Help with specifying a long 4-core SWA run please

[Mark42]

I need to augment the underground supply to remote workshops at my own house.

I’m struggling to understand the practical implications of voltage drop, rather than using blind compliance with recommendations/tables/regulations which do not take account of individual circumstances.

Here’s the spec:

• A three phase x 100A TNC-S supply at the main house
• A 220m run from the main house to the workshops. Will be local TT (wet clay).
• Long-term loads at the workshops are mainly heating panels. If everything were on at once, it would be possible to use about 8kW of heating on any one phase. But never on all three phases at the same time.
• The only three phase loads are trivial roller shutters, a 5kW compressor and a 2kW dust extractor. Both are lightly used, and never run continuously.
• There are many other occasional SP loads, like a power saw, computers, blown-air diesel heater, and lights which are now all LED and relatively trivial.

The question is what size 4-core SWA do I need to use?

Cu prices have gone up massively and I need to consider cost. Or maybe aluminium sectoral cable? 4 x 35mm Cu is OK, 50mm is scary, and 70mm economically impossible. Runs of this size and length are beyond my experience.

The actual voltage on this rural site is 250V+. NOT the 230V we are supposed to use in the tables. Does this affect the practical implications of transient voltage drops?

ie Why the obsession with adiabaticity? If there’s a temporary underground cable ‘overload’, when say the compressor or a saw is used for 20 minutes, and a few degrees of heat is generated in the underground cable, then so what? It’s a waste of energy of course, but it will never cost me the thousands of pounds in additional copper indicated by the tables.

The buildings are currently supplied by a legacy 3 x 10mm, which shows no sign of heating at the terminations. But I am careful and watch the loads. I will in future rent these buildings for short-term research work, so this needs to be done properly, without me worrying about the upstream 32A breaker tripping. I’d much prefer to be able to fit a 40A B-curve, or even a 40C.

 

[buzzlightyear]

Multicore 90°C Armoured Thermosetting Insulated Cables​

Tables Apply to: H6942XL, H6943XL, H6944XL, H6945XL, Tuff Sheath,

Voltage:400V

Load:8kW / 14.43A

Length:220m



Cable Size:10mm²

volt drop 17.6
percentage 4.4
load 20amps
and no smoke paper insight .lol.

 

[telectrix]

this is not one to think about after 6 pints. cu tomorrow.b?

 

[Mark42]

Multicore 90°C Armoured Thermosetting Insulated Cables​

Tables Apply to: H6942XL, H6943XL, H6944XL, H6945XL, Tuff Sheath,

Voltage:400V
Load:8kW / 14.43A
Length:220m
Cable Size:10mm²
volt drop 17.6
percentage 4.4
load 20amps

and no smoke paper insight .lol

Thanks but it's around 8kW of domestic heaters on each phase, not a single 8kW three phase heater! (2.7kW = 10.8A @250V - that would be easy ?)

Plus I need to allow for other loads, so the calcs are best done for say 10kW/phase.

One of the problems is that in winter, all the heating would go on together each morning for a few hours until the rooms warm up, when they would switch in and out intermittently, providing diversity.

I'd plan to blast the place with the very noisy 45kW diesel heater before the client/s turn up, then use the electrics for maintenance.

Everything's going on new 'Heatmiser Neo' thermostats, so complex (and remote) timed control will be easy via the app. That give me a lot of control over the total instantaneous load, but I have to allow for daft clients who bung everything on manually, all at once. And then plug their their machines into the 'wrong' sockets, overloading one phase.

 

[buzzlightyear]

You need to recalculate the loading again for extra loading.

 

[pc1966]

Sticking in 220m and 30kW (i.e. 10kW/phase) an on-line calculator comes up with 25mm, so cost is around £2.5k or so plus VAT for 4C copper SWA.

Going to aluminium cable can save money BUT comes with a lot of extra trouble. It is very much more brittle and you need special care with termination materials/type/compounds to avoid corrosion and hot spots forming. Typically you need 1.5 times the CSA or 1-2 steps up in size (so maybe 35mm Al). However, aluminium is rarely seen under 70mm or so in SWA.

If you want to save a bit on the cable can you arrange for better balance of your heating loads across the 3 phases, etc?

