Voltage drop again

[JohnW2]

Obviously you cannot have a ring final to a kitchen 100 metres from the CU.

Probably not, but if the kitchen were 100 metres from the origin of the supply, you'd have to have either a 100m radial or a 100m sub-main - and, as I've said, the difference is essentially semantic, since either would have to satisfy the VD requirements.

Kind Regards, John.

 

[sparkticus]

I agree - the only reason for a downrated OPD would be to 'achieve compliance' (VD-wise), and I personally think that would be a bit silly. However, that takes me back to the original question - how would you feel about a (very!) long 4mm radial with a 20A MCB which only satisfied VD requirements with, say, the 10A total load which the designed regarded (currently) as the 'probable load'?

Oh - hummmm (as he shuffled uncomfortably around in his chair) Was it Mr Einstein who said that god does not play dice then realised that maybe he did?

Assuming a socket then I would "probabilistically" feel that the telephone call after plugging in an electric kettle or high-wattage vacuum cleaner (plus the probable load) would be somewhat inevitable within a week or four. In the commercial world where it might be seen as acceptable to have a label on a socket saying something like "maximum load not to exceed 2.3KW" then maybe OKish but a bit of a stretch in my book.

I see the following escape routes (some of them not really acceptable):

- fixed load (FCU)
- Label in commercial environment
- 16 amp MCB and "wing it"
- Design it as a sub main since the local CU essentially "resets" the distance for final circuit VD (which in reality is what I would probably do)
- Increase the cable size to 6mm and protect at say 25 amps. But is that a "good design" ?

If the circuit does need to be long (distant bedroom) then I would consider a 4mm radial or a sub-main with local CU.

That's almost a matter of words, isn't it - since we'd then be talking mainly about the CSA of the 'sub-main' rather than the final circuit?

Well, not really semantics because the VD spec is based upon the final circuit not the distribution circuit. True that you use more copper (in the sub-main cable) and you could simply say why not use a 6mm radial anyway. But I would also consider that a local CU of some sort would be reasonable when you are roughly 40 meters away (25 amp MCB) from the main CU so I am applying other considerations too.

 

[JohnW2]

Well, not really semantics because the VD spec is based upon the final circuit not the distribution circuit.

Unfortunately, that doesn't seem to be true - and that's really the main reason for my questions. BS7671 seems to make it very clear that the VD limits relate to the total drop from the origin of the installation, including distribution circuits, not just the final circuit.

Kind Regards, John.

 

[sparkticus]

Well, not really semantics because the VD spec is based upon the final circuit not the distribution circuit.

Unfortunately, that doesn't seem to be true - and that's really the main reason for my questions. BS7671 seems to make it very clear that the VD limits relate to the total drop from the origin of the installation, including distribution circuits, not just the final circuit.

That's correct otherwise you could have multiple distribution circuits in series with perhaps only 10 volts on the end of it (based upon load current) I badly worded my sentence. My thinking would be that it would be prudent to run a distribution circuit (sub-main) with significantly increased CSA to allow for a local CU of some sort. The reason I bring in that additional factor (additional to the VD issue) is that in practice if one socket group are on the ragged edge of the VD limit then most if not all circuits in the vicinity will probably be close also. So I would tend to feel that there would be a reasoned case for a distribution circuit to supply that region (lighting - power etc) Reset was a bad word to use so I guess I really meant re-scaled.

 

[sparkticus]

Obviously you cannot have a ring final to a kitchen 100 metres from the CU.

Probably not, but if the kitchen were 100 metres from the origin of the supply, you'd have to have either a 100m radial or a 100m sub-main - and, as I've said, the difference is essentially semantic, since either would have to satisfy the VD requirements.

Well, semantic in a purely academic sense which I recognise is still semantic but in practice there are other factors that would make a distribution circuit (proper) a reasoned solution. As I mentioned if one circuit is on the ragged edge of VD then other circuits in the vicinity will also be close. A case for -rescaling with a sub-main and CU are in order I think.

 

[JohnW2]

Unfortunately, that doesn't seem to be true - and that's really the main reason for my questions. BS7671 seems to make it very clear that the VD limits relate to the total drop from the origin of the installation, including distribution circuits, not just the final circuit.

That's correct otherwise you could have multiple distribution circuits in series with perhaps only 10 volts on the end of it (based upon load current)

Quite so.

I badly worded my sentence. My thinking would be that it would be prudent to run a distribution circuit (sub-main) with significantly increased CSA to allow for a local CU of some sort. The reason I bring in that additional factor (additional to the VD issue) is that in practice if one socket group are on the ragged edge of the VD limit then most if not all circuits in the vicinity will probably be close also. So I would tend to feel that there would be a reasoned case for a distribution circuit to supply that region (lighting - power etc) Reset was a bad word to use so I guess I really meant re-scaled.

