.

Kind Regards, John

 

[ericmark]

I must say that it has always been my practice to minimise as far as I can the number of active loads on circuits it's protecting when I press the test button of an RCD - so maybe I 'miss out' on discovering that some connected equipment may (normally) be impairing the ability of an RCD to work as intended.

That has also been what I have done in the past, however in view of the video that @flameport posted it would seem that is likely wrong, and one should also test with items running when testing a type AC RCD.

However in that video they had a three phase RCD with DC being fed through one phase and normal AC through the other phase, can't really see how that would happen in real life, only way to get the DC to flow through the RCD would be to earth. Well he injected 250 mA of DC and it tripped with 195 mA so if going to earth rather than a simple loop would it trip it? It is clear the DC is a problem, but not clear is where in the home the DC is coming from.

So in my case if I have some thing swamping the socket supply for front of house, will that affect all the other socket supplies, I have 14 RCBO's so 14 independent circuits, so if the other 13 are OK then an issue yes, but not a big issue, it is only likely to affect 10% of the installation, if I had just two RCD's for whole of house then a bigger issue, but also if just two RCD's it would be a lot cheaper to cure.

So I have told people on this forum many times, there is no point having a 30 mA RCD feeding a 30 mA RCD when taking a supply to garden or outbuilding, however it seems that advice is wrong?

 

[JohnW2]

That has also been what I have done in the past, however in view of the video that @flameport posted it would seem that is likely wrong, and one should also test with items running when testing a type AC RCD.

Maybe - but at best that would be rather 'hit and miss', since you couldn't possibly have every conceivable load 'running' at the time you pressed the test button.

So in my case if I have some thing swamping the socket supply for front of house, will that affect all the other socket supplies, I have 14 RCBO's so 14 independent circuits ...

Maybe I'm missing something but, as I said before, I cannot see how the current flowing through one RCD or RCBO could conceivably effect the performance/behaviour of any other one.

So I have told people on this forum many times, there is no point having a 30 mA RCD feeding a 30 mA RCD when taking a supply to garden or outbuilding, however it seems that advice is wrong?

I'm not sure what you are saying/suggesting here. If they are both 30mA RCDs of the same type, then I don't see what benefit (other than redundancy) could be gained by having two in series.

Kind Regards, John

 

[ericmark]

I'm not sure what you are saying/suggesting here. If they are both 30mA RCDs of the same type, then I don't see what benefit (other than redundancy) could be gained by having two in series.

If a RCD type AC is being swamped by say a washing machine inverter drive so it will not trip a second RCD even is also type AC feeding from that circuit to a hedge cutter is not likely to be swamped as well as only supply to hedge cutter going through that RCD, so that one would still trip.

What of course is required is a risk assessment, what are the chances of a RCD failing to operate because of the DC component? In the main where a RCD has tripped it is either water or a component failure, like a capacitor used in interference suppression. In other words the RCD trips before a human touches a live part, not as he touches a live part, although the latter is of course possible.

It seems the inverter drive, USB socket, switch mode power supply or other item will not stop the RCD working under normal conditions, only when the item goes faulty, I assume a DC path to earth, but not clear on what fault causes DC to run through the RCD iron core, so the question is how likely are these faults likely to happen at the same time as some one touches a live wire? i.e. what is the risk.

We have identified bathrooms and gardens are the high risk areas, the bathroom shower has its own RCBO so that it seems will still protect, so only item in the bathroom is the lights, garden at rear of my house I would use power from the flat, which is really only a store room, so that RCBO will still work, so the only one for power is front of house so if I fit an outside socket with its own RCD for front of house that is covered.

So only remaining problem area is the lights, but there are no earths to most bulbs, so how would there be a DC through the RCD? I don't know which is why I am asking. Now the lights is the odd one out, I can buy B6 type A RCBO's so it would seem three RCBO's type A for lights and a outside socket with RCD built in for front of house. However what is the risk of getting a shock from lights? Is it really worth it, and would it just be simpler not to use electric outside at same time as running washing machine?

So what is the chance of a DC swamping at the same time as also getting a shock?

 

[JohnW2]

If a RCD type AC is being swamped by say a washing machine inverter drive so it will not trip a second RCD even is also type AC feeding from that circuit to a hedge cutter ...

Oh, I see. I thought you literally meant having two RCDs in series (hence both take the circuit's full currents), rather than one being in a 'branch' off the circuit protected by the first one. So, yes, I agree with you, at least in theory.

What of course is required is a risk assessment, what are the chances of a RCD failing to operate because of the DC component?

As Detlef said, that's something about which we have very little information. I haven't a clue, but I suspect that, despite what is being said, in practice the chances are very small. Apart from anything else, if it were common for DC components due to some common loads preventing RCDs from tripping in response to test buttons being pressed, I suspect we would be hearing about it!

Since I would have thought that it should be so simple, and cheap, to avoid it (e.g. by the inclusion of brisdge rectifiers, where appropriate), I can't really see why there should be much need to create equipment which results in DC components in the supply current - even cheapo LED lamps usually have bridge rectifiers.