Unfortunately it is not likely to be at the point where the cost of transformers makes up for the possible saving in cable (that also has all sorts of added complications).

 

[pc1966]

ie Why the obsession with adiabaticity? If there’s a temporary underground cable ‘overload’, when say the compressor or a saw is used for 20 minutes, and a few degrees of heat is generated in the underground

The adiabatic limit is generally used for overloads/faults of around the 5s or less disconnection times typically planned for at the design stage. Its assumption is the temperature rise is due to the short-time fault energy heating the conductor quickly so little heat is lost to the surroundings.

For modest overloads, say in the minutes region, then your typical OCPD trip behaviour is matched to the cable already i.e. if the OCPD is <= cable CCC then all overloads and faults are protected against and it allows for heat escaping.

However, for any modestly long cable run you end up with the volt drop dictating the cable CSA (not heating effects), and usually if you meet the VD limits then your fault current is going to be high enough for safe quick disconnection. But almost never with TT, which is why the RCDs come out...

 

[Mark42]

The adiabatic limit is generally used for overloads/faults of around the 5s or less disconnection times typically planned for at the design stage. Its assumption is the temperature rise is due to the short-time fault energy heating the conductor quickly so little heat is lost to the surroundings.

For modest overloads, say in the minutes region, then your typical OCPD trip behaviour is matched to the cable already i.e. if the OCPD is <= cable CCC then all overloads and faults are protected against and it allows for heat escaping.

However, for any modestly long cable run you end up with the volt drop dictating the cable CSA (not heating effects), and usually if you meet the VD limits then your fault current is going to be high enough for safe quick disconnection. But almost never with TT, which is why the RCDs come out...

Thanks. all good stuff, which I'd long forgotten!

And you're right about Aluminium. I'll forget that idea.

So does that mean that I can choose the OCPD primarily on the cable CSA/CCC alone? So a 40C breaker is (obviously) fine on say 35mm, even if it's over 200m long? Any gross fault downstream would trip a local MCB (none of which exceed 20A) before the supply main cable began to warm up significantly. A fault on the buried supply cable itself is unlikely.

There are three separate TP DBs in the workshops. One has a 30mA RCD incomer, the others have multiple RCBO circuits.

There's currently a historic mixture of 'exported' PME, and multiple local rods earthing the metallic building structures. I haven't yet checked the local earth impedance or PFC.

 

[bigspark17]

Why are you TT ing the outbuilding? Have you taken an earth rod reading? Being that far from the transormer might make it unviable to get a low enough RA..

 

[pc1966]

So does that mean that I can choose the OCPD primarily on the cable CSA/CCC alone? So a 40C breaker is (obviously) fine on say 35mm, even if it's over 200m long? Any gross fault downstream would trip a local MCB (none of which exceed 20A) before the supply main cable began to warm up significantly. A fault on the buried supply cable itself is unlikely.

Not quite.

You have two forms of protection and they are quite separate issues:

• Fault protection must be provided (so any short, etc, disconnects in under 5s/0.4s for a TN sub-main/final circuit)
• Overload protection is usually required, but not always for fixed loads (so the OCPD goes before thermal damage to the cable)

You can have one but not the other. So you can have fault-protection OCPD that clears a short but allows the cable to roast, or overload protection that stops a fire but fails to disconnect a fault in any sane time-scale (typically if you fail the VD requirement for the OCPD rating even if you meet it for the expected load on long cables).

In this case you have around 220m of cable, and if my earlier guess at the VD parameters are right, it is 4C 25mm and you plan a 40A OCPD so from the OSG tables.

• 40A BS88-2 fuse has 1.0 Ohm (5s)
• 40B has Zs max of 0.88 Ohm (0.4s & 5s identical)
• 40C has Zs max of 0.44 Ohm (0.4s & 5s identical)
• 40D has Zs max of 0.44 Ohm (5s)

From the Prysmian SWA data sheet 25mm 4C has R1 = 0.727 ohm/km and R2 = 2.3 ohm/km, so R1+R2 = 0.67 ohms at 220m. Running 16mm CPC in parallel with the armour is another option, so R1+R2 is around 0.33 ohm then if Zs is 0.11 ohm you can use any of the above, otherwise you are forced in to 40A B-curve, or a fused-switch.

