Bi and Omni directional RCD's how can one tell what has been fitted.

[JohnW2]

People seem to be talking about current direction flow, which is clearly meaningless in this context.

Quite so.

It is the direction of power flow that is being addressed by the amendment.

Again, indeed - although I'm not really sure what gthat means (other than 'the obvious'

It would seem that where an RCD is designed as uni-directional ...

... which, as discussed, would beg the question as to why one would want to do that.

..., on tripping, if the device is connected back to front there will be some circuitry and or solenoid that should be disconnected by the operation of the device, but that remains connected, and fries.

True, and that's essentially what I wrote last night about 'which side feeds the electronics'. However that would only relate to powering of things within the device, and should not prevent the device from operating in response to a detected residual current, regardless of 'which way around' it had been installed, should it?

 

[JohnW2]

As far as I can tell, two issues, one inside the MCB ... <image> ... we have 8 Arc divider/extinguisher which is often only on one side ...

We're talking about RCDs, not MCBs.

However, since an arc would form at the same location (the location of the contacts when they started to open), regardless of which way around the device was installed, I would not have expected the physical location of the arc chamber to be dependent on which way around it had been installed, would it?

 

[DetlefSchmitz]

why one would want to do that.

Because it didn't used to matter.

should not prevent the device from operating in response to a detected residual current, regardless of 'which way around' it had been installed, shouldn't it?

As I tried to say, manufacturers have reported that under these conditions, the circuitry in question was liable to overheat if it remained powered. This is what the manufacturers' association were reporting that gave rise to the amendment. Don't shoot the messenger.

 

[DetlefSchmitz]

Again, indeed - although I'm not really sure what gthat means (other than 'the obvious'

That is clear!

 

[JohnW2]

Because it didn't used to matter.

I thinksomething approaching clarity is starting to emerge, along the lines that I suggested last night.

I don't think that any of us believe that there are any RCDs out there which won't respond satisfactorily to residual currents (L-N imbalances) regardless of which way around (supply/load) they are installed and nor do I think that any of us believe that there would be any point in designing (probably with considerable difficulty) a device which had such functionality.

Hence, any difference between uni- and bi-directional ones must relate to 'internal issues' (within the device), not to any potential inability to work satisfactorily as an RCD if 'installed the wrong way around'.

As I tried to say, manufacturers have reported that under these conditions, the circuitry in question was liable to overheat if it remained powered.

I thought that was what you were implying, but I didn't really understand. If internal parts remain powered after the device has operated, that's surely no different from them being continuously powered (as is nearly always the case) when the device has not operated, is it? If so, why should anything overheat when the device has operated if it doesn't overheat during the period (months or years!) during which it hasn't operated?

This is what the manufacturers' association were reporting that gave rise to the amendment. Don't shoot the messenger.

That's the first I've heard of such a "reporting". Do you know where I can find it?

 

[JohnW2]

That is clear!

As I implied, 'the obvious' is clear, but what does it mean electrically?

Whether AC or DC, if current is flowing through a conductor, how can that conductor (or anything else) know which end of the conductor is connected to 'the supply' and which is connected to 'the load' ?

In other words, if I presented you with a loop of wire coming out of a 'back box' which contained 'a supply' and 'a load', would you be able to tell me which end of the wire was connected to the supply and which was connected to the load?

 

[DetlefSchmitz]

In other words, if I presented you with a loop of wire coming out of a 'back box' which contained 'a supply' and 'a load', would you be able to tell me which end of the wire was connected to the supply and which was connected to the load?

Actually, given time to build the circuitry, yes of course. This is something you ought to be able to work out yourself.
As for the rest of it - I give up. It all just an excuse to build a post count. I found the answers on the internet within 10 min. You could do the same.

 

[ebee]

It would seem that where an RCD is designed as uni-directional, on tripping, if the device is connected back to front there will be some circuitry and or solenoid that should be disconnected by the operation of the device, but that remains connected, and fries.

yes that is the only option i see. in other words, it actually works and trips when it should, whichever way power is transferred but some (electronic withinn the RCD housing itself) parts will not disconnect when intended to or is designs too. It is not easy to see what detriment would actually occur should this happen

 

[JohnW2]

Actually, given time to build the circuitry, yes of course. This is something you ought to be able to work out yourself.

I'm afraid I can't, ("work it out myself") and it still seems to me that it would be impossible - so, if I'm wrong, I need help to understand, not 'criticism'.

Consider the simplest case of a battery connected to a light bulb with a bit of 2-core cable. You have access to part of that cable (the separate cores if you wish) but cannot see which end of the accessible bit of cable is connected to the battery and which is connected to the lamp. How on earth (without cutting the cable, hence interrupting the circuit) could you ascertain which end of the bit of the cable was connected to the battery and which to the bulb?

The same would seemingly be true if it were an AC supply and any sort of load.

What possible "circuitry" am I overlooking?

As for the rest of it - I give up. It all just an excuse to build a post count.

What on earth makes you think that I have any interest in "post count"? Yes, mine is high, but that's because I have chosen to post a lot of messages, not because of any interest in how many there have been.

I found the answers on the internet within 10 min. You could do the same.

You must be more clever or more skilful than I am. I've done a lot of 'looking on the internet' and I have yet to find an explanation which I understand and which makes sense to me - whether that be due to my stupidity, lack of intelligence or whatever..

For example I explained that (and why) I couldn't understand what you wrote about over-heating of internal parts of the device.

So, again, I need help in understanding whatever it is that I am 'missing', rather than jibes about 'post count' If, as you say, you found an understandable and credible explanation in 10 minutes, would you be able to share it with me?

