Is this article a Red Haddock ???

[flameport]

Why should that trip a type ac RCD or any RCD?

It won't trip an RCD, but could prevent an RCD from tripping when a fault occurs.

 

[JohnW2]

Why should that trip a type ac RCD or any RCD?

It wouldn't - but see flameport's reply (but also my reply to him which I am about to write).

 

[JohnW2]

It won't trip an RCD, but could prevent an RCD from tripping when a fault occurs.

As you will be aware, I've been trying hard to gain some understanding of the various 'Types' of RCDs, but am coming to wonder whether Type A ones are actually 'good enough' to achieve very much in many situations - so perhaps you can help me to understand? ...

BS7671 says:
.... For RCD Type A, tripping is achieved for residual pulsating direct currents superimposed on a smooth direct current up to 6 mA.

If that is true then the implication is that the RCD might well not trip if the underlying DC component of current is greater than 6mA. In the context of what we're discussing (a motor fed through a diode) I would imagine that the DC component is likely to be a lot more than 6mA, so does that mean that a Type A RCD would offer little/no benefit (over a Type AC) in such a situation?

 

[Harry Bloomfield]

Why should that trip a type ac RCD or any RCD?

It will not cause it to trip, rather it will slightly affect its ability to trip.

 

[JohnW2]

It will not cause it to trip, rather it will slightly affect its ability to trip.

As I've just written, on the basis of what little I know of the characteristics of a Type A RCD, it might even seriously affect the ability of a Type A RCD to trip!

 

[SUNRAY]

Oh wow, it's strange to see someone else asking the same question I asked 100 years ago

 

[JohnW2]

Oh wow, it's strange to see someone else asking the same question I asked 100 years ago

I've been asking questions about the behaviours of these different Types of RCD ever since I first heard of them and, despite having done a fair bit of reading, I'm still not very much the wiser.

On the basis of what little BS7671 says, Type A ones are only guaranteed/'required' to respond to a residual current in the presence of a DC component of up to 6mA - which is presumably relative peanuts in the sort of context we've been discussing here. On the other hand, BS7671 tells us nothing at all about the performance of Type AC RCDs in the presence of DC components (of any magnitude0 - but the implication appears to be that DC components even less than 6mA may impair their functioning - which, if true, would rather surprise me.

 

[fixitflav]

For modern electronics which use DC, there is a small loss converting AC to DC. However this is tiny, and even if properties were connected to DC supplies the conversion would still be required, as whatever voltage was supplied to a building would not be what was required by the electronic device.

Not only that, but the energy used by electronic devices is likely to be a small fraction of the total.

 

[ericmark]

Someone sent me 531.3.3 from BS7671:2018 where it goes through types AC, A, F&C with 0 mA, 6 mA, and 10 mA and I have tried to measure the DC leakage

with the meter shown, with AC not a problem as seen here, but with DC the meter needs to be zeroed, and any slight movement, and it needs zeroing again, so I have failed to measure DC leakage, would be interested to hear how others have got on measuring DC superimposed on the AC supply.

Even 10 mA is not much, and it would be easy for this to be exceeded, and the only other test one could do, is to test the RCD with all loads connected, and confirm if the RCD works.

I know with an EV charging unit we need a RDC-DD, but the question is what about the rest? Well I suppose a RCD is secondary protection with a TN supply, only with items outside and with a TT supply, is the RCD primary protection.

 

[flameport]

whether Type A ones are actually 'good enough' to achieve very much in many situations

They are better than Type AC, but only of use in situations that do not have smooth DC.
For most domestic properties they are good enough today, and certainly far better than the AC types of the past.
They will not be good enough in the future as more electronics, inverter drives and DC items are installed.
Eventually all RCDs will be Type B+ or some equivalent.

BS7671 tells us nothing at all about the performance of Type AC RCDs in the presence of DC components (of any magnitude

It won't, and neither will anything else as they were never designed for or tested with anything other than 50Hz sinusoidal AC.
Their behaviour with anything other than 50Hz AC is undefined and unknown.
Given the options of testing all makes and models of AC RCDs with various other currents or just banning the whole lot, most countries went with the ban several decades ago.

Type A ones are only guaranteed/'required' to respond to a residual current in the presence of a DC component of up to 6mA - which is presumably relative peanuts in the sort of context we've been discussing here.

Type A will work with pulsed DC which is what you get from a diode, there is no 6mA limit there.
The 6mA refers to where a continuous DC current exists, such as that you would get from a battery or solar panel, which is why installations that have those will have other types of RCD designed for DC.

A fixed DC current will saturate the magnetic core in the RCD and over a certain level it will not work, even if the fault current is AC. As most modern RCDs have tiny cores, it only takes a current of a few mA or 10s of mA to saturate the core and completely disable the RCD.
This does not happen with pulsed current as the magnetic field is not constant.

