I have a small (only one 16A circuit) consumer unit piggy-backed onto the main CU. There is only a 3/8" gap between them so this could easily be sealed. The secondary CU is really only there because RCBOs are not available for the main unit, to provide a circuit not subject to all the nuisance trips the main one suffers.
The problem is that the secondary CU is connected at the incoming main switch terminals of the main CU by single 6mm wires. This is plenty for the load, but strikes me as a bit small for the potential fault current from the 100A main fuse.
I don't think two 25mm wires will fit in the main CU switch terminals - what is the minimum I could reasonably use?
(I know I need a new CU, but I fear the results of a general inspection of the wiring and I don't know where I can find an electrician who will change the CU without such inspection. I am cautious about doing it myself.)
IANA electrician, but what conditions are you concerned about drawing an excessive fault current? If you have a fault downstream of the secondary consumer unit, the 16A RCBO should trip. That only leaves the wiring between the secondary CU and the main switch. Is this wiring visible for inspection? How likely are you to get a fault in this wiring?
It would only need to carry a 100A fault current if somehow this cable were to become faulty. Is that risk manageable by inspecting the cable and making sure the terminations at each end are secure?
If you tried to stuff in another chunky wire into terminals not designed for it, maybe that would make the terminations worse (ie more liable to be high resistance and overheat) rather than better?
I don't think someone will do a CU replacement without doing wiring checks. If you did it yourself you'd need building control signoff, so same difference.
If you got an EICR, I don't think you'd be obliged to do anything about the results (unless it was a rental property or something, or a danger to the network). Then you would know the situation. If it's not actively dangerous (C1) then you wouldn't have to remedy it, although it is good to address C2s.
I think the simplest solution is to remove the 6mm cables from the main switch terminals and wire the secondary CU from one of the spare ways (if you have any?). Then you can use an appropriate MCB to protect those wires and everything down stream of that.
The secondary CU already has a modern, type whatever RCD that has never tripped except when tested. But *all* the main CU circuits are protected by a single RCD, so there is no nuisance trip free way on the main CU to connect the secondary CU to.
The thing to do would be check by calculation... There are two classes of protection that the wires need - overload and fault (or "short circuit") protection. The overload is handled by the downstream MCB and the nature of the load. So now you only need worry about the (very very slim) chance of a fault between the secondary CU and the main one.
So start with a guestimate of the supply impedance. If it is a TN-C-S head end (PME), then in the absence of an actual measurement it is normal to take 0.35 ohms as a fairly pessimistic worst case, and lets assume the length of the 6mm tails is negligible in contribution to the loop impedance. So that gives a prospective fault current of 230 / 0.35 = ~ 660A
If you look at the trip curve for a BS88 incomer fuse, that looks like about 2.5 sec to clear a fault:
formatting link
So plug those details into the adiabatic equation and work out the cross sectional area of copper required to withstand the fault current:
s = sqrt( I^2 x t ) / 115 = sqrt( 660^2 x 2.5 ) / 115 = ~9 mm^2
So at that supply impedance the 6mm tails are a bit undersize, and so would likely suffer some insulation damage[1] due to heat while blowing the main fuse. However we are well short of the burst into flames, vaporised copper, arc flash territory!
If the supply impedance is lower (quite likely) - say 0.2 ohms, then that changes the sum such that your fault current goes up to 1150A, which should blow the fuse in 0.2 secs, you would only need ~ 4.5mm^2 tails to be well in spec.
[1] which may be a moot point since most causes of a fault current in the tails are likely to be the result of damaged insulation anyway!
One would normally use a service connector block (aka "Henley") to split the tails.
It is not a particularly difficult job if you plan ahead.
formatting link
It sounds like you already have RCDs (the nuisance trip comment!) so at least there aught not be too many surprises like borrowed neutrals to catch you out on the new board.
If it were my installation, I would probably change the main CU at some point, but would not be unduly concerned about the CU tails.
(one hopes that if installed by an electrician, they would have measured the prospective fault current at the CU and chosen tail sizes appropriately anyway).
While reading about something unrelated I happened upon this discussion:
formatting link
I'm not familiar with the adiabatic equation, but it seems like it covers that situation. The OP does some calculations for their circumstances and comes up with:
"So when testing Zs on the main tails (distribution circuit) connection to a CU supplied via a ≤100A BS 1361 / BS 88-3 main fuse, and the earth conductor is separate and not bunched (143 k factor), then as long as the Zs is ≤0.27Ω (as required for that fuses max Zs), the main earth conductor must be the largest CSA of the following, as applicable:
6mm2 minimum (to comply with the adiabatic equation).
2.5mm2 minimum (if not part of a cable, but has a sheath for mechanical protection).
10mm2 minimum (for ≤35mm copper PEN if a TN-C-S system / PME).
XXmm2 minimum (if buried in ground).
equal to line conductor ≤16mm2 (for not be less than the value determined by selection from table 54.7)
that’s 10mm2 minimum (from the line & neutral adiabatic example above).
Or 16mm2 minimum for line conductor >16mm2 and ≤35mm2."
which suggests 6mm2 might handle the situation. But I'd have to understand this a bit more and do the calculations myself to satisfy myself.
Fair enough. If you are in danger of incrimination I might not choose to add further charges with more DIY :)
But you still have the problem of calling someone in or, if doing it yourself, working live.
In my case, I added an external isolator before the CU, to make future work easier, if I ever need to add a second CU. As my CU is plastic, I felt safe enough removing the tails, one at a time, from the CU and putting them into the new switch - which could be done outside its housing and then clipped in afterwards.
Is it not considered necessary to work out the adiabetic equation for finite resisance faults rather than short circuits? I presume they either don't happen or melt if they do.
We're 50m from a fairly hefty pole transformer supplying five houses, a PME supply. From what you say 10mm might be fairly reputable. Should the 'earth' wire be the same dimension, I presume it could carry the whole fault current?
Thanks for the advice. To be honest I think the house should probably be rewired when I leave it. The original was at best amateurish and 40 years ago. I might give the new CU a miss.
Avoiding too much nuisance tripping on the fridges, computers and boiler is enough for me of an achievement for me, and I believe that is why the secondary CU was probably fitted originally.
The one good bit of my installation is that someone did put an isolator after the meter and before the Henley blocks (and before the Economy 7 switch), I wouldn't attempt it live.
The short answer is no. Faults are assumed to be a short circuit. A high- resistance fault would heat up much more than the (low-resistance) cable, so the fault would burn itself out (or become a short circuit). And a low- resistance fault is close enough to a short circuit to make no difference. A fault that draws a current significantly higher than the normal maximum load but lower than the current to blow the fuse, and is stable for long enough to risk damage, more or less never happens.
Or put a bloody great cover over it and hide it. I'm sorry to be flippant, but it seems to me that what is legal one minute can be made illegal by somee body making regulations not applicable in the case you have. Does it not make a mockery of the whole thing that you can just by endless extension sockets and plug them in. Obviously on my old CU its just circuit breakers and nothing fancy, so the circuited will trip if overloaded. Brian
Bollocks. It simplp renders the circuit essentially dead.
And a low-
Oh, it surely does...
As we found out when our whole office went dark during a powercut, and when it came alive again the inrush of all those CRT monitors blew the main 60A fuse..
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.