Service cut out 80 meters away from the installation. Is it OK and Cable Size

[spikyspud]

Hi All,

I have a client who is developing a plot of land for a single house. Currently, their 100 amp service cut-out is in a cubicle fed by overhead cable, however, this is in the way of the build. Rather than having the cubicle moved before the build and then again once the build is complete with the overhead cable buried, the DNO suggested looking at having their service cut out moved about 80 meters away to the top of the hill, and then power is taken down to the house by "any electrician". This removes the need for DNO involvement once the house is complete and gets rid of the overhead cable that they don't want to look at out of their window.

Having done the manual cable size calculation from Appendix 6 I am coming out at a 50mm2 cable for 100 amps over 80 meters. This drops 6 volts over the run. [in the calculation I am using a soil resistivity of 1.2 as the ground is wet clay even in summer as there is a lot of groundwater, giving a rating factor of 1.4 for direct buried SWA]. This obviously is pretty close to the 3% limit for total voltage drop for any lighting circuits.

I am planning on protecting the SWA with a 100amp cartridge fused switch. The consumer unit would be down in the house which will have a pretty standard install.

Questions I Had (Any help/advice appreciated)

1. What are people's thoughts on using 50mm2 or upping to 70mm2 to give more wiggle room on the voltage drop on the end of any lighting circuits (that aren't installed yet)? 70mm2 seems like an enormous cable given that the DNO would deliver it in 50mm2.
2. Is it allowed to lower the amps in the voltage drop calculation because 1 house is very unlikely to ever pull 100 amps?
3. Any issues with running in 2 x 25mm2 instead of a 50mm2 or 2 x 35mm2 rather than a 70mm2 (as both are easier to handle/terminate)
4. Any other regs that would prevent the service cut from being so far from the installation?

Thanks
A.

 

[pc1966]

1. What are people's thoughts on using 50mm2 or upping to 70mm2 to give more wiggle room on the voltage drop on the end of any lighting circuits (that aren't installed yet)? 70mm2 seems like an enormous cable given that the DNO would deliver it in 50mm2.

The DNO is not bound by 5% drop, just that it stays within ESCQR limits. The cost for allowing more drop is the consumer's meter is at the feed end, so they pay for the I2R losses and over time that could add up!

1. Is it allowed to lower the amps in the voltage drop calculation because 1 house is very unlikely to ever pull 100 amps?

Yes, if you know the supply needed you could do the equivalent of the DNO offering 80A or 60A.

But run through what they want and might need in the future (like ASHP or fast EV charger, but that might start looking 3-phase)

1. Any issues with running in 2 x 25mm2 instead of a 50mm2 or 2 x 35mm2 rather than a 70mm2 (as both are easier to handle/terminate)

Parallel is fine. You could use 25mm 4C or 5C if you want end wires easier to wrangle and are OK with the bulk SWA weight and bend issues.

1. Any other regs that would prevent the service cut from being so far from the installation?

If TN-C-S you need the CPC to meet bonding requirements of 10mm copper equivalent.

Also you need to meet sub-main disconnection times on the cable. Is this a TN supply? If so the fuse should be fine but check the armour R2 in your cable Zs calculation and if needed run parallel CPC or use extra core in SWA.

If this is TT then you are looking at a delay RCD to meet disconnection times and selectivity with downstream circuits. In this case you need N-switching RCBO otherwise you have no selectivity on a N-E final circuit fault.

Of course you need a local means of isolation, but that can be the main switch of the CU. However you might still want to put in a local isolator just in case there is any desire for more than one CU and better selectivity than MCB-feed, etc.

 

[spikyspud]

The DNO is not bound by 5% drop, just that it stays within ESCQR limits. The cost for allowing more drop is the consumer's meter is at the feed end, so they pay for the I2R losses and over time that could add up!


Yes, if you know the supply needed you could do the equivalent of the DNO offering 80A or 60A.

