RCD blowing cutout fuse

[westie101]

Though it's at least done one part of it's job - protect the person operating it from injury!

Yes looks like an internal L - N short, possibly small clearances between the two were part of the cause

 

[RF Lighting]

Looks like it certainly did go with a proper bang

 

[JohnW2]

Conjecture before has been some form of build up (conductive soot/whatever), then a bit of arcing as the switch/breaker operated was the last straw - followed by a flashover with the deposits of soot/carbon acting as the initial conductive path until a good ionised air channel has developed. Once the arc is going, then it's going to self sustain until "something blows" upstream. So the fact that it's an RCD being tested is not relevant, it's a switch breaking a load.

That certainly sounds very credible, albeit I'm assured that there was virtually no load on the CU when the RCD operated. It certainly looks as if the bang probably originated in the vicinity of the main switching mechanism. If that 1.8KΩ resistor (or, indeed, any of the other components on the PCB) is, indeed, the 'test resistor', then it certainly was not current through that which caused the problem.

Mind you, I'm not sure where this "conductive soot/whatever" is meant to have come from. The device certainly seemed to be well sealed, so I doubt it could have come from outside.

If you clean up the parts, is it possible to determine where the arcing occurred ? For something to take out the service fuse, I'd expect there to be "significant" arcing damage visible.

Yep, that's the next step. However, like a true pathologist or forensic scientist, I did not want to do any 'cleaning up' until I had fully documented (i.e. photographed) the situation 'as it was'.

Kind Regards, John

 

[JohnW2]

Though it's at least done one part of it's job - protect the person operating it from injury!

I suppose it's really the cutout fuse which did that. There clearly was quite a lot of current involved, and had that been allowed to persist, I would imagine that there would have been a lot more damage, if not a fire, in the CU.

Kind Regards, John.

 

[JohnW2]

Looks like it certainly did go with a proper bang

I am assured that it did -it certainly frightened my daughter!

Kind Regards, John

 

[bernardgreen]

Is that a cut in the insulation of the red lead ? If that is a cut due to part of the mechanism having been pressed against the cable then was the metal of the mechanism live and did this short to the neutral somehow.

 

[JohnW2]

Is that a cut in the insulation of the red lead ? If that is a cut due to part of the mechanism having been pressed against the cable then was the metal of the mechanism live and did this short to the neutral somehow.

As I said, there appears to be damage to the insulation of both main conductors - and, as you say, interaction with the switch mechanism could be a cause of that. I'll soon be trying to clean it all up, and hopefully then may have a clearer idea of what may have happened.

Kind Regards, John.

 

[JohnW2]

Is that a cut in the insulation of the red lead ? If that is a cut due to part of the mechanism having been pressed against the cable then was the metal of the mechanism live and did this short to the neutral somehow.

It seems that you were right. There is a very clean cut (as if produced by a very sharp knife) through the insulation of the main L conductor. When one folds back the flap of outer sheath so created, one sees that the damage goes right down to bare copper (see pics).

As you say, one can but presume that the conductor somehow got caught up in the main switching mechanism, which 'sliced through' it when the RCD operated. I haven't yet been able to work out what that exposed bare conductor came into contact with, but it clearly must have been something at N potential. Maybe whatever 'sliced through' the cable (one imagines probably something metal, judging by the cleanness of the cut) was itself at neutral potential.

This obviously should not happen, and suggests that some imperfect assembling of the device resulted in the conductor being in a vulnerable position. Why it took~12 years, and dozens of RCD operations (during testing) before this catastrophic failure occurred is anyone's guess. I suppose earlier operations of the RCD may have have progressively moved the (imperfectly positioned) cable nearer to a position of vulnerability. Apart from periodic testing of that RCD, the CU had not been 'touched' in any way for a number of years.

The pics:

Untitled

• JohnW2
• 21 Apr 2013
• 2

Untitled

• JohnW2
• 21 Apr 2013
• 4

Kind Regards, John

 

[bernardgreen]

The damage to the live conductor at the point of the cut insulation looks too small to have been involved in the current needed to blow the mian fuse. Maybe there had been several shorts at that point when the unit was tested ( or was tripped ) and the resulting soot and vapourised copper built up until there was enough to cause the major short.

Is the sensor coil and the damaged Live on the IN ( supply ) side of the contacts or the OUT ( load ) side ?. If on the OUT side it would explain why the damage at the Live wire is mininal and the short there did not take out the supply fuse by itself.

It might be worth contacting the manufacturers to make them aware of the failure.

 

[JohnW2]

The damage to the live conductor at the point of the cut insulation looks too small to have been involved in the current needed to blow the mian fuse. Maybe there had been several shorts at that point when the unit was tested ( or was tripped ) and the resulting soot and vapourised copper built up until there was enough to cause the major short.

Yes, that all makes credible sense. One confusing factor is that there really does appear to have been a minimal load on the CU when the test button was pressed - as far as we can make out, literally just a small fridge, small freezer, combi boiler electronics and two or three wallwarts. I'm a little surprised that breaking such a current would be enough to precipitate what happened.

Is the sensor coil and the damaged Live on the IN ( supply ) side of the contacts or the OUT ( load ) side ?. If on the OUT side it would explain why the damage at the Live wire is mininal and the short there did not take out the supply fuse by itself.

It is, as you, surmise on the load side of the 'switch'.

Kind Regards, John

 

[bernardgreen]

The amount of load will have no effect if the fault was an arc from Live to Neutral via soot and/or splattered copper deposited around the contact assembly.

It appears that the supply to the PCB board is from the sensor side of the contacts. In which case the supply to the PCB will cease when the contacts start to open. If the capacitor on the PCB has gone low value it might result in the trip coil losing power during the operation of the mechanism. Whether this would result in slower opening of the contacts and hence prolonged arcing is another thing to think about.

 

[Spark123]

Contacts are usually spring loaded on some sort of a latch so open rapidly, doesn't look like the PCB has anything to do with it. I don't think you're a million miles away with something going wrong in the arc arresting chamber causing an arc, ionisation and eventually it to go with a big bang.

 

[bernardgreen]

I assume it is a spring driven opening but I was looking at all options as I find it hard to believe that the design would allow copper spluttering ( from the damaged Live ) to get that close to the contacts.

 

[JohnW2]

The amount of load will have no effect if the fault was an arc from Live to Neutral via soot and/or splattered copper deposited around the contact assembly.

True, but something associated with the operation of the device must have initiated the arcing, and I was hypothesising that this something could have been a small arc at the opening contacts.

It appears that the supply to the PCB board is from the sensor side of the contacts. ...

Yes, as far as I can make out, virtually everything within the device is 'on the sensor (load) side' of the contacts - which are very close to the supply terminals.

... In which case the supply to the PCB will cease when the contacts start to open. If the capacitor on the PCB has gone low value it might result in the trip coil losing power during the operation of the mechanism. Whether this would result in slower opening of the contacts and hence prolonged arcing is another thing to think about.

Not so sure about that. I would have thought that the trip mechanism was simply yes/no and that, once triggered, the speed of opening of the contacts would be down to a spring. Until the main contacts opened, the PCB would remain powered, and hence not reliant on a power-sustaining capacitor.

Kind Regards, John

 

[Spark123]

I suppose if the spring was weak and the contacts not forced apart far enough then the arc may have not been passed to the arc chamber - from the original pics the arc chamber looks relatively clean.

Another option is that the cable (connected to one side of the contact) passed closer to the other side of the contact than the contact opening, the insulation being damaged caused a shorter air gap hence the arc not being passed to the arc chamber. When the contacts opened an arc formed between the two closest points and didn't quench.
That would be my favoured guess now.