A2 and A3 radials for sockets

[Pretty Mouth]

Browsing the OSG, I notice in section 7 that the data for ring finals assumes the load is distributed throughout the circuit, whereas the data for 20A and 32A radials assumes a terminal load.

So my question (particularly to those who install radial socket circuits regularly, in preference to rings):

When calculating voltage drop for A2 and A3 radials powering sockets, how are you treating the load: as if all the load is at the terminal point of the circuit, or as if it is distributed throughout the circuit? And if the latter, how do you assume it is distributed?

 

[littlespark]

When you say a terminal load, do you mean the load is at the furthest point from source?

So all the cables would be carrying the same load. (ignoring any branches off at any point)
If the load was say, half way along, then the cable downstream of where the load is would have no current flow. only from the load, back to source.

With a ring, the current has two routes back to source, so wherever the load is on a ring, there is some current at all points. The load may as well be evenly distributed.

Confused as what the A2 and A3 means? Is that something only in that OSG... I cant find anything in the regs being named A2 and A3

 

[pc1966]

I would go with the OSG values that assume the load is at the end for a couple of reasons:

• I personally would only look at a radial for a handful of sockets in one place, not a whole floor, so the loads are all biased to the end anyway.
• Often a radial for sockets is 20A / 2.5mm and you could have all of that from a double socket at the very end by design anyway.
• The OSG length limits also take in to account max Zs and adiabatic limits, they should be checked for the end socket case and if you assume a distributed VD to approx double the length you need to re-calculate them

Worst-case for VD/Zs on the RFC is also for the balanced case (i.e. half way round the ring) so nothing else to worry about, bar too much load on one leg by site layout / room use.

 

[Pretty Mouth]

When you say a terminal load, do you mean the load is at the furthest point from source?

Thanks for the reply.

Yes, Table 7.1(i) in the OSG is split into 3 parts: rings, lighting and radials. The radials have been calculated as a terminal load, whereas the rings and lighting have been calculated with the load distributed around the circuit. It occurred to me that perhaps a radial feeding sockets could be calculated as having a distributed load too.

Confused as what the A2 and A3 means? Is that something only in that OSG... I cant find anything in the regs being named A2 and A3

This must just be in the OSG because I can't find it in the regs either. Table H2.1 on p188, plus in other places in app H

 

[Pretty Mouth]

I would go with the OSG values that assume the load is at the end for a couple of reasons:

• I personally would only look at a radial for a handful of sockets in one place, not a whole floor, so the loads are all biased to the end anyway.
• Often a radial for sockets is 20A / 2.5mm and you could have all of that from a double socket at the very end by design anyway.
• The OSG length limits also take in to account max Zs and adiabatic limits, they should be checked for the end socket case and if you assume a distributed VD to approx double the length you need to re-calculate them

Worst-case for VD/Zs on the RFC is also for the balanced case (i.e. half way round the ring) so nothing else to worry about, bar too much load on one leg by site layout / room use.

Thanks @pc1966 .

I'm trying to work out how it's feasible to install these instead of rings, as the circuit lengths for either a 20A / 2.5mm or 32A / 4.0mm from the OSG are not very generous (about 42m). I could see these limits being exceeded easily.

 

[pc1966]

circuit lengths for either a 20A / 2.5mm or 32A / 4.0mm from the OSG are not very generous (about 42m). I could see these limits being exceeded easily.

If it is a couple of rooms then that is exactly what the RFC was created for!

Otherwise you can look at using 4mm / 6mm feed to the first socket then 2.5mm/4mm for the rest if that is a significant portion of the overall length, but you have more calculations to check and the "what do I put down in the EIC for size?" sort of question. Or you up a size so 20A and 4mm for whole length (69m), etc, but at somewhat higher cost.

 

[telectrix]

i'd rarely consider a RFC for sockets in bedrooms. max. load is usually < 10A , couple of lamps, chargers etc. and of course separate radials for the kids' bedrooms, so you can turn off each MCB at will.

 

[pc1966]

i'd rarely consider a RFC for sockets in bedrooms. max. load is usually < 10A , couple of lamps, chargers etc. and of course separate radials for the kids' bedrooms, so you can turn off each MCB at will.

Well there is your answer, put them on 10A C-curve MCBs and teach them about load management in to the bargain!