Looking at the Hager catalogue, a 20A B is selective with a 40A B upstream MCB to around 0.4kA, and to around 0.5kA with a 40A BS88 fuse, so not much difference. But oddly a 20A RCBO is only selective with 40A B MCB to around 0.18kA.

If none of them are usable due to Ze being too high then you are looking at a supply-side delay RCD for fault-to-earth disconnection and then you might even go with a D-curve MCB for overload protection if final circuits all are RCBO so selective with the supply delay-RCD.

 

[pc1966]

There's currently a historic mixture of 'exported' PME, and multiple local rods earthing the metallic building structures. I haven't yet checked the local earth impedance or PFC.

There are others who usually point out that you do not 'export PME' unless you are TN-C between DB which in the UK wiring regs is a strict no-no, only the DNO seem to be permitted to use TN-C.

If it is a farm on PME, or once was a TT supply, then multiple earth rods are likely to keep any stray potential down as farm animals REALLY don't like that. Is there anything likely to be used that actually would be best on a TT supply like livestock handling or hot-tub, etc?

If not, you would be best simply to keep it all one the one TN earth system along with the rods, etc, just in case of an open PEN fault or lightning surges, etc. But check all the estraneous earth bonding is at least 10mm copper equivalent if it is a TN-C-S (PME) supply.

You might still need to look at a delay RCD for sub-main feed if you can't meet disconnection on the end of cable Zs with the chosen OCPD, but even a fairly poor rural supply Ze and long sub-main cable is going to be an order of magnitude or two lower Zs than a local rod or two.

 

[Mark42]

There are others who usually point out that you do not 'export PME' unless you are TN-C between DB which in the UK wiring regs is a strict no-no, only the DNO seem to be permitted to use TN-C.
...
If not, you would be best simply to keep it all one the one TN earth system along with the rods, etc, just in case of an open PEN fault or lightning surges, etc. But check all the extraneous earth bonding is at least 10mm copper equivalent if it is a TN-C-S (PME) supply.

Your replies above are brilliant help: thanks for spending the time to detail all this.

Shame you’re the other end of the country – what I really need is an electrical engineer here for a day's consultancy to go through the current install with a second pair of eyes, and modern test kit, then define the best way forward before I backfill 200m of trench. (Anyone in Norfolk?)

For the last 30 years it’s been only me here, with full control of what goes on - electrically and otherwise. Now I’m starting to rent the workshops to others I must ensure everything’s right.

I’m effectively a DIY-er, I did (easily) pass the 17th ed’n exam years ago, but as I have ten other careers in parallel and have since been doing other things, I have forgotten a lot … It’s nothing to do with getting older, of course ?

Your comments on keeping the ‘exported’ PME hooked up chime with my own ideas.

I’ll make an admission: that’s exactly what I have always done on all the buildings. Some even have 5-core supplies with the ‘PME’ earth still connected, not ending in a plastic box as is the usual practice.

Plus I enjoy driving rods here and there to ground the metallic building structures. As you suggest, all the bonding is in at least 10mm, usually 16 or 25.

Although there are no animals, and certainly no big quadrupeds with legs far apart (this is an ex-farm site) and the risk of an external loss of neutral is low: the brand new transformer is right on the edge of my site, and all LV is now underground. It serves only two other properties.

I wouldn’t normally admit this practice on here as it may produce a storm of abuse, and ‘you can’t do that because you’re breaking the regulations’ (but without technical details of why not) type of comment.

I see it like this: everyone bonds metallic buildings, and as far as I am aware, metallic buildings are not built on big rubber pads. Rebar is often bonded to the structure: the whole thing is a giant earth rod. Therefore any bonded metallic building provides another path to ground for the DNO’s TNC-S arrangement. As do (did?) bonded metallic service pipes. So why not go the whole hog and make the entire site one huge equipotential zone?

It also means that earth impedances everywhere are kept low.

Maybe I’m misunderstanding something, but I can see no theoretical or practical reason why it’s wrong. It provides so many paths to earth, all over the place, that in the event of a neutral loss, it would pull any touch voltages down to safe levels, and if I were providing the ‘whole street’s’ return path I reckon there’s enough copper to handle it. But there is no street anyway.

Over ?