RCD TD & Single pole RCBO

[AndyPRK]

The first similar topic at the bottom says I need a double pole earth leakage device to seperate N and E

 

[stillp]

Is that correct?

No.
(Please bear in mind I am not an RCD designer, so this is a simplified version of something I was told by one!)
In each half-cycle of a sine wave, the current rises from zero to a peak and then decays, and rises in the other direction. A steady current will be pretty well equal in each half-cycle, so there is no imbalance. Now consider a sudden drop in circuit impedance during the rising part of one half-cycle. The rate of rise of current will be much greater than expected, and that rate of rise will not be achieved during the next half-cycle. This results in an imbalance over the full cycle.
I think that was the basis of the explanation.

 

[JohnW2]

The first similar topic at the bottom says I need a double pole earth leakage device to separate N and E

Well, yes - as has been said, the problem only exists because virtually all RCBOs used in domestic installations are single-pole. With a double-pole RCBO (or, equivalently, with an RCD, inevitably DP, plus an MCB), the potential problem would go away. However, although they exist and (as eric said) some even in single module width, you will find very very few in domestic installations.

Kind Regards, John

 

[JohnW2]

Is that correct?

No. (Please bear in mind I am not an RCD designer, so this is a simplified version of something I was told by one!) .... In each half-cycle of a sine wave, the current rises from zero to a peak and then decays, and rises in the other direction. A steady current will be pretty well equal in each half-cycle, so there is no imbalance. Now consider a sudden drop in circuit impedance during the rising part of one half-cycle. The rate of rise of current will be much greater than expected, and that rate of rise will not be achieved during the next half-cycle. This results in an imbalance over the full cycle.
I think that was the basis of the explanation.

That's a bit like some of the arguments put forward in regard to RCD operation in response to switching things off etc. However, the explanation 'fails' for the same reason. Everything you say is clearly correct until we get to the last sentence. Unless Mr Kirchoff has re-thought his ideas, in the absence of anywhere else for current to go (which might possibly haven by virtue of inductive or capacitive coupling to external conductors etc.), what you describe will happen identically (at any point in the cycle) in both L and N conductors (everywhere in the circuit, including in the RCD) - so that there will not be any L-N imbalance (at any point during the cycle, or averaged over a cycle, or whatever) for the RCD to sense.

Kind Regards, John

 

[stillp]

Remember that a high rate of rise of current is the same as injecting high frequencies, so those inductive and capacitive effects become significant.

[Are you using predictive text, or did something odd happen in your haven?]

 

[ban-all-sheds]

Unless Mr Kirchoff has re-thought his ideas, in the absence of anywhere else for current to go (which might possibly haven by virtue of inductive or capacitive coupling to external conductors etc.), what you describe will happen identically (at any point in the cycle) in both L and N conductors (everywhere in the circuit, including in the RCD) - so that there will not be any L-N imbalance (at any point during the cycle, or averaged over a cycle, or whatever) for the RCD to sense.

Mr Kirchoff has not rethought. And neither have Messrs Clausius or Rankine.

If a rapidly changing voltage wrt earth, not neutral, induces a voltage in a separate conductor, wrt earth (or neutral), then that's not free energy. The RCD will see it as some current leaking to earth.

 

[JohnW2]

Remember that a high rate of rise of current is the same as injecting high frequencies, so those inductive and capacitive effects become significant.

Yes, of course - that's why I mentioned them. However, to create an L-N current imbalance at the RCD, those effects have to result in substantially more 'current loss' (to somewhere, probably earth) from L than from N or vice versa. That's very unlikely with inductive effects, since about the same current should be flowing through L and N (at any point in time), but it seems probable that there will be more capacitive 'current loss' from L than N (to E), which could result in an L-N current imbalance at the RCD that would, if large enough, cause it to operate.

However, those are phenomenon of the circuit (wiring), not the RCD. I think I still stick to my view that ( now qualified by 'in the absence of other phenomenon occurring in the wiring of the circuit'), a 'perfect' RCD would not operate solely as a result of a very rapidly rising fault current - i.e. if the fault were very close to the RCD, thereby eliminating those capacitive effects in wiring.

[Are you using predictive text, or did something odd happen in your haven?]

I don't go within a mile of 'predictive text', which drives me mad. What you saw was an interesting bit of 'subconscious semi-phonetic typing' (well, "haven" doesn't sound that much different from "happen", does it?!), maybe aided by the Merlot

Kind Regards, John

 

[JohnW2]

Mr Kirchoff has not rethought. ... If a rapidly changing voltage wrt earth, not neutral, induces a voltage in a separate conductor, wrt earth (or neutral), then that's not free energy. The RCD will see it as some current leaking to earth.