Nasty (but symmetrical about zero - hence no 'DC component') waveforms are undoubtedly much more common (than 'DC components') but another thing I don't know is the extent to which they will impair RCD functionality.

As Detlef said, there appears to be a paucity of (at least, easily found) information.

Kind Regards, John

 

[flameport]

if it were common for DC components due to some common loads preventing RCDs from tripping in response to test buttons being pressed, I suspect we would be hearing about it!

Most people never test RCDs, so the majority of those RCDs which don't work in certain circumstances will never be noticed by anyone.

which it seems only makes type A RCBO as 6 amp,

All ratings are available asType A, here: https://www.consumerunitworld.co.uk/fusebox-type-a---b-curve-rcbos-409-c.asp
Slightly more expensive that the AC: https://www.consumerunitworld.co.uk/fusebox-type-ac---b-curve-rcbos-408-c.asp

However in that video they had a three phase RCD with DC being fed through one phase and normal AC through the other phase, can't really see how that would happen in real life,

It won't - that test rig was built to demonstrate the Blakley ELS device in the centre, and having a 3 phase RCD with a single phase supply and separate DC supply was the simplest method to show how DC can prevent a Type AC RCD from working.

Maybe - but at best that would be rather 'hit and miss', since you couldn't possibly have every conceivable load 'running' at the time you pressed the test button.

The concept of testing with loads connected is valid, that originates from commercial installations where the usual methods in the past were to do all testing out of hours with the facility completely shut down, and unsurprisingly the RCDs all worked perfectly every time. When the same testing was done with the facility energised during the day, none of the RCDs worked.

 

[JohnW2]

Most people never test RCDs, so the majority of those RCDs which don't work in certain circumstances will never be noticed by anyone.

That's obviously true up to a point, but some do test them, so if it were a common issue, I would have expected us to at least occasionally hear reports that they've pressed the test button and nothing happened - but that seems very rare.

Of course, if they did report that, they would probably be told to get the RCD 'tested properly' (which might show no problem) and/or to have the RCD replaced (which might not solve the problem)!

The concept of testing with loads connected is valid, that originates from commercial installations where the usual methods in the past were to do all testing out of hours with the facility completely shut down, and unsurprisingly the RCDs all worked perfectly every time. When the same testing was done with the facility energised during the day, none of the RCDs worked.

That makes sense, certainly in a commercial or industrial environment, where the 'energised facility' would presumably often/usually consist of a fairly consistent set of loads. However, as I implied, in a domestic environment, in which anything could be 'plugged in', any day, it is much more difficult to test with 'all the relevant potential loads' energised.

There still seems to be a lot more I need to try to learn about this business, since I doubt that the core saturation you previously discussed can be anything like the whole story - and, as I've said, I've also yet to discover how the Type A (and B/F) RCDs deal, mechanistically, with these issues - can you point me towards any useful aids to my education?

As I wrote in my previous post, I also still wonder how much of an issue this really is in domestic installations. As I said, whilst I can understand 'horrible waveforms', significant DC components in the current drawn seem much less likely?

Kind Regards, John

 

[ericmark]

With the three phase RCD it is common for the test button only to test between one phase and neutral I have found many a three phase RCD used with a single phase installation where the test button would never work as connected to wrong phase, but no report received in spite being installed for years, and the MK 10 mA RCD socket the test button would take out the 100 mA RCD, so one it must have been phase to earth rather than phase to neutral and two must have been drawing well over the 10 mA required to trip it.

As to CP RCBO thank you is does seem now you can get the type A version, since most of my inverter drive stuff is on one circuit it would make sense to change that RCBO although at £16 each changing all 14 gets a little expensive.

 

[JohnW2]

With the three phase RCD it is common for the test button only to test between one phase and neutral I have found many a three phase RCD used with a single phase installation where the test button would never work as connected to wrong phase ...

I think 'common' is probably an understatement, since I find it hard to think of any simple/safe/practical alternative. Even if one had everything triplicated, with three separate test buttons, if only one of the phases were energised, then two of the buttons would do nothing, without any obvious indication as to why..

.... but no report received in spite being installed for years ...

That's obviously just plain silly. There is no point in wasting one's time pressing a test button if one is not going to do something about it if the test 'fails'!

... and the MK 10 mA RCD socket the test button would take out the 100 mA RCD, so one it must have been phase to earth rather than phase to neutral ...

The test current surely has to be 'phase to earth' in order to test RCD functionality. 'Phase to neutral' would be like any other load, and would not trip an RCD.

... and two must have been drawing well over the 10 mA required to trip it.

A bit over, yes (≥15 mA). I seem to recall from previous discussions (and dissections of RCDs) that the test buttons usually generate a test current about 50% higher than the devices IΔn. If so, the test current generated by a 10 mA RCD might sometimes be just about enough to trip a 30 mA one.

Kind Regards, John

 

[flameport]

The test current surely has to be 'phase to earth' in order to test RCD functionality. 'Phase to neutral' would be like any other load, and would not trip an RCD.