Indeed, and that's precisely the arrangement I have in my (fairly large) house. There are a total of seven CUs, five of which are on sub-mains, for the very reason you describe. In fact, contrary to what I think you said before, at least in terms of current-carrying capacity, this approach facilitates some saving of copper, by application of diversity to the submain - in other words, with diversity, the design current for the sub-main can be (perhaps considerably) less than the sum of the design currents (themselves with diversity applied) of all the final circuits it feeds.

The point is, of course, that if one does apply diversity in terms of CCC when sizing the submain, one could fall foul of VD if one didn't also apply diversity to that calculation (and maybe with lighting circuits even if one did) - and, as I've been saying, it's not really clear (at least not to me) whether the regs intend diversity to be allowable in a VD calculation - since they appear to be silent on this question.

Of course, and this was one of the things in my mind when I wrote about 'semantics' before, if there is essentially only one final circuit, as in an outbuilding, then it doesn't really make any difference (VD-wise) whether one is calling the cable feeding it a radial circuit or a sub-main.

Kind Regards, John.

 

[sparkticus]

how would you feel about a (very!) long 4mm radial with a 20A MCB which only satisfied VD requirements with, say, the 10A total load which the designed regarded (currently) as the 'probable load'?

for all that I failed to read your question correctly and answered a completely different question. apologies.

Now to answer your question:

In such a case I would "let it go" would tick all the cert boxes and would still sleep well at night.

 

[JohnW2]

Well, semantic in a purely academic sense which I recognise is still semantic but in practice there are other factors that would make a distribution circuit (proper) a reasoned solution. As I mentioned if one circuit is on the ragged edge of VD then other circuits in the vicinity will also be close. A case for -rescaling with a sub-main and CU are in order I think.

I've addessed this in my recent response you previous post. In essence, I definitely agree with you IF there are 'other circuits in the vicinity'.

Kind Regards, John.

 

[JohnW2]

for all that I failed to read your question correctly and answered a completely different question. apologies. Now to answer your question:
In such a case I would "let it go" would tick all the cert boxes and would still sleep well at night.

Thanks.

Kind Regards, John.

 

[sparkticus]

The point is, of course, that if one does apply diversity in terms of CCC when sizing the submain, one could fall foul of VD if one didn't also apply diversity to that calculation (and maybe with lighting circuits even if one did) - and, as I've been saying, it's not really clear (at least not to me) whether the regs intend diversity to be allowable in a VD calculation - since they appear to be silent on this question.

Yes, lighting circuits on one hand are invariably fixed max-load so max loading pretty much predictable but on the other hand a 3% VD limit and most often smaller CSA conductors. In terms of diversity I look at specific individual circumstances and assume nothing before doing that. In fact my initial thoughts/concerns with lighting tend towards max Zs (especially with TN-S) despite the fact that you can hit the VD limit earlier.

Having said that, the invention of low energy lighting is taking the stress out of that situation a little. Traditionally it was standard practice to assume 100 watts per lighting point (actually it still is as far as I know) but in reality in a domestic environment and in some commercial/industrial environments you are often looking at only 15 watts or so per point.
Though circuit inductance is another matter. Having just written that my mind is immediately drawn to a regular domestic customer we have who runs 5.5KW of halogen lighting (yes all ceiling torches) left on most of the time

With some fixed motor/high inrush equipment I have found Max Zs becomes a significant consideration/worry with TYPE-C CBs on a TN-S supply.

 

[sparkticus]

- in other words, with diversity, the design current for the sub-main can be (perhaps considerably) less than the sum of the design currents (themselves with diversity applied) of all the final circuits it feeds.

Yes, I see where you are going with that and I had such a situation some time ago. I designed a very distributed installation (lots of outbuildings split across 3 phases) and found that I had to look in intimate detail at the diversity of the various sub-main loadings VS the diversity loadings of individual final circuits. In fact I spent hours generating a spread sheet to keep track of everything. Amongst other things the main supply intake (3 phase 4 wire TN-C-S) was not in the main dwelling but in the piggery of all places!!!

 

[JohnW2]

- in other words, with diversity, the design current for the sub-main can be (perhaps considerably) less than the sum of the design currents (themselves with diversity applied) of all the final circuits it feeds.

Yes, I see where you are going with that and I had such a situation some time ago. ....

Of course, the lower VD allowed for lighting can sometimes destroy my argument for a sub-main being the better way to go. Consider the following - which, admittedly has been deliberately contrived to have marginal figures.

.... consisder a 'distant bathroom' with a 9.75KW/230V shower (42.4A) and 2A of lighting, such that cable runs from 'source' would be 37 metres. Considering just the shower, clipped direct we could manage with 6mm² cable (CCC 46A), and the VD just scrapes in as acceptable at 11.45V. A dedicated circuit for just the lighting would be fine with 1mm² cable (VD 3.25V).

If we used a sub-main for both shower and lighting (total 44.4A), the VD, at 7.2V, of 10mm² cable would not be acceptable for the lighting - so we'd have to move up a further notch to a 16mm² sub-main.

