Electric boiler wiring

Certainly the power for the boiler and controls should be on different circuits (just make sure there is the correct separation between the two - e.g. the relay/contactor on the so called "volt free contacts" on the boiler). Introducing an MCB or fuse, by definition does create a new circuit though, where ever it's fed from.

Running heating controls off a fused connection unit (with a 3A or 5A fuse) is pretty normal in the wider world (and often provides a convenient isolation point) - I can't see that using a 6A MCB makes any practical difference (other than easier to reset if there's a fault, but if there is a fault, there would be a lot of investigating to do first so the replacement of a fuse would be the least of your worries). Tapping off the 63A circuit is a bit odd, while there's nothing particularly wrong in principle, 16mm² wires aren't going to be exactly a nice fit into fused connection unit terminals and using smaller cables for that bit will need care and some detailed calculations to prove it's OK (a bit of 2.5mm² T&E almost certainly won't be). A Fused connection unit off a local socket circuit might be a more common arrangement. Or a dedicated 6A circuit from the consumer unit - all seem valid options to me.

Manufacturer's instructions are often a bit, let us say, simplified - perhaps targeting one particular situation and ignoring other valid ones (these seem to presume the "split load" consumer units that were popular in the past, but are falling out of favour these days). We used to be constrained to "follow manufacturer's instructions" but after a lot of complaints about instructions not making an awful lot of sense (or often being originally written for non-UK markets so not considering UK practice and regulations) the regs these days only require that we "take account of" manufacturer's instructions - which gives a bit more leeway - at least as far as the wiring regs are concerned (manufacturer's guarantee dept might take a different view of course).  (The reference to "IEE Wiring Regulations" is a classic indicator of instructions not being well written, copied from something else or not kept up to date - it's been the IET rather than the IEE for several decades now). 

  - Andy.

I think that Andy has covered the situation nicely.

This seems odd: "The 12kW boiler must be installed in premises having a system impedance of not more than 0.1939+ 0.1939Ω."

I assume that means the earth-fault loop impedance at the boiler. The maximum Z_s for a type B MCB is 0.69 Ω (BS 7671 Table 41.3). 0.1939 Ω may not be achievable with a supplier's earth (my own Ze is around 0.23 Ω), and certainly not in a TT installation.

I think that Andy has covered the situation nicely.

This seems odd: "The 12kW boiler must be installed in premises having a system impedance of not more than 0.1939+ 0.1939Ω."

I assume that means the earth-fault loop impedance at the boiler. The maximum Z_s for a type B MCB is 0.69 Ω (BS 7671 Table 41.3). 0.1939 Ω may not be achievable with a supplier's earth (my own Ze is around 0.23 Ω), and certainly not in a TT installation.

The combination  version of the same thing (link) has a similar clause, and I agree many supplies won't make that,  but adds a note about a 30mA RCD. In anywhere but the UK, where high current single phase is more common than 3 phases, the elements would normally be shared around the phases.  The real problem is that once you have a load that big on for some hours at a time, a really large fraction of a 'normal' single phase supply is unavailable for anything else.

A 63A RCBO will be fine for the boiler in terms of overload, and handling earth faults, and to fuse down to 5A locally to create a new circuit  for the controls makes some sense in terms of ease of isolation for servicing.  If the hop of thin wire from the 63A submain to the input side of the FCU  is short (advice is less than 3m) and well contained, so a dead short there is unlikely, then it can be rated for the load, rather than the supply, and be as thin as 1.5mm2. A short length of 4mm2 or 2.5mm2 would feel better however. For a longer length then some adiabatic sums would be needed, and makers data about that RCBO. 

Mike.

Thanks Chris - indeed one of the electricians called the manufacturer to check this point as it didn’t make sense (the Ze was around 0.25Ω from memory). I wasn’t party to that conversation so couldn’t say why but the conclusion was that the manufacturer was “fine with it” being above this threshold.

mapj1 said:

The real problem is that once you have a load that big on for some hours at a time, a really large fraction of a 'normal' single phase supply is unavailable for anything else.

Indeed. 12 kW ÷ 230 V = 52 A. Add one heavy duty shower and you just about have enough left to keep the bathroom light on. :-)

12 kW might seem a lot, but it is right at the bottom end of gas boilers.

Chris Pearson said:

This seems odd: "The 12kW boiler must be installed in premises having a system impedance of not more than 0.1939+ 0.1939Ω."

I've seem similar on a few high-current appliances - I think it's meant to be L-N loop impedance and is to try and ensure there isn't excessive voltage drop due to the high current draw. It's usually written with a "j" on one part (separate resistive and reactive element) so maybe something like 0.27 Ohms overall.

   - Andy.

This raises another question - the manufacturer’s warranty engineer said 12kw seems oversized for the property size (and heating only), so suggested deactivating some of the elements to rate the boiler down to 6kw and see how we get on with that. In hindsight I suspect the installer didn’t do a heat loss calculation and just went for a like for like replacement.

What implications would this then have for, for example for appropriate breaker size if the draw is now halved?

The warranty engineer was referring to deactivating the elements via an option on the boiler’s control menu, rather than permanently disconnecting them. So from a safety perspective there’s nothing to stop someone in the future reactivating all the elements.

What implications would this then have for, for example for appropriate breaker size if the draw is now halved?

Could be "none at all" - i.e. you can leave the existing cable and MCB in place - no harm there. Or it could allow you to reduce things if that was advantageous. It all depends on what the problem is...

   - Andy.

KMA1987 said:

This raises another question - the manufacturer’s warranty engineer said 12kw seems oversized for the property size (and heating only), so suggested deactivating some of the elements to rate the boiler down to 6kw and see how we get on with that. In hindsight I suspect the installer didn’t do a heat loss calculation and just went for a like for like replacement.

A quick flick through Screwfix's catalogue shows that their least powerful gas boiler is 11 kW, so 6 kW implies a very well insulated, or very small dwelling (or a combination of the two).

AJJewsbury said:

I think it's meant to be L-N loop impedance

I see. So, if it is PME, it will be much the same.

0.1939 is clearly not to an accuracy of 4 s.f., but I cannot quite work out what fraction gives that value. (19/98 ≈ 0.1939 (4 s.f.))

Incidentally, is the 230 V +10%/-6% in R. 27 of ESQCR on or off load?

Incidentally, is the 230 V +10%/-6% in R. 27 of ESQCR on or off load?

The local load is not supposed to matter, as this is at the point of metering, and dominated, probably, by everyone else who shares the transformer secondary, not the load of the installation. Of course it does a bit...  especially with a 50A boiler. 
Drops within the installation are of greater concern, especially if, as likely it is soon, we are allowed down to 207V at the origin , as as that could mean as low as 190 or so at the far point in a large building on a fully loaded final circuit - which may be fun for some devices, as the product standards are also usually based on 230+/-10%.

You were not expecting joined up technical thinking were you ?
Mike.

Chris Pearson said:

Incidentally, is the 230 V +10%/-6% in R. 27 of ESQCR on or off load?

I thought both - it not being 100 miles away from 0.35Ω * 100A gives 35V ... whereas 16% of 230V gives 36.8V.

Or to put it another way if you had simple system feeding just one consumer and a constant 253V at the transformer and the DNO's lines had a resistance of 0.35Ω - the consumer would see 253V (230V+10%) at their supply terminals when drawing no load, and 218V (i.e. just a tad over 230V-6%) when drawing 100A.

    - Andy.