Diversity and or multiple CUs

[Fash]

Just sorting out the calculations for the rewire now that the Part P route is agreed with building control, and have come across a problem(?) with specifiying the CU. Basically given that I'm not supposed to allow for diversity at the CU I can't wire the house on a single consumer unit. Any electrical installation experts advice on reasonable cost solution very welcome. The reality is that I'm never going to exceed the 100A supply as no electric showers and immersion heater is emergency only if the boiler (in fact both boilers) breaks down (pretty unlikely) and no electric heating apart from comfort undertile heating in the cellar.

I want to wire the house in 4 'zones' with the following circuits:

Zone 1 = 2nd floor - 1x socket ring (32A) + 1x lighting (6A) Zone 1 = 38A total

Zone 2 = 1st floor - 1xsocket ring (32A) + 1x lighting (6A) + central heating control (6A) + immersion (13A) Zone 2 = 57A total

Zone 3 = Ground floor RHS - 1x socket ring (32A) + 1x lighting (6A) Zone 3 = 38A total

Zone 4 = Ground floor LHS - 1x socket ring (32A) + 1x socket radial (20A) + lighting (6A) + range cooker (32A) + refridgeration circuit (6A) Zone 4 = 96A total

Zone 5 = Basement - 1x socket ring (32A) + 1x lighting (6A) + kiln (16A) + tile heater (13A) + smoke alarms (6A) Zone 5 = 71A total

Nominal Installation Total = 300A

Clearly this is ridiculous given that the total number of circuits is

18 the realistic diversity is very high. The only way I can see to get the nominal total down is by combining circuits but this still won't get me all the way as the house is 270sqm so I need at least 3 ring circuits.

My original plan was to use an MK sentry CU with 21 ways (19 usable) and use separate RCBOs for the socket circuits to avoid the risk of nusiance trips from earth leakage on electronic kit (PCs etc.). This would have the required number of ways and is in reality adequate for my needs.

To meet the supposed design requirements I would need to use 3 CUs fed from a Henley block which I can do but this seems like overkill. The cost difference isn't huge it's just the hassle factor and having to run meter tails over longer distances to the 3 dis boards.

Any advice on what is reasonable/acceptable in practice.

Sorry the post is so long but it's a big installation!

Fash

Zone 5

[James Salisbury]

Takeing a look at The Electrcians guide to wireing regs by John Whitfield

You have 5 rings 5 lighting ciruits one radial 20A one cooker one killn one tile heater and one imersions

so rings are you realy going to have 12 fan heaters pluged at the same time ??

The book suggests 85A for power

16A for the cooker the water heater kiln and tile heater can not have diversity applied too them so 42A what is the wattage of these appliances??? 19.8 for lighting (is this a light house?) 1A for fire and smoke alarms the Fridge must not be on a 6A breaker, it will trip due to the starting current, run the circuit in 2.5mm and fit a standard 13 socket and 16A breaker ( so you can plug the kettle in when the RCBO has taken the kitchen out)

SO

Suggested total is 163.8A however this is a little excessive you are unliklly (very!) to have a main fuse that is over 100A

I would suggest that you take a look at what you are going to use and say put the cooker kiln tile heater and imersion in one 80A CU and the other in the rest and fit a 100A fuse switch to keep it legal

page 26 part 110M

When you starts blowing the electrcity companies fuse then think of upgradeing, this will cost £££££££££

total

[Owain]

Okay, but I would not put the immersion heater on the "basement" CU as this might be confusing, and I would put the smoke alarms on one of the house lighting circuits as it's more obvious if the circuit fails then.

Having 2 CUs has the advantage that if max demand is too high for the

100A incomer it is now fairly straightforward (perhaps not cheap) to bring in a second phase for the basement.

Owain

[dom]

p.86/87

Allowance for diversity,

1. Lighting - 66% of total current demand.
2. Standatad arrangements of final circuits - 100% of largest circuit +

40% of every other.

[Fash]

Thanks for the advice, so you reckon it is worth going with 2 CUs rather than 1? Would you just put them next to each other? If I put them next to each other then this gets round the problem of running long tails and needing switch-fuse units next to the meter for the remote CU, plus then the fact that the CU loads are not split geographically doesn't matter.

As I said before the reality is that the diversity on the socket circuits is much greater than the rule of thumb in the regs and since this is guidance it's not really an issue.

Since I haven't blown the company fuse yet and in the current installation I do have an electric shower I'm not particularly worried about this. Certainly wouldn't want to get a bigger supply in since it took me 18 months to negotiate to get rid of my 1920's overhead TT service and have it replaced with an underground TN-C-S!

Take your point about the fridge, I was always going to run as a 2.5mm radial but wasn't sure on the breaker size since they could all be hard wired in which case it wouldn't be a socket circuit at all. Some other comments/questions below.

Kiln is 3.2kW so 14A and recommendation is for 16A supply. Clearly not diversified but pretty low usage. Tile heater depends on whether it's just comfort heating or whether I spec it to provide full heating given that I will need to apply some heat all year round when the rest of the house won't be heated. The manufacturers suggest it can be run off a fused spur (provided it's 1A for fire and smoke alarms Haven't checked this yet but Fire Brigade recommended putting in 8 smoke alarms (interlinked)! Sounds excessive but is because the property was originally 2 houses (until 1862) so ceiling heights don't match across floors leaving dead areas that require additional alarm points.

Fair enough as discussed above.

