Extreme non-discrimination or faulty MCB?

[JohnW2]

As others have said, time to replace it if suspect, NHX are the current range and are readily available and at a low price

Indeed. As I said/implied at the very start, that is my intention. I'll probably do it later today.

That said, I'd be tempted to try and find a way to test the old one with a high current low voltage AC source before slinging it into the dustbin.

I might, just as a matter of interest, but it would make no sense not to 'throw it into the bin' afterwards. As you say, they cost virtually nothing, and if I'm replacing it because of 'uncertainties' about it, it would make no sense to use it again, regardless of what testing might show. In fact, given that it's 'only' a B6, I could theoretically test it 'in situ', since there is a small (8.5kW) shower (virtually never used) supplied by a B40 in that CU - and that ought to draw just about enough current to relatively safely test the B6's magnetic trip. I wouldn't personally attempt to 'test', by any method, the magnetic trip of an MCB requiring a higher trip current than a B6!

I've actually seen a fault (caused by a crushed cable) on a 1.5mm blow itself clear before a breaker tripped, but that was on a federal stablok type 4 Zs was out of spec at end of circuit, however probbaly wasn't at the point the fault occurred. Backup protection was BS88 100A which also stayed intact.

There is seemingly nothing wrong with the circuit. Once the fuse was replaced, it works fine, with a plenty low enough Zs and it IRs OK.

Kind Regards, John

 

[mfarrow]

Did you perhaps also mean "...should have low enough circuit impedance..."??

No no, that part was quite correct. My thinking was that the fault current from source (Ze) may be low enough for a high-current short-duration fault which may have "crippled" the MCB (in whatever fashion), but that actually the impedance of the circuit would be high enough to dampen and elongate the fault to a duration suitable to allow the MCB to unlatch, and without damage.

The fact that it takes only half a cycle anyway to unlatch an MCB means my thinking at near midnight that it may not have time to do so may be slightly flawed.

I've no idea where my post went though, did the mods not take to kindly to my testing MCBs with shower switches? They're designed to make fault current you know!

 

[JohnW2]

Did you perhaps also mean "...should have low enough circuit impedance..."??

No no, that part was quite correct. My thinking was that the fault current from source (Ze) may be low enough for a high-current short-duration fault which may have "crippled" the MCB (in whatever fashion), but that actually the impedance of the circuit would be high enough to dampen and elongate the fault to a duration suitable to allow the MCB to unlatch, and without damage. ... The fact that it takes only half a cycle anyway to unlatch an MCB means my thinking at near midnight that it may not have time to do so may be slightly flawed.

OK. I think your late-night thinking may, indeed, have been flawed - since you really seem to be postulating a loop impedance which was at the same time, both very low and reasonably high! As I've said before, the distinction between the external and internal parts of the loop impedance is irrelevant - it's the sum of the two which matters [on a technical issue, as I said before, it was almost certainly the external L-N loop impedance, not "Ze" (external L-E loop impedance), which represented the fault path in my case].

I've no idea where my post went though, did the mods not take to kindly to my testing MCBs with shower switches? They're designed to make fault current you know!

I presume that it (and my confused response thereto) was removed because it was irresponsibly suggesting something potentially diabolically dangerous. To suggest that anyone should use a switch (or anything else) to deliberately introduce a short circuit on a low-impedance 230V circuit which may only be protected by a 100A fuse is just crazy, and potentially very dangerous. When I responded, I had misread your suggestion - again late at night, I thought you were talking about using a shower, not a shower switch, to test an MCB. As I've subsequently said, I regard using a shower to test a B6 (but no other MCB) as OK (given sensible precautions) (and may well be about to do it!), but would never regarded it as sensible or acceptable to attempt to deliberately introduce a short circuit across 'the mains' - nd definitely not responsible to put the idea into the minds of others.

Kind Regards, John

 

[mfarrow]

OK. I think your late-night thinking may, indeed, have been flawed - since you really seem to be postulating a loop impedance which was at the same time, both very low and reasonably high! As I've said before, the distinction between the external and internal parts of the loop impedance is irrelevant - it's the sum of the two which matters [on a technical issue, as I said before, it was almost certainly the external L-N loop impedance, not "Ze" (external L-E loop impedance), which represented the fault path in my case].

