New wiki article - diversity

[John Rumm]

adapt adopt improve as usual:

[Roger Mills]

Oh, *that* sort of diversity - rather than the usual PC clap-trap!

[Andy Wade]

Perhaps the title should be Diversity (Electrical).

Also:

- I think John needs to make it clear that the first table is used for estimating the load for a particular final circuit and that the second table is for the diversity for a group of circuits fed by a distribution board (or, in the case of a house with a single consumer unit, for the whole installation);

- "If the Power factor of the lighting units is unknown, then use the VA rating times 1.8", err, no, use the rated lamp watts times 1.8 (and divided by 230 V). If you had a VA figure for a luminaire there'd be no need to multiply it by 1.8 :~)

[John Rumm]

Yup will do...

Yup, so much for shortening note 2 of table 1A - made a mucking fiddle of that! ;-)

[John Rumm]

Yeah well bollocks to PC ;-)

(did think about including "electrical" in the title, but could not think of any other type that would apply to DIY)

[Andy Wade]

More suggestions, based on the page as it stood a few mins ago:

I can't see why that shouldn't also apply to mixed tungsten and discharge loads. (It would err on the safe side re phasorial addition in in-phase and lagging currents.)

Also the meaning of "discharge" might not be clear to some of the audience. I suggest "fluorescent" or "discharge (fluorescent)".

Suggest: "for fluorescent and similar fittings use the total VA load rating (if known). This allows for ballast losses and power factor."

"If no VA load figure is available simply use 1.8 times the total nominal lamp wattage." (It's lamp, not fitting, since the 1.8 factor includes an allowance for ballast losses.)

Delete that. The wording suggested above makes it fairly clear that this is to get a 'presumed' VA figure, I think. Certainly don't talk about "peak current" - that could have two meanings: peak of waveform as opposed to RMS, or maximum value of a time varying current. Both are red herrings in the present context. "Full load current (FLC)" or "maximum demand" are the pertinent terms, the former for a point of use or a final circuit and the latter for the dis-board/CU/whole installation.

Agreed. Although of course if you know they've got modern HF ballasts the PF will be very close to unity and the tubes will be running at about 50 W, so the FLC is much less - more like 2.2 A. Best to err high though...

Suggest you add an example for the second table too, say two lighting circuits, two rings, immersion heater and cooker. This is to get across the point that you use the design currents of the final circuits (with diversity where allowed) and not the MCB ratings. Something like this, perhaps:

--- --- --------------- --- --------- ---- Cct MCB Load type Ib Diversity ADMD*

--- --- --------------- --- --------- ----

1. 32A Cooker (12kW) 23A 100% 23A
2. 32A Ring 1 (kitchen) 32A 100% 32A
3. 32A Ring 2 32A 40% 13A
4. 16A Imm htr (3kW) 13A 100% 13A
5. 6A Lights (1kVA) 4.3A 66% 3A
6. 6A Lights (600VA) 2.6A 66% 2A . --- . Total ADMD 86A . ===

• ADMD = after diversity maximum demand

[Robin]

Helpful as ever.

One question. Would it be practicable and helpful to say something about the way the maximum demand often ends up greater than the supply fuse even after applying the factors in the 2nd table? That did my head in for a while; and I'm still unclear what the regs. require/permit by way of design. At the risk of displaying yet again the depths of my ignorance I had in mind something vaguely along the lines of

"Even after making the allowances for diversity in the second table the maximum demand on a consumer unit often works out as more than the supplier's main fuse. This is not necessarily a cause to worry. The allowances for diversity are conservative. [But it may help to use a split load, dual RCD consumer unit so the maximum demand on each side is no more than their rated 80A or 100A; or to use 2 consumer units so the maximum demand on each is within the (usually) 100A capacity of the main switch.]"

ISTR finding the advice in square brackets but I'm not sure how the first works if there is still just one main switch; or how either works if the whole thing hangs off a 100A REC. In other words, I'm not waving but drowning.