 

[JohnW2]

They are better than Type AC, but only of use in situations that do not have smooth DC. ... For most domestic properties they are good enough today, and certainly far better than the AC types of the past.

Thanks for the explanations. My first reaction when reading your ......

It won't, and neither will anything else as they were never designed for or tested with anything other than 50Hz sinusoidal AC. Their behaviour with anything other than 50Hz AC is undefined and unknown.

..... was that if the behaviour of Type AC RCDs with anything other than 50Hz AC is "undefined and unknown", then that would suggest that it might be possible that they actually work as well as Type A's with pulsating DC. However, having re-visited the video of your experiments about this (here), I see that such is not the case.

Type A will work with pulsed DC which is what you get from a diode, there is no 6mA limit there. The 6mA refers to where a continuous DC current exists, such as that you would get from a battery or solar panel ....

Fair enough. It sounds as if I misunderstood what was being said - since I was considering the fact that that half-wave (or full-wave) rectified AC can be analysed/regarded as a complex mixture of AC components (probably spanning a theoretically infinite range of frequencies) superimposed on a "smooth DC component" (the average of the waveform over time). However it seems that, at least in this context, no waveform which has varying voltage all on one side of zero is regarded as having a 'smooth DC component', even though analytically it certainly does!

A fixed DC current will saturate the magnetic core in the RCD and over a certain level it will not work, even if the fault current is AC. As most modern RCDs have tiny cores, it only takes a current of a few mA or 10s of mA to saturate the core and completely disable the RCD.

That is obviously inevitable, qualitatively, but my uncertainties have always related to the quantitative aspects. In the video of your investigations of this (here), the smallest DC component you use is 50 mA, and that did, indeed, increase trip threshold of a 30 mA Type AC RCD to 70 mA (essentially 'completely disabled'). However, 50 mA is a lot higher than the apparent design spec of a Type A RCD, (≤6 mA 'smooth DC') so I wonder whether a Type A would have done much/any better with 50 mA of DC? It would be interesting to know what results you would get with 'smooth DC' components a lot lower than 50 mA.

This does not happen with pulsed current as the magnetic field is not constant.

As a generalisation, that does surprise me, but maybe you're thinking of a fairly constrained definition of "pulsed DC"? Probably the most obvious literal meaning of 'Pulsed DC' is a square wave, and as the mark-space ratio of such a waveform increases, the waveform gets closer and closer to being 'true smooth DC'.

 

[ericmark]

I remember these

back on the building of Sizewell 'B' and see them in

@flameport video of 2020 clearly from sticker 2013 and think these are around 30 years old.

But back then there was no sign to say what type they were, they sent a signal to the moulded breaker, and it activated the solenoid I can't remember how many mA they were or even if fixed or variable, it was a long time ago, and we were lucky to have any RCD protection.

As to time, I remember a nail in the cable taking out 40 mS, type S, and variable set to a minute, and 30 mA, 100mA, 1 amp and 3 amp. Back in the days of Sizewell 'B' we did not have a tester. I think it was around 2000 when I got my hands on the first RCD tester, I suspect the old ELCB-v testers were expensive. And I have never used one. The loss of metal water pipes resulted in the ELCB-v, but I tried testing a RCD with a ELCB-v further down the line, and the RCD failed, and I wondered who fitted it, as it clearly never worked.

 

[JohnW2]

I remember these View attachment 405934 back on the building of Sizewell 'B' and see them in

@flameport video of 2020 clearly from sticker 2013 and think these are around 30 years old.
But back then there was no sign to say what type they were, they sent a signal to the moulded breaker, and it activated the solenoid ...

I somewhat doubt that many people knew anything about RCD 'Types' back then. However, the performance of the one in flameport's video obviously indicates that the one he had was, in current parlance, 'at least Type A'.

 

[ebee]

I never had or used a "Robin D-Lok" tester but I did hear that the concept usually worked but not always, the recommendation was not to test an RCD with the outgoing circuit attached, personally I preffered to do both with and without the circuit attached if possible and compare the results in an existing installation and likewise in a new installtion too, three sets of tests each time if possible.

 

[ericmark]

I think that is how they should be tested, even if the IET has reduced the tests, and also test results should be recorded on the EIC/EICR the entering of the maximum figures for tripping or loop impedance is unhelpful.

Why we test a RCD but never test a MCB seems a bit odd too. And with a RCBO would be helpful if they showed why they have tripped.

I now have both a ramp test, and a clamp-on to measure back-ground leakage, So with circuit connected if it trips at 20 mA one can reasonably assume the back-ground leakage is below the permitted 9 mA, but in the past only test was not tripping at 15 mA which was often done with circuits disconnected, I found early working with RCD's that strain on the cables could stop them tripping, so always want to test with all cables connected, with RCD / MCB turning off the MCB to test was reasonable, but with RCBOs you can't test with cables connected and load removed. So you don't really know if it will trip at 30 mA only it will trip with an extra 30 mA which is likely good enough, but does not really show it complies.