 

[Pretty Mouth]

If it is a couple of rooms then that is exactly what the RFC was created for!

Otherwise you can look at using 4mm / 6mm feed to the first socket then 2.5mm/4mm for the rest if that is a significant portion of the overall length, but you have more calculations to check and the "what do I put down in the EIC for size?" sort of question. Or you up a size so 20A and 4mm for whole length (69m), etc, but at somewhat higher cost.

It's a lot of mucking around isn't it? It seems then that for standard socket circuits the ring is still king.

As @telectrix points out, power consumption in living areas isn't that great: 20A would be adequate, and this is why I'm looking into radials as an option. But with the VD limits, it's likely 2 radials would have to be installed where 1 ring would do. May as well just install a ring.

 

[pc1966]

As soon as you need two radials to cover a floor then you have to ask why not a RFC? More or less the same cable used and wiring effort, but in one case 2*MCB/RCBO, the other case a bit more 2.5mm to link end-end to form a ring.

Yes, it makes sense to have the kitchen sockets separate as that is likely the majority of household load and might be an unacceptable imbalance on the RFC depending on cable route/lengths. But in a small flat 1 RFC is normally fine for everything 13A (my own flat has that, and on 20A for historical reasons, though I don't have a dishwasher in the mix).

I know some folk raise the issue of future muppetry in terms of broken rings and/or spurs from the Devil's wiring guide, but to some extend that is not a design issue, that is an idiot meddling issue. Even then, a 20A RCBO (even C-curve if desired) would deal with that future scenario and allow something over 100m in 2.5mm for a RFC on VD/Zs

 

[Pretty Mouth]

As soon as you need two radials to cover a floor then you have to ask why not a RFC? More or less the same cable used and wiring effort, but in one case 2*MCB/RCBO, the other case a bit more 2.5mm to link end-end to form a ring.

Yes, it makes sense to have the kitchen sockets separate as that is likely the majority of household load and might be an unacceptable imbalance on the RFC depending on cable route/lengths. But in a small flat 1 RFC is normally fine for everything 13A (my own flat has that, and on 20A for historical reasons, though I don't have a dishwasher in the mix).

I know some folk raise the issue of future muppetry in terms of broken rings and/or spurs from the Devil's wiring guide, but to some extend that is not a design issue, that is an idiot meddling issue. Even then, a 20A RCBO (even C-curve if desired) would deal with that future scenario and allow something over 100m in 2.5mm for a RFC on VD/Zs

The Devils wiring guide ??

Out of interest, I had a play around with the VD calculations for the above radials. The data in the onsite guide assumes a load of 26A for a 32A circuit, and 16A for a 20A circuit. Putting half of the load at the furthest point, and half at the mid point, it increases the max length of both circuits to ~54m. A little better I guess, but as you pointed out in real life it may be that the load is concentrated close to the furthest point. I'm still not seeing a reason to move away from rings.

 

[pc1966]

The Devils wiring guide ??

Out of interest, I had a play around with the VD calculations for the above radials. The data in the onsite guide assumes a load of 26A for a 32A circuit, and 16A for a 20A circuit. Putting half of the load at the furthest point, and half at the mid point, it increases the max length of both circuits to ~54m. A little better I guess, but as you pointed out in real life it may be that the load is concentrated close to the furthest point. I'm still not seeing a reason to move away from rings.

I wonder if that is a 0.8 factor for cable heating and VD increasing with R, or maybe a factor for long-term operation of MCB that has been applied?

It would be nice to see the details of just how they are calculated, as usually you just accept them as standard practice,

 

[Pretty Mouth]

I wonder if that is a 0.8 factor for cable heating and VD increasing with R, or maybe a factor for long-term operation of MCB that has been applied?

It would be nice to see the details of just how they are calculated, as usually you just accept them as standard practice,

I had a bit more of a play with volt drop calculations for these standard circuits.

For the radials with terminal loads, a straightforward volt drop calculation (using the max length and assumed max load from the OSG data) gives a drop a little over the 11.5v allowed. Applying the correction for operating temperature from p382 of the BBB brings it down to almost exactly 11.5v in all cases that I tried.

For ring finals, I calculated it as if it was a radial, half the maximum length (53m) and half the assumed max load (13A), placed at the furthest point. Again, once the max operating temp correction had been applied, the volt drop came to 11.44v, very close to the permitted limit.