Hence, with seperate circuits for shower and lighting, we could use 6mm² and 1mm² cables. If we wanted to run a sub-main to a local CU, we would have to use 16mm² - more than twice the amount of copper (and probably cost) ... which all goes to remind us that one size does not fit all, and that each individual case has to be designed on its own merits.

Kind Regards, John

 

[bernardgreen]

which all goes to remind us that one size does not fit all, and that each individual case has to be designed on its own merits

Going a bit deeper into the maths

The heater rated 9.75 Kw at 230 has a resistance of 5.42 ohms

R = V*V/W = 230 * 230 / 9750 = 5.42

The cable ( based on quoted volt drop ) has a resistance of 0.27 ohms

R = V / I = 11.45 / 42 = 11.45 / 42 = 0.272

Total circuit resistance is therefore 5.69 ohms

With 230 volts applied the current will be 40.42 amps

I = V / R = 230 / 5.69 = 40.42

Power dissipated as heat in shower will be 8855 watts

W = R*I*I = 5.42 * 40.42 * 40.42 = 8855.068088

Power dissipated as heat in the cable will be 441.12 watts

W = R*I*I = 0.27 * 40.42 * 40.42 = 441.119

Total power = 8855 + 441 = 9296 watts

% useful in heater 95.25 %

% wasted in cable 4.74 %

Shower is running at 91 % of rated heat output

8.85 Kw and not at 9.75 Kw as expected.

Power loss in cable is significant when the designed cable size is just inside the voltage drop limits. For a shower the duration of the loss is fixed by the person in the shower, but for applications such as heating a kettle or immersion heated tank of water the duration depends on how much heat is reaching the water. Some one else can do the financial maths of what would be the pay back time for investing in cable larger than the minimum size the voltage drop regulations require to be used.

And the thought of 12 watts per metre being used to heat a copper wire inside two layers of PVC insulation (441 / 37 = 11.91 ) tends to worry me. The idea of that amount of heat depending on being clipped to a wall to help dissipate the heat leads me to use larger cable,

 

[ericmark]

I asked the same question at an IET local talk that was given. I was questioning the 106 meters he claimed we could now use in a final ring circuit.

What I was told is for a final ring circuit of 32A we consider 20A to be draw at a central point and the remaining 12A to be even throughout the circuit. So we use 26A as the design current when working out volt drop.

This however means for example a 20A radial would still if final socket was a double be calculated for a 20A load at it's end.

It is due to this accepted method of calculating the load that the ring is still accepted as a good method of wiring a house.

As the design engineer you can of course take into account all sorts of local variations and use a different method. However you then need to show in your paperwork how you have arrived at your figures and do a complete risk assessment.

As the sad old get I am I made up an Excel work sheet to experiment with different cable sizes etc using the formula given. It became apparent that using a B32 MCB the cable lengths allowed for ELI were nearly the same as allowed with a 5% volt drop. So if one works as if there is no RCD in circuit and limits the lengths of cable to ensure the ELI will be within limits then in the main the volt drop will also be within limits.

 

[JohnW2]

Going a bit deeper into the maths ... Power dissipated as heat in shower will be 8855 watts ... Power dissipated as heat in the cable will be 441.12 watts ...Shower is running at 91 % of rated heat output
Power loss in cable is significant when the designed cable size is just inside the voltage drop limits. For a shower the duration of the loss is fixed by the person in the shower, but for applications such as heating a kettle or immersion heated tank of water the duration depends on how much heat is reaching the water. Some one else can do the financial maths of what would be the pay back time for investing in cable larger than the minimum size the voltage drop regulations require to be used.

Yes, that is certainly a consideration, but I think the pay-back time could often be extremely long. With just the shower, as you imply, there would probably be no financial cost due to the lost power, because of human behaviour - and, even if it weren't for that, the duration of showers is pretty brief. A full financial analysis would, of course, also have to take into account the fact that the power dissipated in the cable was not necessarily 'wasted' - since, at least in winter, it would contribute to heating of the property.

As for kettles etc., (i.e. a socket circuit), I think the issue is a bit more general that you imply. With, say, a socket circuit with a design current of 32A (on which VD calcs will have been done), the kettle alone itself would obvioulsy get one nowhere near the limit imposed by VD. Hence, the issue is not really anything to do with the situation "when the designed cable size is just inside the voltage drop limits" - but, rather, simply a statement of the undeniable fact that 'the fatter the cable the better' in terms of reducing power loss.

And the thought of 12 watts per metre being used to heat a copper wire inside two layers of PVC insulation (441 / 37 = 11.91 ) tends to worry me.

For ease of calculation, one can obvioulsy get directly to the '(milli)watts per metre' for any cable figure simply by multiplying the tabulated VD (in mV/A/m) by I².

I'm not too sure why you are worried, unless you're just talking about 'wasted power' again. We are obviously talking about cables which are being used within their specified current-carrying-capacity ratings, and those ratings are based on a knowledge of the mW/metre that will be dissipated in them at the specified currents

Kind Regards, John