Yep, that's roughly what I was typing whilst you posted that - although I don't think your explanation is very clear. A rapidly changing voltage relative to earth will also be a rapidly changing voltage relative to neutral. As I said, it would probably have to be capacitive, not inductive, coupling to something (presumably something earthed, like the CPC of a cable), since essentially the same (rapidly changing) current ought to be flowing in both L and N, so that any 'inductive loss' would probably effect current in N and L similarly, hence not create an L-N imbalance for the RCD to see.

I have to say that, no matter how high the frequency of the transients, I'm a little surprised that as much as 30mA could/would flow through the tiny capacitance that we would normally be talking about.

Kind Regards, John

 

[RF Lighting]

I guess that the phenomenon of failing lamps taking out RCDs which I have seen happen, is due to the the huge currents involved when a plasma arc forms. A tiny tiny amount of power relative to the current involved is leaked, but a tiny tiny percentage of many KAs could easily exceed 30mA.

But whether this power is lost due to capacitive effects, ionisation of the air, leakage to true earth or something different altogether I don't know.

 

[JohnW2]

I guess that the phenomenon of failing lamps taking out RCDs which I have seen happen, is due to the the huge currents involved when a plasma arc forms. A tiny tiny amount of power relative to the current involved is leaked, but a tiny tiny percentage of many KAs could easily exceed 30mA.

That's all true, but it obviously relies on the supply being able to supply 'many kAs'. In my domestic installation, the PSSC, even at the origin, is only about 600A, and is more like 200A-300A at most light fittings.

But whether this power is lost due to capacitive effects, ionisation of the air, leakage to true earth or something different altogether I don't know.

To be pedantic, it's current, not power that would be being 'lost' (from one of the intended conductors, probably the L) and, in practice, I imagine that it has to be a loss to earth (either to a CPC/ something connected to CPC or to 'true earth) - and the only mechanisms I can really think of are either capacitive or, as you say, leakage through ionised air.

As I've said, it's not something I can recall ever having experienced personally in my domestic environment, and I lived with countless RCD-protected incandescent lamps/bulbs for many years.

Kind Regards, John

 

[bernardgreen]

Like John I cannot recall any RCD being tripped by a lamp blowing. MCBs have tripped but not the RCD protecting the circuit.

The sensor in an RCD is a toroidal core with the Live and Neutral passing through it in opposite directions. Each create a magnetic field in the core but as these are in opposite directions they cancel each other out so there is no magnetic field in the core and therefore no output from the sense winding.

In reality the amount of magnetic field per amp is affected by the position of the wire in the aperture of the core. If it is touching the core the field created is larger than if the wire was dead centre of the core.

With equal currents in Live and Neutral the difference in field strengths due to the positions of the wires in the cores is negligible compared to that field created by a 30mA difference in the currents. Therefor for normal use with currents with in the rating of he RCD the positions of the wires in the core is not critical.

When a high current pulse from a lamp in plasma mode passes on both Live and Neutral the differences in field strengths due to the positions of the wires in the aperture of the core could be large enough to trip the RCD even though the currents in the Live and Neutral are equal and opposite.

 

[JohnW2]

Like John I cannot recall any RCD being tripped by a lamp blowing. MCBs have tripped but not the RCD protecting the circuit.

I'm glad to hear that I'm not alone!

When a high current pulse from a lamp in plasma mode passes on both Live and Neutral the differences in field strengths due to the positions of the wires in the aperture of the core could be large enough to trip the RCD even though the currents in the Live and Neutral are equal and opposite.

We're all essentially scraping the intellectual barrel in attempts to think of a mechanism, but I suppose that one is as credible a contender as any other suggestion.

We are obviously talking about fairly extreme situations. Even if (as is probably common in domestic installations), the PSCC at light fittings is only 'a couple of hundred amps', if such a fault current arises within, say, half a cycle, that would represent a rate of change of current of around 20,000 amps/second.

There is, of course, nothing particularly special about lamps/bulbs - they are just very common things. One assumes that if this phenomenon occasionally occurs, it could be in response to any suddenly-appearing negligible impedance L-N short.

Kind Regards, John

 

[Risteard]

Like John I cannot recall any RCD being tripped by a lamp blowing.

I witnessed it in my own house.

 

[JohnW2]

Like John I cannot recall any RCD being tripped by a lamp blowing.

I witnessed it in my own house.

It clearly does happen, at least occasionally - it's the why/how that is less clear. In some situations (e.g. with earthed metal lamp holders), there clearly is more scope for it happening, particularly if the envelope explodes.

Kind Regards, John

 

[Risteard]

Like John I cannot recall any RCD being tripped by a lamp blowing.

I witnessed it in my own house.

It clearly does happen, at least occasionally - it's the why/how that is less clear. In some situations (e.g. with earthed metal lamp holders), there clearly is more scope for it happening, particularly if the envelope explodes.

Kind Regards, John

In my case it happened with an insulated lampholder on a standard ceiling rose. It was a couple of years ago with my cousin staying in the house too at the time. I nearly lost all of the food in my freezer as a result but noticed in time!