[John Rumm]

Yes, perhaps "discharge (Fluorescent tubes, Sodium / Mercury Vapour lamps, or any lamp technology that requires a current limiting ballast to operate)"

a bit less chewy, ta.

As the lamp wattage is not so intimately tied to the length of the fittings these days, that is better anyway.

Its surprising just how many recent off the shelf fittings still have magnetic ballasts though. Not sure if the new rules on PFC will affect that though?

I wonder which is cheaper to make, HF ballast with built in PFC, or mag ballast, and separate cap?

Yup, very good idea. I realised he wording allowed for ambiguity as to which current one uses, and was thinking about adding a note, but an example will do nicely ;-)

[John Rumm]

Yup, I think Andy''s point about adding an example to the second table ought to address that.

There is as written (for the next few minutes anyway ;-) a avenue for misinterpretation (its in the OSG as well in fact) that does not really make clear which numbers you use on the second table. If you just apply the second table rules to a list of MCB values, then you will almost certainly end up with a number way too large.

So the first table helps assess the design load of a circuit, and then the second one applies further diversity to a group of circuits.

I can see the logic of spreading circuits out over several RCDs (for many reasons rather than just load), but the point that they tend to have smaller maximum switch ratings is also valid. Not sure that the logic really applies to split load in general though, since the same main switch typically switches both sides at once.

Have a read again after the next update and see if it is any better...

[dom]

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D\

=A0 =A0 =A0 =A0|

=A0 =A0 =A0 =A0|

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D/

I'm assuming when you mention sockets other than 2A lighting and 13A power, you mean things like 15A round-pin, and 16A commando - where they're occasionally used in domestic settings?

What about 5A round-pin lighting sockets? Hardly appropriate to rate the 2A version and 0.5A load (115W) and the 5A at 5A (1150W).

[John Rumm]

I believe that is the intention, yes.

The OSG does not mention them explicitly. I expect this is a case rather like with 15A round pin and 16A commando connectors mentioned above, where you need to apply actual knowledge of the installation to come up with a sensible answer. For lighting purposes, 2A sockets would seem more than adequate, so its harder to think of applications for 5A ones that you can make blanket assumptions about.

[The Medway Handyman]

EMT's get 3 days training on delivering babies, but 5 days on 'diversity awareness'.

[Andy Wade]

For 2A or 5A on a lighting circuit I'd treat as per any other lighting point - 100 W per outlet - unless indicated otherwise. No-one's going to plug any other appliance into them these days.

I saw some guidance on IEC 60309 outlets somewhere recently (might have been in /Wiring Matters/). One or two 16 A outlets per 16 A circuit, AFAICR, again in the absence any contra-indication. They're unshuttered of course so aren't allowed for general use in a domestic installation - OK if out of reach of younger children in a workshop though, IMHO.

[dom]

Commando connector hook-ups for caravans camper vans whilst at home is the more common usage I've seen - and soon, no doubt, electric car charging points.

Advice about the latter here:

[Robin]

I like the example but it doesn't really deal with the point I had in mind

Yes thanks - I made that mistake early on but have got beyond it.

It is getting better and better. But I am still left with the problem that some modern installations - especially but not only where people cook electric and have an electric shower - seem to have maximum demands in excess of the fuse/isolator. I am sad enough to have looked at friends' and neighbours and tried to do the sums (making best guesses about numbers of lights etc). Several end up well over 100A.

I don't think I've I mucked up the sums. Mine look much like the example in the Wiki. (Indeed, taking the example there, if the owners asked for an electric shower and there was a spare way in the CU should they be told it's not practicable because it'd take the maximum demand over 100A?)

