What type RCD for energy saving lights?

[ericmark]

It's not 'DC injection' at all, so long as the waveform remains symmetrical around zero.

Kind Regards, John

I really don't understand "PLUS pulsating DC components." as Hager puts it.

Single phase inverters, class 1 IT & multimedia equipment, power supplies for class 2 equipment, appliances such as washing machines, lighting controls, induction hobs & EV charging.

This would included my fridge/freezer, washing machine, freezer, TV, auto system in fact easier to list what does not have some kind of inverter.

Fact that type A RCBO's are not made for my consumer unit at 32A means I am not going to fit type A, not changing the consumer unit again, and so what is more to the point is in real terms will they protect me? So assuming my RCD tester is old and does not ramp up so 6 tests at half should not trip and full must trip but does not need to be in 40 mS however it seems it always is, and at 5 times must be within 40 mS and all tested on pos and neg cycle, had the tester for years and all pass.

But I assume we now have a means to measure the pulsating DC component and also a option to select that DC component on the tester so we know if the RCD fitted will work, but can't find out what this equipment is, and no point asking another electrician to test the RCD if his tester is basic same as mine.

So how do we know if a RCD is suitable. What is pulsating DC components?

 

[JohnW2]

I really don't understand "PLUS pulsating DC components." as Hager puts it.

I'm not sure what you are asking, since I presume you know what 'pulsating DC' is as well as I do - i.e. pulses which are all in the same direction relative to 'zero volts' (i.e. all pulse has the same polarity), such that the waveform has a net DC component (the 'average' over time of the unidirectional pulses).

I would think that is very rare these days. In the earliest days of, say, dimmers, they used thryristors (SCRs), rather than Triacs - so that the output pulses, when dimmed, were all of the same polarity (hence, even when turned up full, the load would only receive current during alternate half-cycles). However, with triacs (as with any PSU etc. using full-wave rectification), the pulses have both polarities, such that the waveform is symmetrical around zero - hence no 'DC component'.

But I assume we now have a means to measure the pulsating DC component and also a option to select that DC component on the tester so we know if the RCD fitted will work ...

I doubt that any such devices exist, not the least because we have no idea as to what degree of 'DC component' they would have to be able to tolerate whilst still working satisfactorily.

As I have said, I would imagine that the imbalance currents which one wants an RCD to detect will virtually always be AC, with the question apparently being as to whether the ('additional') presence of a 'DC component' will impair the ability of the RCD to detect that AC current imbalance.

I think you are worrying unnecessarily. Your standard RCD tester will tell you whether, in the absence of a 'DC component', the device operates as expected in response to an (AC) current imbalance. RCDs other than Type AC ones are apparently designed so that they will also continue to work satisfactory even if there is also a DC current component present, but (to the best of my knowledge) we have no means of testing that.

You should be used to the concept of there being aspects of protective device performance that we cannot test - we cannot test MCBs (or fuses ).

Kind Regards, John

 

[ericmark]

Yes I remember testing 6A MCB's with a fan heater, 1kW should hold 2kW should trip on a caravan site as people were doing fiddles to get more power, but other than that never tested a MCB.

What I am assuming is with DC residual magnetism can build up so the device sticks in some way, but I really fail to see how it would there are some electronic components in them today

looking at Wikipedia they are no longer pure mechanical devices, but I really fail to see how switching power would make them less sensitive, I have in the past found turning a whole consumer unit on will trip the RCD and you have to turn off MCB's first then slowly add the load, but not really worried if it makes it more sensitive it would fail safe, only worried if it would not work the same as a 100 mA S type, not really worried if will not work at 30 mA or in 40 mS as in real life water will normally allow more current to flow, and in the main they protect from wet electrical items. Basic idea is the RCD trips before you touch anything live, not as you touch some thing live, although it will also trip then but not before you get a shock.

It seems the only thing most worry about is the RCD on an electric car charging point, I will not call it simply an electric vehicle as my mobility scooter and e-bike are electric vehicles but don't use the special charging points that cars use.

I remember the talk about the passive and active, where the RCD needs power to trip the using it is important the voltage does not drop below the point when it will work, it does not matter with active type as once voltage drops below that point it will auto open (fail safe) I note you can get 110 volt RCD's but I would have thought they all should work down to 50 volt, at which point RCD protection is not required, but fitting a MCB to even a 12 volt unit they still work, with an RCD not so sure, even the mechanical type not sure if they would work, never tried would be pointless as not required.

But question is if you go into a house full of inverter drives and switch mode power supplies can you accept a type AC RCD? In my old BS7671:2008 I see nothing about type A, AC, B or F, and if there were no A, F, or B types I would be happy to use type AC, and maybe the type A is for use after the half wave rectifier and it does not matter before the half wave rectifier. I have seen some RCD's fitted where they would never work, like on an IT supply, he said it was required by the regulations so fitted it, but it would never trip.

 

[bernardgreen]

Many sets got the CRT heater supply from an extra winding on the line output transformer

How would the line output transformer be energised before the line oscillator was operating.