Main fuse is 100A and all nice and brand spanking new, including tiny digital meter with customer accessible tails (to CU not main fuse obviously). So I can turn it all off back to the meter without needing my friendly electicity board to remove their fuse. I do have a spare main fuse for the first time I blow it anyway!

If I put them next to each other and run to each from the Henley block (about 500mm away) can I ignore the switchfuse?

Never going to happen.

[ironer]

There's more information on this general topic, and much else, relevant to the OP's task in

Institution of Electrical Engineers Wiring Regulations: Commentary on BS 7671 (IEE Wiring Regulations) (Hardcover) by Paul Cook

Hardcover: 411 pages Publisher: Institution of Engineering and Technology (April 1997) ISBN: 0863413161

[data from Amazon.co.uk]

Price 50smackers new, or 85 old!

My copy is out of reach just now, but ISTR the book indicates a 240% allowance for diversity on a 100A supply (though, no doubt with qualifications that I forget just now).

The book is written from a more theoretical electrical engineering approach than the On-Site Guide &, as the OP says in a parallel posting that he's an elec engineer, it may not be such a daunting read as it was to some not far from here.

In doing what the OP is setting out to do a few years ago I found that no single book covered all the topics needed. Some topics are just not adequately covered in any of them. Would suggest for starters the OP needs to check out the following books: Whitfield mentioned by another poster; IEE On Site Guide; Cook's book above & several books by B Scaddan; plus a look in the local library.

Invaluable was "A guide to selection of electrical cable" by Kevin Boone - on web at

all the books are short on applicartion of the many rules & practical considerations as to routeing cable in a domestic building - topics such as cable support, bunching, route through joists and walls, presence of insulation and so on.

HTH .

[ironer]

/snip/

The reality is that I'm never

/snip again/

/snip/

The underlying problem here is that a UK ring circuit depends on statistical chance [aka 'diversity'] for security against overload.

Each socket has an uncertainty as to its load - it may be over or under norm. String the sockets together in a ring (ie in series) & the uncertainties (over or under) tend to cancel out, keeping the circuit maximum load within bounds.

Divide the single ring into 2 separate ones however and loading uncertainty no longer cancels out in quite the same way. Hence a large number of rings, being in parallel, tends to escalate the nominal total load even if the same applicances are going to be used either way. Which is as you have found.

One way out of the dilemma is to consider statistical loading from a higher viewpoint - ie that of the whole house rather than individual rings. The basis of this is already in the diversity rules (100% of

1st ring + then 40% of subsequent circuits). You need to justify reducing the 40% and perhaps the 100%. IIRC there is a clause somewhere (maybe in the On-Site guide?) allowing a suitably qualified elec engineer to make that sort of allowance in design calculations.

IIRC the topic is covered partly in P Cook's book " Commentary on BS

7671 (IEE Wiring Regulations)" (see my other post).

In the case of high power fittings such as showers, washing machines, and so on, it may be that you can argue they are only used for short time spans intermittently over 24 hours & substantial diversity reductions can be made. Cook's book explores something of diversity from this angle. Also try Googling this group on the subject of cooker wiring and fusing. IIRC there's been a few queries on this & the related subject of diversity needed.

I struggled in this area, so would be most interested to hear how you resolve the issues: please do post the details of your solution.

HTH

[Fash]

I've spoken to a number of CU manufacturers and advice is pretty similar from all. Since the supply fuse is 100A and a 21 way/18way CU is also 100A there's no point having 2 since it just moves the bottleneck. Also clear that it's not more dangerous to have more rings given that it has no real effect on total load in the house. The way forward is to justify increasing the diversity based on the fact that the installation loads are the same as if I had only 2 rings. Since the diversity is advised not absolute provided it is justified it is OK. In my case the absolute minimum rings I could have is 4 based on floor area (unless I started mixing floors which would get it down to 3). Advice is that it's better to be correct on the ring circuits and worry less about the diversity issue at the CU. So it's back to a 21 module MK Sentry. They all validated my decision to use single module RCBOs rather than a split load board as with 5 rings I would only need

6mA/ring to get a nuisance trip on the RCD. THis isn't a problem now, but once the kids all get playstations and PCs and other things with switch mode supplies the chances of tripping a single RCD over all the circuits is quite high.

Fash

[ironer]

TFT info: illuminating.

Also clear that it's not more dangerous to have more rings

Not sure I'd say it was more dangerous as such. Provided suitable MCBs are in place each ring will current limit safely. But what is happening is that as the no of parallel rings increase, the chance of one or more ring becoming overloaded statistically increases. Which means more chance of an overload trip. String the sockets out in series and the individual risks tend to cancel out.

However despite the statistics I agree that multiple rings are preferable. This is especially so if the effect is to limit the average current in each ring. That does 2 things -it keeps I^2R losses down & reduces voiltage drop so you get to use more of the KW-hrs you pay for. Downside is that it may increase the length of cable used which will have the opposite effect.

When I did my rewire I kept a check (via a spreadsheet) on in-circuit I^2R losses (and annual cost) as well as voltage drop. In some cases it can pay to use a larger cable to cut the loss, OTOH cable larger than 2.5mmsq is hard to fit in most ring sockets and fittings.

[ironer]

plus it gives you a power source for maintenance on the other CU & circuits & vice versa.

[James Salisbury]

There is still the small issue of if you stuff all those MCBs into one CU of how you will provide overload protection to the CU. You should not rely on the DNO cutout (meter fuse) as it is not your property..

[Andy Wade]

Two points:

1. For a given number of final circuits and given set of diversity assumptions the ADMD (after-diversity max. demand) of the installation will not depend on whether one, two or N consumer units are used. For