 

[JohnW2]

yes that is the only option i see. in other words, it actually works and trips when it should, whichever way power is transferred but some (electronic withinn the RCD housing itself) parts will not disconnect when intended to or is designs too. It is not easy to see what detriment would actually occur should this happen

That was the point I made.

Yes, as I had suggested previously, it's possible that installing a device 'the unintended way around' could result in some of it's internal parts remaining powered after the device had operated. However, as I wrote more recently, that's surely no different from the continuous powering of those internal parts which usually goes on, uninterrupted, for months or years whilst the device is in-service - so, as you ask, how can that 'failure to disconnect the supply from internal parts' (after device has operated) possibly do any harm?

Kind Regards, John

 

[ebee]

Yes Indeed John, that was my point. I can`t see you can`t see, some others might not see yet some others again might see. Can the secret be shared to enlighten us, it may well be that we are missing something here and if so we need to know, if only to correct our knowledge/understanding. I`m not aware of anyone of us claiming to be infallible .

I`m the type of person that likes to get to the bottom of things, an arr yes so it is/is not, sort of person.
That`s why I was bugged by the Monty Hall conundrum, a couple of weeks of not thinking about it and the answer came out of the blue one day, I thought "Oh you twit, it`s easy, why did I not see it". That was after I`d spent a bit of time for 2 or 3 days trying to reason it out.

 

[flameport]

The proposed amendment is available to view here: https://standardsdevelopment.bsigroup.com/projects/2024-01247#/section
Login required, but you can register for free.

Not much too see, less than half a page of text and most of that is headings and titles.

 

[JohnW2]

Yes Indeed John, that was my point. I can`t see you can`t see, some others might not see yet some others again might see. Can the secret be shared to enlighten us, it may well be that we are missing something here and if so we need to know, if only to correct our knowledge/understanding. I`m not aware of anyone of us claiming to be infallible .

Exactly. I never make any claims of infallibility or omniscience, and quite often 'miss' things or get things wrong - so I am reliant on others 'educating' me when I need it.

Much as I currently find it 'hard to believe', it could be that those who believe they know 'the simple (or 'easily found') answer' are right - but, if so, it would be nice if they could share their knowledge and understanding, rather than criticising, or almost ridiculing, those who have not (perhaps!) 'seen the light'! With the tables turned, I would certainly try to help people understand, rather than making an issue out of the fact that they didn't understand.

I`m the type of person that likes to get to the bottom of things, an arr yes so it is/is not, sort of person.

As is presumably apparent, same here!

That`s why I was bugged by the Monty Hall conundrum, a couple of weeks of not thinking about it and the answer came out of the blue one day, ....

Only 2 weeks?

.... I thought "Oh you twit, it`s easy, why did I not see it". That was after I`d spent a bit of time for 2 or 3 days trying to reason it out.

I would say that you maligned yourself unnecessarily. A few years ago I saw a group of eminent academic statisticians at an international conference get very close to 'coming to blows' over it, and a fair bit of money was lost by some of them in making 'rash bets' with members of the other camp

I wouldn't say that it's as 'easy' as you suggest. Even for statisticians who eventually come to accept (and understand the logic and maths of) the true answer, it often still remains very counter-intuitive. In my personal case, even when I came to understand and sort-of accept that 'logic and maths', I still had to undertake computer simulations before I really accepted that 'the maths was actually correct'!

Kind Regards, John

 

[JohnW2]

The proposed amendment is available to view here: https://standardsdevelopment.bsigroup.com/projects/2024-01247#/section
Login required, but you can register for free. ... Not much too see, less than half a page of text and most of that is headings and titles.

Thanks - but, as you say, that does not really help with we've been discussing here, since the proposed new 530.3.201 consists of just two sentences, and says nothing that we do not already know.

If you've been following this thread, you will be aware of the fact that a good few of us are having difficulty in understanding what this is really all about - do you have an understanding or view that might help us?

I presume that no-one is disagreeing with the fact that, if 'power may be flowing in either direction' any RCD must be able to detect, and operate in response to, a residual current no matter what is the direction of power flow (which itself is not an easy concept to get one's {at least my!} head around!). What at least some of us are struggling with is the question of how any residual current device could fail to have that functionality.

 

[JohnW2]

I thought I had just had a 'Eureka Moment' but, although interesting and relevant, my excitement was rather short-lived - so I continue to 'not really understand' - I guess I qualify as "Confused of Buckinghamshire"

I still find it very hard to believe that any real-world RCD will fail to detect (and act upon) a residual current (L-N imbalance of current passing through it) resulting from an L-E fault on the load side of the device, regardless of which side of the device is connected to the load and which side is connected to the load.

However, it is apparent that no RCD could ever detect an L-E fault on the supply side of the RCD, since the L & N currents through the RCD would then not being unbalanced. It therefore follows that if the situation is such that the load and supply can 'swap places', in one of the two possible situations no RCD could ever protect against L-E faults on one side of the device (the side that was currently the 'supply' side).

That sounds like a very 'directional situation' but the problem is that it's an inevitability (regardless of anything) with any RCD - i.e. it would simply not be possible to create a device which could protect against L-E faults which were, at the time the fault arose, on the 'supply side' of the RCD, since the 'unbalanced current' would then not be flowing through the device.

The proposed 530.3.201 in Amd 3 says:

... Where bidirectional power flow is possible, only a device suitable for bidirectional power flow shall be used

If, as I would expect (and see *** below), "bidirectional power flow" means that either side could, at a particular point in time' be the 'supply' side then, as above, it would seem impossible to create a device which would always protect against L-E faults on a particular physical side of the device, wouldn't it? If so, I have to wonder what sort of device "suitable for bidirectional power" the proposed reg is thinking of.

[ *** Amd 3 also proposes to add this definition to Part 2 ..

Bidirectional protective device. A protective device where a power supply may be connected to either set of connection terminals ...

]