Of course they may just be bad installations. But I am still left wondering about the OSG. I say that in part because I have now tracked where I read about it being conservative and about using a split board. It was in "Advanced Electrical Installation Calculations" published by Newnes. They had an example where the MD after applying all the OSG factors was 143A. And the text which followed was:

"In this case as in many other cases, the assumed maximum demand is still greater than the supply fuse.You will find that the supply fuse has been in place for many years and never been a problem and should not give rise for concern. This is because the diversity allowed is usually quite a conservative figure. The greatest problem with this situation is that the switched disconnector is usually 100A. In these cases consideration should be given to using a split load board with possibly the power circuits on an RCD protected side of the board."

I'm sorry if I'm flogging a dead wotsit. Should I just give up worrying on the basis I'll never learn the guild's secret handshake either :-) ?

[Andy Wade]

I was avoiding the issue, although the type of installation it describes would often have only a 60 A main fuse.

I think the underlying problem is that the OSG guidance is out of date. I don't think it's changed since the 14th edition, possibly even the

13th! Modern installations tend to have more final circuits, an approach encouraged by the 17th edition and RCD-related issues.

To a good approximation domestic MD is probably more related to the size of the house than anything else. Normal size houses (up to 4 or 5 bedrooms, say, seldom seem to have any problems on a 60 or 100 A supply (cannabis farms excepted).

[ARWadsworth]

It would not be unreasonable to bring down the 40% value on remaining socket circuits. I don't think that I could load my upstairs sockets to 13A if I plugged in all my electrical appliances that I keep upstairs.

A 4/5 bedroomed house with teenagers having showers and using hairdryers whilst dinner is been prepared and the washer and tumble dryer are on may draw over 100A for a short time, however I doubt the main switch or fuse will have a problem with that - we are talking a few minutes here not hours.

[John Rumm]

I suspect Andy chose values deliberately to avoid the question! ;-)

I might go and add another circuit to deliberately provoke the point!

I think this is one of those occasions where you need to apply local knowledge to override the guideline. For example, the calculations on socket circuit loadings are pessimistic in two ways... Firstly in terms of load per circuit; where with the possible exception of kitchen circuits it is going to be exceedingly rare to reach 32A of load. Secondly, the calculation for load over the remaining circuits is also rather restrictive... saying 100% of first, 40% of second - is ok IMO, but then I would tend to ignore all the remaining socket circuits or count them at a nominal values based on actual load rather than also add

40% of those.

So in my house for example I currently have 4 ring circuits protected by

32A MCBs. Kitchen / utility room, might be able to get close to that - but probably only by including very short term loads like kettles etc. Ground floor ring will typically struggle to reach 10A, and only then manage it when running the vacuum for example - the rest of the load is electronic niff naff. Upstairs will be even lower. The workshop one could potentially get up to about 25A load in the winter with some heating running, dust extraction, and a power tool - but even then would have difficulty sustaining it.

I still don't follow the logic of the advice - since split load boards will typically connect all "sides" back through the main switch. Splitting across multiple CUs however makes more sense.

No I quite like it when folks do flog dead wotsits, since it tends to elicit the sort of answers that are not typically provided in the more formal publications, and yet are known as a matter of lore to some of the trade. After all the purpose of the wiki is not to save people buying the OSG etc, but provide the info required by DIYers in a quickly digestible form with practical advice in there as well as the theory and the "legality".

[John Rumm]

OK, I have significantly extended the bit at the end to better cover the cases where you have apparent overloads even after applying diversity.

[Robin]

I'd be relaxed about that approach for a Scout badge but as the basis for a BS which has statutory force by virtue of the building regs. I'm unimpressed by the combined might of the IET and ODPM as was. Anyone installing a new CU under a building notice and not working within the very conservative OSG seems to risk a nasty surprise from the appointed inspector.

Yes, thanks. (I started worrying about this after splitting a downstairs ring into 2 rings - one for the kitchen and one for living room and hall. It seemed bizarre that as a result of that alone my MD had increased beyond the meter's rated capacity.)

Many thanks. (And if I may say so I like very much the way you bring the total in just below 100A.) I'll shut up now.