The equivalent of Ze then (I need the book). I know the internal and external parts are irrelevant, I was just considering them both and bringing my thinking into the context of a lighting circuit. A cooker circuit in a building with a low Ze for example may be reasonably assumed to have a fairly low internal loop R1 and Rn, and therefore would carry a more significant fault current than your lighting circuit with the same low Zdoodah.

To suggest that anyone should use a switch (or anything else) to deliberately introduce a short circuit on a low-impedance 230V circuit which may only be protected by a 100A fuse is just crazy, and potentially very dangerous. When I responded, I had misread your suggestion - again late at night, I thought you were talking about using a shower, not a shower switch, to test an MCB. As I've subsequently said, I regard using a shower to test a B6 (but no other MCB) as OK (given sensible precautions) (and may well be about to do it!), but would never regarded it as sensible or acceptable to attempt to deliberately introduce a short circuit across 'the mains' - nd definitely not responsible to put the idea into the minds of others.

It's done quite regularly in other industries, perhaps I was a little out of touch for suggesting it in an industry where it is clearly not normal. They would put something suitably sized in the circuit to protect the main bomb (as you've got), and it's usually a rotary isolator that's used.

It's done to:

- ensure discrimination
- verify disconnection times
- confirm proper operation of the devices

Being able to verify all these things in a controlled manner is much preferable to those industries than having the inconvenience of a fire, poor discrimination or a failed device. It also ensures the sort of "guess which of my 5 devices operated when my lamp blew" style of questions asked here do not occur.

I've thrown away properly sourced (and expensive) MCCBs in the past because they failed to operate when expected.

You've seen first hand what happens when a device fails to operate as expected, and it's lucky that no one was under the bulb and no molten metal dropped onto the bedsheets and set them on fire. However clearly this is a rare event which has not been proven to be a repeatable problem so do not try any of the above at home.

 

[bernardgreen]

However clearly this is a rare event which has not been proven to be a repeatable problem so do not try any of the above at home.

I am not so sure it is a rare event, I know of at least four similar incidents where capsule tungsten halogen lamps have exploded. With a 6 amp current there is 1.5 kW of energy in the plasma as the lamp melts out. And the small separation between the filament supports means the plasma can be maintained on the 230 volts until the lamp falls apart.

 

[JohnW2]

The equivalent of Ze then (I need the book).

Indeed. Of course, with TN-C-S it will be the same as Ze.

I know the internal and external parts are irrelevant, I was just considering them both and bringing my thinking into the context of a lighting circuit. A cooker circuit in a building with a low Ze for example may be reasonably assumed to have a fairly low internal loop R1 and Rn, and therefore would carry a more significant fault current than your lighting circuit with the same low Zdoodah.

That is all obviously true, but I'm still not really sure what point you are making. In view of the difference in cable sizes, one obviously would expect a lower loop impedance on a cooker circuit than a lighting one, so the PSCC would be appreciably higher for the shower circuit. However, in my case, the overall (external+internal) L-N loop impedance was under 1Ω**, hence a PSCC of at least 230A at 230V (at least 240A with my ~240V supply) - obviously far more than enough to magnetically trip a B6, and also probably enough to blow the 60A fuse pretty quickly. However, I would not have expected an 'instant fuse blow' at that current (albeit I'm not really sure exactly how quickly the fuse blew)
[** the 1mm² lighting circuit, from local CU to the offending light fitting, was very short - nearly all of the 'internal' L-N loop impedance related to the (16mm² or 25mm² - I haven't yet checked) distribution circuit - hence probably a lower impedance than even your cooker circuit.]

Kind Regards, John

 

[JohnW2]

I am not so sure it is a rare event, I know of at least four similar incidents where capsule tungsten halogen lamps have exploded. With a 6 amp current there is 1.5 kW of energy in the plasma as the lamp melts out. And the small separation between the filament supports means the plasma can be maintained on the 230 volts until the lamp falls apart.

In my case, of course, unless the fuse was 'faulty', it was not 6A but, rather, something appreciably in excess of 60A. However, the capsule did not explode - it simply 'fell off'. The base of the lamp/bulb was totally intact, as were the contacts of the lampholder, not even showing any 'blackening', let alone any damage. The only defect was the complete absence (in both base and capsule) of any signs of the two wires that had once connected capsule and base:

Untitled

• JohnW2
• 30 Dec 2014
• 1

Kind Regards, John

 

[JohnW2]

OK, I've now replaced the B6. No doubt some will criticise me but, before removing it, I 'tested' it by temporarily moving the 8.5kW shower onto that MCB. On two occasions, about five minutes apart, application of that load (about 35A) resulted, as expected, in 'instant' operation of the B6.