But run through what they want and might need in the future (like ASHP or fast EV charger, but that might start looking 3-phase)

Parallel is fine. You could use 25mm 4C or 5C if you want end wires easier to wrangle and are OK with the bulk SWA weight and bend issues.

If TN-C-S you need the CPC to meet bonding requirements of 10mm copper equivalent.

Also you need to meet sub-main disconnection times on the cable. Is this a TN supply? If so the fuse should be fine but check the armour R2 in your cable Zs calculation and if needed run parallel CPC or use extra core in SWA.

If this is TT then you are looking at a delay RCD to meet disconnection times and selectivity with downstream circuits. In this case you need N-switching RCBO otherwise you have no selectivity on a N-E final circuit fault.

Of course you need a local means of isolation, but that can be the main switch of the CU. However you might still want to put in a local isolator just in case there is any desire for more than one CU and better selectivity than MCB-feed, etc.

Thanks pc1996, very useful.

The installation is a TN-C-S and I was planning on using a core in the SWA as the CPC (as well as the armour). And there will be an isolator going at the house end as well, saves walking 80 meters up the hill if you need to isolate the whole supply for any reason as well!

 

[pc1966]

I would just add that is 3P is even a remote possibility in the future then going 4C 35mm plus 16mm CPC in parallel is, from memory, fine for 100A 3P at distances to about 100m and you can have parallel for 70mm equivalent now, and usual for 3P in the future, just with changes to the end fused-switch & local isolator. No need for a new cable run and all the civil works!

 

[davesparks]

4C 35mm plus 16mm CPC in parallel is, from memory, fine for 100A 3P at distances to about 100m

I don't think it would be, the voltage drop still needs to comply for 230V (unless all loads are 400V)

 

[pc1966]

I don't think it would be, the voltage drop still needs to comply for 230V (unless all loads are 400V)

You don't need all loads to be 400V / 3P, just that enough of the loads are distributed over the phases so sufficiently balanced that it works out OK.

If you run the calculations for 35mm 90C cable (1.35mV/A/m drop single phase) then you get 5% drop met for 100A and 85m single phase (exceeding the 80m of the OP's post), and 3% drop met for 100A at 51m or 64A at 80m.

If you are fully balanced 3P then the VD values are half of that due to neutral current cancellation, alternatively the distances are double. Basically the same as paralleling up conductors for the initial SP supply.

In reality the drops will be slightly less, as the 90C temperature for duct/burial is for 115A which exceeds the expected max demand & DNO fuses, and for 70C cable it is 98A (closer to design value for 100A supply) leading to 8% less drop as cable cooler so resistance less.

Basically if it was 35mm 4C initially paralleled, and then a change to 3P in the future, then they should make sure that existing loads get a bit of redistribution so there is not more than say 50A of expected single-phase load on any one phase, and not more than 100A for combined 3P + SP loads per phase.

 

[davesparks]

If you run the calculations for 35mm 90C cable (1.35mV/A/m drop single phase) then you get 5% drop met for 100A and 85m single phase (exceeding the 80m of the OP's post), and 3% drop met for 100A at 51m or 64A at 80m.

That doesn't leave much room for voltage drop on the final circuits.

There's a rule of thumb I've often heard mentioned, and tend to follow, that you should aim to have a VD of no more than 1% in a distribution circuit.

 

[pc1966]

That doesn't leave much room for voltage drop on the final circuits.

There's a rule of thumb I've often heard mentioned, and tend to follow, that you should aim to have a VD of no more than 1% in a distribution circuit.

But trying that here would be bonkers economically!

The VD on typical lighting circuits now is negligible due to LED lamps, so aiming for just under 3% would still allow you to meet 3% overall for that. And realistically the 3% drop for lights is something long overdue for amending considering how much less LED lights vary with supply voltage compared to filament lights. Then you have just over 2% for most other circuits, might need care for some but achievable for less than going 150mm feed cable!.

Also if the DNO ran the feed that extra 80m they are not going to try to keep to 5% overall budget, just to meet ESQCR at the cutout.