Type A/B RCDs can you help explain the component parts?

See table A53.1 on p192 of the BBB: it has a series of electronics diagrams where a fault has been been introduced, and shows for each what the fault waveform looks like and which types of RCDs can cope.


There is however (IMHO) still a lack of good information available which clearly distinguishes between an RCD tripping on a particular waveform and merely not being blinded by it. This has implications for co-ordination with upstream devices.

I think at the most basic level, what is important is what the product is tested and certified to.


I understand Zs asking the question about whether any particular product may or may not be blinded, to perhaps inform a decision, or even technical curiosity.


However, from the point of view of answering whether an RCD may or may not protect someone [within the bounds of BS 7671 and the product standard of course] is it not more pertinent to understand what the product has been designed and tested for (e.g. Type A residual current detection), not what a product designed and tested for something else might or might not do when subjected to residual currents it outside the specification [standard]?


I'm sure we've all come across even Type AC RCDs tripping during continuity tests, for example, which are carried out with DC test currents often in excess of 6 mA - but that sort of anecdotal evidence is at best very subjective and probably unreliable, because:

1. When the test current is applied, it's not "steady DC" at the moment the current is applied.

• Anecdotal evidence also shows that the trip doesn't always occur when the DC continuity test current is passed through the same device on the same day, shortly before or afterwards.

Thank you for that Wallywombat,  those are interesting as is the whole of  Annex A53 which I had quite forgotten about and ought to read more often.  Good revision.

Zs

Thank you Graham, 


Technical curiosity only,  and wanting better to understand. I am not about to start installing EV charging points.  But you know those conversations we overhear where person A is giving person B advice?  After hearing one such, just the other day, I came home with a few question marks. Am looking it up and I was right to question. Let me reassure you that I'm not looking to find a product that does the specified job of another. Bit surprised by that suggestion.


We have the component parts of a standard type A RCD drummed into us so well that we can all draw a diagram and explain what is going on to make it trip and what happens when it trips. 10 points in an exam for that.  But the type B, presumably, is operating differently?


For an installer or end-user; yes, I agree completely about the pertinence of knowing what to buy and where to apply it. that's usually enough for most and that's fine.  IMHO though it doesn't have to be a mystery regarding any difference in what's going on inside. But I sure can't find anything.


Yes, I have anecdotal on those tests - one of my Meggers trips RCDs on various tests unless I swap the probes over which was mentioned on here some years back.  Yes, it can stop tripping them in the blink of a tea break. The other Megger has never tripped an RCD unless I tell it to.


Zs

The sensing of an AC current, with ring cores as a current transformer and so forth is well established. how that core is scaled to handle waveforms that are full wave rectified bumps rather than sine waves is a detail of size of core, and perhaps change of magnetic material.

The sensing of a DC that we must assume comes on slowly (perhaps ramped over several seconds) so not to look 'AC-like' though the transformer, has to be detected rather differently - we are really now looking to detect a static magnetic field.

so the main part is like a normal RCD, plus a sensor for static magnetic fields.


Two common ways to detect an un changing magnetic field.. I have not got  a B type in front of me, and if I did, I think given the price, I'd think hard about reaching for a hacksaw, but it will be based on one or other of these.


Hall effect devices do this by deflecting the flow of a measurement current (much as the magnetic filed of the deflection coils of the old CRT telly moved the electron beam about, but the electrons are not a 'gas' of charges in vacuum, but instead diffuse  through  a doped semiconductor,and the amount the current is deflected more to one electrode than the other is measured in microns per mm.)

Hall effect devices are used in  the modern car to detect notches in camshafts as they whizz  round, and are mass produced cheap items.


The other method is the 'flux gate' where an AC signal, perhaps at a few kHz sweeps the magnetic core to saturation in both directions.

When a DC generated magnetic field is present in addition to the test waveform, the clipping at the two extremes from the core saturation becomes asymmetric, and that can be detected. This is a surprisingly sensitive technique, and crossed pairs of such sensors are often  used in electronic compasses in smart-phones, where it costs a few pence.


the  third and final component of the B type RCD must be a gold bar given the price, but I am at a loss to see what it does in there.

Mike.

To some extent it's all down to "rigidly defined areas of doubt and uncertainty" if you like.


AC types have no defined response to d.c. or partial waveform a.c. residual currents - some certainly can be disabled by d.c. (remember the D-LOC loop testers?) but others may (and sometimes do) behave differently - more modern ones seem very close to A types - but the standard doesn't guarantee that behaviour so you can't rely on it in any design.


A types, as well as responding to partial a.c. waveforms, are required to behave properly with a residual d.c. current of up to 6mA - which gives you some leeway (e.g. where EV charge points have their own trips that'll disconnect at 6mA d.c.) ... provided of course you don't have multiple possible sources of d.c. downstream... What happens if you feed them more than 6mA of d.c. is again "undefined" ... they might work as expected, or they might not.


B types have much smaller areas of doubt and uncertainly - but still some.


(Then add in F-types, K-types, EV-types and B+ types just for extra confusion)


    - Andy.

Page 9 of this ABB guide to RCDs indicates that their ones at least do use the flux gate method to detect the DC currents.


Mike