 

[ericmark]

All I know was there was a big green tube in the back of the TV which cracked in two, and I could not find a replacement, however worked out the resistance and it was the same as the iron, so fitted a 13A socket and plugged in iron to watch TV, OK if wife had been ironing cotton, but if she turned the heat down, it always seemed to be the exciting bit where iron thermostat would open and TV went off.

So there was more heat than used in a smoothing iron lost from the back of an old TV, it was very wasteful, but no one seemed to worry. Not sure when the ELCB first started to be fitted, seem to remember it when between the earth rod and the house bonding, and would trip at 50 volt, however if anything linked the house bonding to earth it stopped working, so the ELCB-v was replaced with the ELCB-c.

 

[JohnW2]

What I am assuming is with DC residual magnetism can build up so the device sticks in some way, but I really fail to see how it would ...

Quite. As I've said, I don't understand, either - which is why I was hoping that someone here might be able to enlighten me!

But question is if you go into a house full of inverter drives and switch mode power supplies can you accept a type AC RCD? In my old BS7671:2008 I see nothing about type A, AC, B or F, and if there were no A, F, or B types I would be happy to use type AC ...

BS7671:2018 has (in 531.3.3) got quite a lot to say about the various Types of RCD, and says that an 'appropriate type should be selected (but, as you can see, it also says "For general purposes, Type AC RCDs may be used"). However, I am now even more confused since, as you can see, it talks about the differences in relation to them being able to trip in response to non-sinusoidal-AC residual currents, rather than such currents impairing their ability to trip with AC imbalances) ....

531.3.3 Types of RCD
Different types of RCD exist, depending on their behaviour in the presence of DC components and frequencies. The
appropriate RCD shall be selected from the following:
(i) RCD Type AC: RCD tripping on alternating sinusoidal residual current, suddenly applied or smoothly
increasing
(ii) RCD Type A: RCD tripping on alternating sinusoidal residual current and on residual pulsating direct current,
su ddenly applied or smoothly increasing.
NOTE 1: For RCD Type A, tripping is achieved for residual pulsating direct currents superimposed on a smooth direct
current up to 6 mA.
(iii) RCD Type F: RCD for which tripping is achieved as for Type A and in addition:
(a) for composite residual currents, whether suddenly applied or slowly rising, intended for circuit supplied
between line and neutral or line and earthed middle conductor
(b) for residual pulsating direct currents superimposed on smooth direct current.
155
NOTE 2: For RCD Type F, tripping is achieved for residual pulsating direct currents superimposed on a smooth direct
current up to 10 mA.
(iv) RCD Type B: RCD for which tripping is achieved as for Type F and in addition:
(a) for residual sinusoidal alternating currents up to 1 kHz
(b) for residual alternating currents superimposed on a smooth direct current
(c) for residual pulsating direct currents superimposed on a smooth direct current
(d) for residual pulsating rectified direct current which results from two or more phases
(e) for residual smooth direct currents, whether suddenly applied or slowly increased, independent of
polarity.
NOTE 3: For RCD Type B, tripping is achieved for residual pulsating direct currents superimposed on a smooth direct current
up to 0.4 times the rated residual current (IΔn) or 10 mA, whichever is the highest value.
For general purposes, Type AC RCDs may be used.
NOTE 4: For guidance on the correct use of RCDs for household and similar use, see PD IEC/TR 62350.
NOTE 5: Some typical fault currents in circuits comprising semiconductors are given in Annex A53, Figure A53.1.

However, in terms of testing, it merely says:

643.8 Additional protection
The verification of the effectiveness of the measures applied for additional protection is fulfilled by visual inspection
and testing. Where RCDs are required for additional protection, the effectiveness of automatic disconnection of
supply by RCDs shall be verified using suitable test equipment according to BS EN 61557-6 (see Regulation
643.1) to confirm that the relevant requirements of Chapter 41 are met.
NOTE: Effectiveness is deemed to have been verified where an RCD meeting the requirements of Regulation 415.1.1
disconnects within 40 ms when tested at a current equal to or higher than five times its rated residual operating
current.

... and you may be interested in (if you can read it!) the following (from 'Annex A53) ...

Kind Regards, John

 

[winston1]

How would the line output transformer be energised before the line oscillator was operating.

It wasn't and didn't need to be.

 

[winston1]

In the earliest days of, say, dimmers, they used thyristors (SCRs), rather than Triacs - so that the output pulses, when dimmed, were all of the same polarity (hence, even when turned up full, the load would only receive current during alternate half-cycles).

They did indeed. I actually built one from an article in PW. But it contained a bridge rectifier for the thyristor so it did work on both polarities.

 

[JohnW2]

They did indeed. I actually built one from an article in PW. But it contained a bridge rectifier for the thyristor so it did work on both polarities.

I still have several which I built, but all are as I described, producing output pulses during just one half cycle. Indeed, a bridge rectifier would not make any difference, unless one had two thyristors/SCRs.

