DC Cable buried <50mm in wall

[BruceB]

Yes I agree, I was concerned that he was pursuaded by my line of reasoning, but I did give him plenty of opportunity to come up with something else. It is of course possible that if someone else asked a different helpline guy the question a slightly different way they would get a different answer.

 

[SolarCity]

Interesting thread.

I suppose an ideal solution would be to have a device at the PV end of the DC cable which would operate under short circuit conditions between the + and - and would detect leakage to the outer sheath.

However, in my opinion I would say that the risk of electric shock is negligible in any case.

 

[dougalthedog]

The DTI guide recommends that long (no definition) DC runs should be labelled - this would seem like good practice if the cable is hidden.

The new draft guide says DC cables should not be buried, but then goes on to say that if this is unavoidable, SWA should be used.

My concern would be that a non electrically aware person could drill into the cable, shorting earth to one of the DC lines, then drills another hole and makes a circuit via tool, himself, ladder and earth!

Surely the purpose of equipotential bonding is to cause the supply to be disconnected in the event of a fault - in the scenarios we are discussing here the supply will not be disconnected which could result in shock or fire.

 

[SolarCity]

My concern would be that a non electrically aware person could drill into the cable, shorting earth to one of the DC lines, then drills another hole and makes a circuit via tool, himself, ladder and earth!

Surely the purpose of equipotential bonding is to cause the supply to be disconnected in the event of a fault - in the scenarios we are discussing here the supply will not be disconnected which could result in shock or fire.

The purpose of bonding is to limit the magnitude of any shock - by keeping everything at the same potential as much as possible.

It is earthing which gives us disconnetion.

Without an additional protective device there is no way to achieve disconnection from the DC side.

 

[moggy1968]

The DTI guide recommends that long (no definition) DC runs should be labelled - this would seem like good practice if the cable is hidden.

The new draft guide says DC cables should not be buried, but then goes on to say that if this is unavoidable, SWA should be used.

My concern would be that a non electrically aware person could drill into the cable, shorting earth to one of the DC lines, then drills another hole and makes a circuit via tool, himself, ladder and earth!

Surely the purpose of equipotential bonding is to cause the supply to be disconnected in the event of a fault - in the scenarios we are discussing here the supply will not be disconnected which could result in shock or fire.

Earth is a poor conductor, so in your example there wouldn't be a shock hazard. Try holding a + & - cable in each hand and touch them to earth with a bit of seperation and I don't think you'll get a shock! Just don't hold the 2 bare ends!
I was pinged on my first inspection for not labelling visible DC cable along it's length, which was all of 3 metres in a loft. The reason, perfectly reasonably, is it looks too much lile sat cable or coax. Any DC run should be labelled. We also label conduit which contains DC cable.
I think when you say burried cable should be SWA what the guide is refering to is buried in the ground, not buried at less than 50mm in a wall.
You can't create equal potential between DC and AC so equipotential bonding wouldn't help and is a different matter to earthing a fault current.

 

[dougalthedog]

"Because PV array cables almost exclusively rely on double or reinforced insulation as their means of shock protection they should not be buried in walls or otherwise hidden in the building structure as mechanical damage would be very difficult to detect and may lead to increase instances of shock and fire risk.
Where this cannot be avoided conductors should be suitably protected from mechanical damage, suitable methods may include the use of metallic trunking or conduit or the use of steel wire armoured cable."

Here's the text from the draft 3rd edition - it is referring to walls.

I'm not sure I get the bit about touching the cables to earth - in this case your body forms no part of the circuit. However, if one line of the DC run has been earthed and you then touch the other line, surely you are liable to get a shock (perhaps of a lesser magnitude due to the earth resistance) but if you've ever had a DC shock of even a few mA I suspect you'll agree that this could be somewhat dangerous if up a ladder! It would of course require you to touch the second line without shorting it to earth via the SWA armour / braid but if drilling / inserting rawplugs etc. surely this is a possibility. Perhaps I am simply coming up with a scenario which is too unlikely to worry about?

Noted that bonding is about limiting the shock, but bonding is usually used in conjunction with earting to achieve disconnection and in this case there will be no disconnection unless the inverter flags an error which someone then investigates.

It would seem that clear labelling would address most of the issues, however the labelling would need to be done in such a way that it will last 20 years and to go back to the original poster who wanted to bury the cable for aesthetical reasons, I suspect a well installed conduit (hidden behind a downpipe) would probably look better than a load of big WARNING labels!

 

[moggy1968]

I don't beleieve you will get a shock, that was the point of my post. Earth resistance is too high. I think you are getting into the realms of 'well if this happens under these circumstances when this is happening and it's the third tuesday of the month'!
The regs require protection under single fault conditions so lets not over complicate things.

