Cable ratings - constant load?

[danieleither]

For argument sake, lets say a cable is rated at 100A clipped direct, ambient temp 30 deg. Does this mean it can take a constant load of 100A in those conditions or is the rating based on cyclic current?

The reason I ask is on a recent install, following current cable guidelines we fitted a 35mm 4-core SWA XLPE cable to provide 150A per phase over a distance of 15 metres. At 150A constant load the cable has over-heated after 12 hours of operation (insulation melting).

I can't seem to find anything online about whether cable ratings are based on constant load or cyclic. Have we under-specced? If not, perhaps it's a damaged cable.

 

[pc1966]

Cable ratings are continuous. How was the cable routed?

Looking up table 4E4A for 90C insulation 35mm three-phase it has 154A clipped direct, down to 115A buried direct/in ducting. For 70C insulation in 4D4A it is 125A clipped direct, down to 98A buried direct / in duct.

It seems pretty marginal for 150A. Also if you have a very non-linear load you may see additional heating of the neutral as 3rd harmonics don't balance out but circulate via N.
[automerge]1590498764[/automerge]
Also remember if you are planning to run 90C cable close to its thermal limits you must also ensure that terminating accessories / joints/ etc are also rated to 90C.
[automerge]1590499084[/automerge]
Out of interest, what make of cable is it?

 

[danieleither]

The cable was clipped to a brick internal wall, cable cleats. Ambient temp around 23-25 degrees. The load was (theoretically) perfectly balanced as the load is 24 x identical high-power server units, 8 on each phase.

 

[pc1966]

The cable was clipped to a brick internal wall, cable cleats. Ambient temp around 23-25 degrees. The load was (theoretically) perfectly balanced as the load is 24 x identical high-power server units, 8 on each phase.

Did anyone measure the currents when it was running? Also what current was in the neutral?

Good power supply units now try to control harmonics, but in the past that would be a likely source of extra neutral heating.

 

[danieleither]

Did anyone measure the currents when it was running? Also what current was in the neutral?

Good power supply units now try to control harmonics, but in the past that would be a likely source of extra neutral heating.

Yes - we noticed the cable was hot after a few hours so used a 3-phase current logger to check what was going on (on each of the 50mm tails entering switch-fuse). Draw was pretty stable, fluctuating between 149.6A - 150.1A per phase. We did not monitor the neutral phase.

 

[pc1966]

What make of cable? Are you sure it was 90C rated and not 70C?

Has it melted inside or just at terminal ends?

Can you do an accurate DC ohms measurement to see what sort of conductor resistance it has? (just in case under-sized conductor or similar)

At full load it would run hot, after all if 90C inside then outer is going to be only a few tens of C below that. But if it has failed then it comes down to:

• Current too high (or not accounted for - neutral harmonic load)
• Resistance too high (under spec cable, high-R terminals overheating ends)
• Insulation not rated for temp
• Cooling inadequate

 

[danieleither]

What make of cable? Are you sure it was 90C rated and not 70C?

Has it melted inside or just at terminal ends?

Can you do an accurate DC ohms measurement to see what sort of conductor resistance it has? (just in case under-sized conductor or similar)

At full load it would run hot, after all if 90C inside then outer is going to be only a few tens of C below that. But if it has failed then it comes down to:

• Current too high (or not accounted for - neutral harmonic load)
• Resistance too high (under spec cable, high-R terminals overheating ends)
• Insulation not rated for temp
• Cooling inadequate

The cable didn't fail, but the insulation on the inner cores started to melt where it is connected in the switch fuse. Where two (insulated) cores were touching each other, the insulation started to melt between them. If we hadn't caught it, they would have melted through and shorted between 2 phases eventually.

Switch fuse is rated 160A. Connector screws torqued correctly, and checked again after a few hours of load. Initially it was conductor straight into the terminals, afterwards we tried crimping into copper blades to ensure full contact with all of the conductor.

Cable used: 35mm² 4 Core LSOH SWA XLPE Armoured Cable (Cut Length Sold By The Mtr) - https://www.cef.co.uk/catalogue/products/2412134-35mm-4-core-lsf-swa-xlpe-armoured-cable-cut-length-sold-by-the-mtr

 

[Marvo]

I'd also check the power factor

 

[pc1966]

Can you try the photos again? I can only see the thumbnail photos.
[automerge]1590504342[/automerge]
Curious, his post with photos has vanished completely!

 

[danieleither]

Can you try the photos again? I can only see the thumbnail photos.
[automerge]1590504342[/automerge]
Curious, his post with photos has vanished completely!

I edited the post to add the cable link. Try a refresh - it's showing here

 

[pc1966]

Nope, sorry no photos seen from my end

 

[Wilko]

Hi - as far as I know the intent of the BS7671 is to provide tables of ratings that the cable can withstand continuously within the temperature and installation constraints.

That an XLPE cable‘s insulation has melted indicates to me that either it isn’t really XLPE or the ratings have been significantly exceeded. It takes more than a small overload to upset XPLE, just saying.

I’m keen to know about the cable - can you confirm from it’s markings that it really is what you reckon?

 

[westward10]

XLPE insulation is rated not to degrade even at a constant 90°C temperature, for any deterioration to take place you have to consider temperatures in excess of 140°C. Overall the insulation appears to be visually okay aside from the damage shown. As already asked who manufactured the cable.

 

[Julie.]

If it’s server load as in a data centre type load then the current does not balance the same, servers can generate triplen harmonics, plus several others.

When you measure the current with a typical meter, it doesn’t always measure harmonics correctly, you need specialist equipment – basically many instruments merely measure the peak, and recalculate for a normal sine wave, other so-called true RMS actually only calculate of the fundamental and filter out the higher harmonics. Also – if you are using a clamp meter this often limits the higher frequencies.

If your load is circ 150A @230V then it’s about 103.5kW in total.

The IEC allows third harmonic current of 3.4A per kW – so up to 352A

The IEC allows 5th harmonic current of 1.9A per kW – so up to 197A

The IEC allows 7th harmonic current of 1.0A per kW – so up to 103A

The IEC allows 9th harmonic current of 0.5A per kW – so up to 51A



These would add to the fundamental current - so if you had full 150A fundamental (zero in neutral) you could have an additional 117A of third harmonic in each phase, plus 352A in the neutral etc.



Whenever I had to deal with data centre type loads I have always requested the actual harmonic content from the server people – they usually have this because it’s a known issue, and usually they like to show their equipment is well below the limits above!

 

[pc1966]

As already asked who manufactured the cable.

They linked to the CEF site which says nothing about the manufacturer.

However, I do happen to have a short bit of the non-LSF version of that cable from CEF and it says on the side "EL SEWEDY CABLES" which I think are an Egyptian manufacturer.

It is interesting that the neutral is one showing the melting action after what Julie just posted!