...none of which really helps me to understand 'what happened'. I suppose that the fuse may have been on its last legs (or faulty from birth) and actually blew at a current appreciably below 30A. ... or, or course (per title of thread) I suppose that it may have just been a fairly extreme case of non-discrimination.

Kind Regards, John

 

[mfarrow]

However, in my case, the overall (external+internal) L-N loop impedance was under 1Ω**, hence probably a lower impedance than even your cooker circuit.]

In which case that settles it. I knew you had a mansion with TT supply and 3 submains, I was expecting the final circuits to be as monstrous.

It's peculiar that your circuit breaker is now OK. Fault current would prove it though*.

Are you sure there's nothing like a shared line going on somewhere? Two lighting circuits jointed together somewhere or some such?

* (runs and hides)!

 

[JohnW2]

In which case that settles it. I knew you had a mansion with TT supply and 3 submains, I was expecting the final circuits to be as monstrous.

It's no mansion, but it is a large house, with several (more than 3) submains. However, the whole point of that arrangement was so that the final circuits from each of the 'local' CUs would be very short, hence allowing Zs (and L-N loop impedance) to be kept low. The cable route from the offending lamp to the local CU is probably no more than about 5 metres.

It's peculiar that your circuit breaker is now OK. Fault current would prove it though*.

35A really is effectively 'fault current' as far as a B6 is concerned - in that it should be adequate to result in a magnetic trip. Are you postulating a very odd-sounding possibility that an MCB would trip with 35A but not with, say, 240A?

Are you sure there's nothing like a shared line going on somewhere? Two lighting circuits jointed together somewhere or some such?

As sure as I can be. For a start, switching off that one MCB certainly kills the supply to the offending light fitting. Furthermore, it's a very simple CU (IIRC, just 1 x lighting, 1 x ring final, 1 x shower and 1 x immersion), so there's little opportunity for any cross-connections.

Kind Regards, John

 

[AndyPRK]

All that on one rcd?

 

[JohnW2]

All that on one rcd?

Nope - two. Fortunately one serves the lighting and the shower, which made my 'test' simple to undertake.

Kind Regards, John

 

[deadshort]

Shower and lighting ?

Regards,

DS

 

[JohnW2]

Shower and lighting ?

Yep. Are you worried about someone being plunged into darkness whilst in the shower?

I presume that the thinking at the time was an attempt to vaguely balance the potential load between the two RCDs. Although people don't often seem to think about such things, to have the (~35A) shower and a 32A ring final on the same RCD is theoretically 'pushing' a 63A RCD. To have 40A and 6A circuits on one RCD and 32A and 16A ones on the other is relatively 'balanced'.

This CU is actually a bit of an anomaly in my house. Other ones serving small areas around the place have the lighting on an RCBO and the remainder of the circuits all on one RCD. I can't remember why this one was different - I suspect there was probably 'no particular reason' (maybe just a question of what components were 'to hand'!).

Kind Regards, John

 

[JohnW2]

OK - bench testing and autopsy of the B6 now completed.

It operates ‘instantly’ at ~26A (2x3kW loads) - about 4.3In (repeated several times with ‘decent intervals’ in between). The thermal part also seems OK - at ~13A it operates in about 1.5 minutes and with ~8.7A (2kW load, about 1.45In) it operates in about 50 minutes - all more-or-less per spec. Repeated manual operation of the lever works fine.

Internally, everything expected is present and there are absolutely no signs of any problem.

I also forgot to mention previously that whilst still it was still 'in situ', I also tested it by running the shower load through the MCB whilst it was also still connected to the lighting circuit. It again operated ‘instantly’ at this 35A load, thereby dispelling any suggestions of ‘cross-connections’ resulting in the current from the lighting circuit being shared by more than one MCB.

There was seemingly nothing wrong with this MCB. I think I’m left with either an iffy fuse (which blew at a relatively low current) or else a fluky case of ‘extreme non-discrimination’, the fuse somehow operating (correctly) in response to a very high current before the MCB managed to operate. Any other thoughts?

Kind Regards, John