SWA Cable Sizing Chart

What the numbers in this chart are

Worth stating plainly, because most sizing charts do not. The current ratings and mV/A/m figures here are representative values for PVC-insulated armoured cable at a 40 °C reference — the same set the sizing calculator uses, so the two can never disagree with each other. They are not lifted from a standard, and they are not any one manufacturer's.

The dimensions and the armour area are a different kind of number again: they are computed, layer by layer, from the conductor outwards — area to diameter, insulation wall, the exact geometry of three or four circles laid up together, bedding, armour wires counted round the circumference, sheath. The overall diameters that model produces sit within a few per cent of typical published figures across the whole range, and the drawing below is a view of the same numbers rather than an illustration beside them.

That is worth doing for a reason beyond the picture. This page originally carried a hand-written table of armour areas, and drawing the cable to scale is what proved it wrong — it claimed 146 mm² of steel on a 240 mm² three-core where the geometry gives 248 mm². You cannot lay thirty millimetres of circumference with 1.6 mm wire and end up with that little steel, but you can certainly write it in a table and have it look reasonable. Where the armour figure matters, type the real one from your datasheet into the check and it will be used instead.

All of which makes these right for preliminary selection and wrong for a design you are about to sign. Real ratings differ between makers by a few per cent, and between construction types by more than that — XLPE insulation buys you roughly twenty per cent over PVC because the conductor is allowed to run at 90 °C instead of 70 °C. Take the size off this chart, then confirm it against the datasheet for the cable you are actually buying.

The derating is applied to every row at once rather than to one answer, which is the point of reading a chart instead of a calculator: you can see that 95 mm² is marginal and 120 mm² has room, and decide how much margin the job deserves.

What the armour is for

Steel wire armour exists to survive things. A spade, a fork-lift, a rat, a reversing lorry, the weight of its own vertical drop down a riser. It is the reason SWA can be buried direct, pulled into a duct, or run across a yard where a conduit would be pointless and a tray would be in the way.

But the armour is also doing a second job that nobody specified and everybody relies on: it is the earth. In the overwhelming majority of SWA installations the steel is the circuit protective conductor, connected at both ends through the gland, and there is no separate earth core at all. That is a perfectly good arrangement — and it is the one thing about armoured cable that wants checking rather than assuming.

The armour as the earth conductor

Two facts decide whether the steel is up to the job, and they pull in opposite directions.

The first is that there is a lot of it. On a small cable the armour dwarfs the conductors: a 2.5 mm² three-core carries about 20 mm² of steel around it, eight times the phase area. Nobody worries about the CPC on a 2.5 mm² circuit, for good reason.

The second is that steel is poor at this. Its adiabatic constant is about 51 against copper's 143, so a given area of steel absorbs roughly a third of the fault energy the same area of copper would. Effectively, that 8× of steel is worth under 3× of copper before you start.

And the ratio that saved you on the small cable erodes as the cable grows. Armour area rises with the circumference of the bundle while conductor area rises with the square of its radius, so steel cannot keep up with copper — the ratio falls from about 8× at 2.5 mm² to 1.4× at 95 mm², reaching parity around 240 mm² and dropping below it beyond that. Set the chart above to copper three-core and read the armour column down: that decline is the whole argument for checking large cables and not bothering with small ones.

Those two together are why the check belongs on large cables and short, hard faults. The calculation itself is the ordinary adiabatic one — where k comes from and how to get the clearing time right is worked through elsewhere and not repeated here. What matters on this page is the input people get wrong.

Use the earth fault current, not the three-phase fault level. They are different numbers and the earth figure is usually the smaller one, because the return path runs through the armour rather than a phase conductor. Reaching for the fault level off the transformer because it was on the drawing will oversize the answer, which sounds safe until it is the reason somebody specifies a separate earth core that was never needed.

When the armour does fail the check, the fixes in order of sensibleness are: clear the fault faster, run a supplementary earth alongside, or go up a cable size. Going up a size works — the armour does grow with the cable — but it grows slowly, and paying for copper you do not need in order to buy steel you do is a poor trade above about 120 mm².

The gland is the earth connection

This is where armoured installations actually fail, and it has nothing to do with the calculation above.

The armour only earths the far end if the gland clamps it and the gland is bonded to the enclosure. A CW gland grips the wires between a cone and a ring and is the type to specify for anything structural; a BW gland does the same job indoors. Either way the path runs steel → gland body → enclosure, and every interface in that chain is a mechanical joint that somebody made in a hurry.

• Fit the earth tag. A gland landing in a painted gland plate is not bonded to it — paint is an insulator, and the locknut biting through it is a hope, not a connection. The tag (banjo) and a short copper tail to the earth bar is what makes it real.
• Bond both ends. An armour earthed at one end only is not a CPC. It also becomes an aerial for anything inducing into it.
• Check the cone is seated. An armour cone that has not pulled the wires up against the ring will pass a continuity test at 200 mA and burn open at 6 kA. Continuity testing an armour path proves it is connected, not that it can carry a fault — the adiabatic check is what proves the second thing.

Cores, and picking between them

The 3.5-core is worth knowing if you are working to Indian practice and reading British charts, because it barely exists in the latter. It is a three-core with a reduced neutral, and it is the standard LT distribution cable here — cheaper than a four-core and perfectly adequate while the neutral only carries imbalance. The moment the load is single-phase heavy, or full of switch-mode supplies putting triplen harmonics into the neutral, the reduced neutral stops being adequate and you want the four-core.

Never put steel armour on a single-core AC cable. The alternating field induces current in the steel and it heats up with nothing to cool it. Single-core armoured cable exists, and its armour is aluminium for exactly this reason.

The constraints that are not electrical

On a real job the size on the drawing is as often set by what will physically go in as by what the current demands.

Bending radius. Six times the overall diameter is the usual figure for armoured cable, and on a 240 mm² four-core that is most of a metre of swept radius. Gland plates get laid out on the assumption a cable arrives straight at them; a cable that has to turn immediately inside the panel needs the room designed in, or it arrives on site and does not fit.

Burial. Direct burial wants depth, a sand surround free of stones, and marker tape above — the armour protects against a spade, not against being laid on rubble and then driven over. Where it crosses a road it goes in a duct regardless of the armour.

Pulling tension and vertical runs. Round wire armour takes tension; flat strip armour does not, which is why risers and duct pulls specify the round type. On a long vertical drop the cable also has to be cleated at intervals so the conductors are not hanging on their own terminations.

Reading this chart against a real datasheet

When you go to confirm the size, the figures will not line up exactly, and knowing which differences matter saves a phone call.

• Check the reference temperature the maker rated at before comparing anything. A rating at 30 °C ambient looks generous next to one at 40 °C and is the same cable.
• Check the insulation. PVC and XLPE ratings for the same conductor size differ by around a fifth.
• Check whether derating is already applied. Some datasheets tabulate in-air and in-ground ratings separately rather than giving you a factor, and applying a factor on top of a ground rating derates it twice.

The guide to reading a cable datasheet covers the designation codes — what YWY and A2XFY actually spell out — and the derating factor tables cover the corrections in more depth than the chart's four inputs allow, including the ones with no obvious box to tick.