 

[flameport]

Quite. As I've said, I don't understand, either - which is why I was hoping that someone here might be able to enlighten me!

There are several different issues, all of which are at least partly related.

The basic concept is that a typical AC RCD contains a ferromagnetic core with a sense winding around it, and the circuit conductors pass through the centre once, or multiple times in some devices.
To detect residual current, a changing magnetic field must be created in the core, which then induces a current in the sense winding and is used to trip the mechanism directly via a solenoid, or on more modern devices passed to an electronic amplifier or other circuit.

Constant DC:
If DC current is passed through one of the circuit conductors, that magnetises the core but is invisible to the sense coil as that can only detect AC.

If enough DC current flows, it magnetises the core to the point of saturation, so that any additional residual current (including AC) cannot create any additional magnetic field, and therefore no current can be induced in the sense winding, and the device will not trip.
For typical AC RCDs, the size of the core is tiny and the current required to completely prevent the RCD working is in the range of a few 100 milliamps.

Pulsed DC:
A similar effect can occur with the DC saturating the core. Plenty of AC RCDs don't react to pulsed DC (half wave rectified AC). Some will trip but only on pulses of one polarity, not the other.
Even when they do trip, the trip time and/or current required can be far in excess of the normal specifications of the device - at least 10x more in some tests that I have done.

The rest of it is far more complex, as real installations do not have nice clean 50Hz AC - it's a horrid distorted mess of various frequencies. Many AC RCDs only work properly at 50Hz - any higher frequency components in the waveform will result in it not working properly or at all. There are many other problem situations such as pulsed DC that's in phase with the AC, or out of phase, differing polarities and so on.
The main sources of this are switching power supplies, which unfortunately are found in the vast majority of modern electronic equipment, including things which traditionally never had them such as cookers, fridges and washing machines.

Several countries including Germany banned the use of type AC RCDs several decades ago for these reasons - they are totally unsuitable for modern electrical installations.

This video explains the basics and includes a demonstration of how little DC is required to totally defeat an AC RCD.

 

[JohnW2]

There are several different issues, all of which are at least partly related.

Many thanks for your input.

The basic concept is that a typical AC RCD contains a ferromagnetic core ... If enough DC current flows, it magnetises the core to the point of saturation, so that any additional residual current (including AC) cannot create any additional magnetic field ...

Yes, that is the one thing which I (and, it seems, also eric) had thought of. However, as I recently wrote, I got a bit more confused when I read what BS7671 has to say about the different types of RCD, in that, rather than talking about the ability of DC currents to inhibit response to AC current imbalances, it seems to major on the ability of the non-AC types to detect (and be tripped by) imbalances in pulsed/asymmetric etc. currents.

Several countries including Germany banned the use of type AC RCDs several decades ago for these reasons - they are totally unsuitable for modern electrical installations.

If the other types are equally good at reacting to simple AC current imbalances, then it's difficult to see why the AC types even exist - I presume it must be a matter of cost?

So, how do the other types circumvent the problem - presumably not just by avoiding having a magnetisable core, since there would only be one non-AC type if that were the only difference?

Kind Regards, John

 

[winston1]

I still have several which I built, but all are as I described, producing output pulses during just one half cycle. Indeed, a bridge rectifier would not make any difference, unless one had two thyristors/SCRs.

The thyrister is fed by the bridge so it produces outputs on both half cycles.

 

[ericmark]

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

If a fault with washing machine can stop the RCD working then swapping 2 RCBO's for type A assuming I can find one to fit the board, on the sockets supply in main house would mean the sockets which is likely the main thing where we a likely to get a shock is covered, but will a fault on the sockets affect the lights?

So taking my house, if I fit a 10 mA RCD socket on the socket supply and on another socket there is a fault injecting DC will my 10 mA RCD socket type AC still work?

 

[Risteard]

Type AC should be banned (as most of the world have). There is no good reason to fit them anymore.

 

[ericmark]

Type AC should be banned (as most of the world have). There is no good reason to fit them anymore.

After reading it seems you are right, however your not permitted to fit other manufacturers equipment into a Consumer Unit and retain it's type testing, so I have a CP (Electric) Limited Kilmarnock KA2 0BA www.cpelectric.co.uk consumer unit only fitted last year from Link electrical which it seems only makes type A RCBO as 6 amp, I am sure many people has a similar set up. I have 14 RCBO's fitted so at £10 each not too bad at £140 but if I need to replace with type A at around £25 each that's £350, that is a lot of money to upgrade a consumer unit less than one year old.

So three questions.
1) Does it really need upgrading?
2) If so do all RCBO's need changing?
3) Is there any other options?

In the main the danger zone is the garden, so if I fitted a socket for use in the garden then the main danger is removed, I have a 10 mA active RCD socket, but yes you have guessed it, it's type AC.

But how do you find a type A socket outlet, when I bought the RCBO's they were marked type B but it seems they are curve B not type B, the only thing on them that shows they are type AC is a very small logo

which shows it is type AC, for type A it should be

and to be frank I needed two pairs of glasses to read the logo it is that small.