I want to run power to a shed. I think my best option is to export the house earth but I'd like a second opinion.
Summary of proposed installation
incomming supply to house - Ze 0.15 ohms, TN-S
I plan to take power from spare slot on 30mA RCD side of consumer unit (20 A mcb should be more that enough).
Total length of cable run is around 20m. I plan to take a length of
4mm2 T&E to outside wall and then connect to 4mm2 SWA (roughly half SWA half T&E).
Shed is wooden with concrete floor.
I don't intend to fit a cu in the shed, just fit a fcu (5A) for lights and wire a few sockets directly to the incoming SWA as a radial circuit (+ maybe a 20A DP switch as an isolator).
I've not yet checked the disconection times, voltage drop etc although my feeling is that for this distance with a 20A mcb it should be OK (don't have resistance/m data for swa to hand). Are there any issues with exporting the earth through SWA armor (as opposed to 3 core) assuming the earth loop impedance is sufficiently low.
1) Anyone think I would be wise to install TT earthing (+ cu etc) in the shed?
2)When I take the 4mm T&E through the exterior wall presumably it needs to be enclosed in conduit (plastic/steel?) and taken to a weather proof junction box IP???? where I can join it to the swa.
3)Are there any issues with exporting the earth through SWA armor (as opposed to 3 core) assuming the earth loop impedance is sufficiently low?
The usual problem of discriminating between circuits in the shed.
Voltage drop is 11 mV/A/m, or about 4.4 V in this case.
R per meter will be a factor of the conductor CSA rather than the cable type as such.
If you were using 3 core SWA with the third core as CPC, then you get
9.22 mOhms/meter on 4mm sq or about 0.2 Ohms. So a total loop impedance of 0.35 ohm. So a prospective fault current of over 650A. That ought to open a 20A type B MCB pretty sharpish! (I expect Andy will be along shortly to do the calc for using the screen in place of 3 core ;-)
Not really... need to make sure the gland is well done up, and stick a boot over it to stop corrosion.
It would give you the option of taking the submain feed from the non RCD size of the house CU, and then having split load in the shed. It also saves you a trek back to the house when you trip the RCD.
If you go for the 3 core SWA then you could terminate the screen in a metal (waterproof) box but continue the inner part through the wall to an internal JB. Or if you don't have to have the SWA flat against the wall you can take the whole thing through.
I refer the honerable gentleman to the previous answer ;-)
Do you think this will be a significant problem as i'm essentially planning a single radial circuit with a fused 5A spur to the lights. I'm not actually sure I care if a fault on the light circuit blows the 5A fuse or trips the 20A mcb first. If I was planning anything more elaborate I'd fit another cu.
One thing I did forget to ask was that I was intending to go from
4mm2 cable up to the shed and then to 2.5mm2 to sockets inside the shed. Is it OK to mix different size cables on a circuit, provided disconection times etc are met.
impedance
ought to
I found some values on the net which worked out a bit higher than yours (9mOhm/m looks like 6mm2 on the table I found), but still gave a fault current of 500A.
sufficiently
stick a
Surely I could do this anyway and take the house earth to the split load cu in the shed?
Because of where the T&E cable will have to run - under floorboards upstairs. I would rather join the cables outside (although I take your point - probably a more reliable connection). The weatherproof junction boxes I've seen are all plastic?
Voltage drop: live conductors are 4mm^2 all the way, drop is 11 mV/A/m [Table 4D2B] hence 4.4 V max. drop for Ib = 20A and a 20 m run. This is under 2% and leaves you another 2% to play with for wiring within the shed. All OK here.
Earth fault loop impedance (Zs):
(a) contribution from 10 m of 4mm^2 T&E with 1.5mm^2 CPC - total resistance R1+R2 = 16.7 mohm/m [OSG Table 9A], hence contribution to Zs is 0.17 ohm for 10m run;
(b) ditto for 10m of SWA (XLPE insulated to BS 5467): phase conductor
4.6 mohm/m, armour (~19mm^2) 7.9 mohm/m, total 12.5 mohm/m, say 0.13 ohm total contribution to Zs for a 10m run.
(a) plus (b) gives about 0.3 ohm total contribution to Zs. Taking Ze as
0.8 ohm (the recommended design basis, rather than relying on your measured value of 0.15 ohm) gives Zs ~ 1.1 ohm. This is a 'cold' value by the way based on resistances at 20 deg. C. Consulting Table 2D of the OSG for a B20A MCB tells us that the maximum measured (cold) Zs is
1.92 ohm, so you're miles in, and are not relying on your RCD for earth fault clearance. (You could consider feeding from the non-RCD side of the consumer unit and putting another RCD in the shed, for the sockets.) With a Type C MCB we need Zs < 0.96 ohm, so there's a potential problem problem there in principle - although not in practice because of your nice low Ze, which is most unlikely in practice to increase.
CPC adiabatic compliance - with MCB protection tripping in < 0.1 s it's hardly necessary to check this - but here goes anyway:
Worst case is with Zs = 1.1 ohm (see above), earth fault current is ~
230 V / 1.1 ohm - say 200 A (rounding down to allow for a bit of heating). For the T&E cable with a copper CPC the min. CPC size is sqrt(I^2*t)/k with k = 115, i.e. 200*sqrt(0.1)/115 which is about
0.55mm^2. The cable has a 1.5mm^2 CPC, so no problems here. Now for the SWA armour: same calculation but with a k value of 46, giving a min. CPC of ~1.4mm^2. As we have 19mm^2 of armour there's no problem here either - this confirms that it's OK to use the armour as the CPC and there's no need to use a 3-core SWA.
Dry inside, no incoming metal service pipes coming out of the ground, no intention regularly to use Class 1 appliances outdoors? - if all those apply you're OK with the exported earth. If not, consider the TT option.
That's OK, provided you accept the inconvenience of losing both light and power if anything trips the MCB or RCD in the house.
See calculations above.
Google for previous lengthy debates.
You don't need to protect the T&E in the wall, but you do need to be sure that water can't get in to the IP66 (say) box through your cable hole at the back.
That's not usually an issue with the smaller SWA sizes (greater ratio of armour area to copper area). For larger sizes a separate CPC or third core may become necessary.
It was more a case of if a fault on the socket circuit taking out the lights was going to give you a problem. i.e. power tools can carry on spinning even with not power and the lights off.
It would be more usual to run a radial in 4mm T&E in this circumstance. If you go to 2.5mm then you have to watch the cable load.
According to the OSG my figures are right for a 4mm sq phase with 4mm sq CPC....
However thinking about it, my figurs are half wrong since you are not going to have the 4mm sq CPC for the whole length - only the SWA bit. The T&E will typically have a 2.5mm sq CPC which will raise R1+R2/m to
12.02 ohms. As you say though, it does not make much difference since you still get a decent fault current.
Yup, looks like it.
You can use plastic - you will just need a earthing tag connection to the gland to provide connection to it rather than relying on the box.
There's no problem with 2.5mm^2 here - the whole circuit is only rated at 20A. As I showed, with a B20A MCB and the 4mm^2 submain, there's 2% voltage drop and about 0.8 of an ohm of Zs in hand - so you you could run up to about 12m of 2.5 T&E before you hit any limit. At 20A (for a double socket) you hit the voltage drop limit first. A branch in 2.5 feeding a single socket (13A) could go for 20m (still voltage-drop limited). How big's the shed...?
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.