Whatever the draft says is irrelevant. We're installing to current requirements, not what may be allowed in the future. Unfortunately the current guidence seems a bit ambiguous. Hard to believe I know!!!

 

[Glen Powell]

Sorry, not read all the thread, We have to put a 'B' type RCD in on the AC side of the circuit and I am sure it would be good practice to put it on the DC side of it in this case along with all the other protection.

DC current has come under HSE in many cases as there is not enough provision for it in BS7671:2008 although the BS7671:2008 including amendment 1 2011 has a little more but nothing to help in this case I do not think, again the book is in the office so cannot scan through it.

 

[moggy1968]

not sure how an RCD designed for AC will help on the DC side.
Wether or not you need a B type RCD depends on the inverter used. An A type may be adequate.

 

[Glen Powell]

A "B" type is used on TL inverters because it is capable of detecting fault on AC and DC should there be a bridge between the 2, an A type would be better as this is based on pulsating currents.

A "B" Type RCD is able to deal with AC and DC however it is based on sinusoidal currents.

 

[moggy1968]

If you can get a cert from the manufacturer though saying the inverter cannot dump Dc onto the Ac side then you don't need a B type, as I understand it (although there have been lots of threads on here arguing either way that seems to be the consensus)
As I understand it though a B type is for use on AC circuits where DC interference is possible because it is unaffected by them. As far as I am aware it doesn't provide earth leakage protection on purely DC circuits, but am prepared to be proved wrong of course!!

 

[Gavin A]

hmm... I'm sure the installation manual, or some other SMA technical document stated that the 4000TL wasn't capable by design of outputting DC current, or some such wording that meant it didn't need a type B RCD, but I can't actually find the wording now.

I believe it's because it has an internal RCMU that automatically cuts the inverter out if it detects over 50mA DC fault current on the AC earth circuit.

 

[Glen Powell]

As far as I am aware it doesn't provide earth leakage protection on purely DC circuits, but am prepared to be proved wrong of course!!

I will be honest and at this point and time have no evidence to prove it, my suggestion is just that, a suggestion. I have never come up with the instance where we have had to bury a cable in a wall so was not really offering expert advice, however if you formulate an "A" type RCD is better and will do the job then the goal is achieved.

If you feel RCD protection of any kind is not suitable then more fool me for suggesting it.

However between us I would say a lot are not aware they are meant to use a type "B" RCD and on the quiet pooing their pants, even though (i not looked) there are probably many threads on this.

 

[Glen Powell]

hmm... I'm sure the installation manual, or some other SMA technical document stated that the 4000TL wasn't capable by design of outputting DC current.

If it TL then it needs a Type "B" RCD, you can chance it with the fact that the manufacturer said, its called best practice. IMO all TL inverters must have a type "B" RCD.

Under the new DTI decision tree you will not have to earth all TL arrays bit I am not sure it is published yet...

 

[Gavin A]

If it TL then it needs a Type "B" RCD, you can chance it with the fact that the manufacturer said, its called best practice. IMO all TL inverters must have a type "B" RCD.

best practice is to follow the regulations and manufacturers guidance, and I've just found the relevant document, which doesn't support your case.

It seems I was wrong about the RCMU, and SMA isn't relying on this to prevent the DC fault current feeding in to the grid, but uses this as an additional safety measure.

4.1 Requirement as per DIN VDE 0100-712 (IEC60364-7-712:2002)
If intended as fault protection (see Section 2.2 "Automatic Disconnection via Residual-Current Device" (page 6)), DIN VDE 0100-712 requires a type B residual-current device for transformerless inverters.

This requirement also applies to inverters with HF transformers, since there is no galvanic isolation between the
AC current side and the DC voltage side.


One exception to this is if the manufacturer of the inverter can exclude the possibility of DC residual currents in the system. If necessary, type A residual-current devices can be used.


All SMA inverters with transformer, including SB 2000HF-30, SB 2500HF-30, SB 3000HF-30, and the
transformerless SMA inverters listed below are not capable of feeding-in DC residual currents due to their
design. They fulfill this requirement in accordance with DIN VDE 0100-712 (IEC60364-7-712:2002).


Sunny Boy:
SB 1600TL-10, SB 2100TL, SB 3000TL-20, SB 3000TL-21, SB 4000TL-20, SB 4000TL-21, SB 5000TL-20,
SB 5000TL-21
Sunny Mini Central:
SMC 6000TL, SMC 7000TL, SMC 8000TL, SMC 9000TL-10, SMC 9000TLRP-10, SMC 10000TL-10,
SMC 10000TLRP-10, SMC 11000TL-10, SMC 11000TLRP-10
Sunny Tripower:
STP 8000TL-10, STP 10000TL-10, STP 12000TL-10, STP 15000TL-10, STP 17000TL-10, STP 15000TLHE-10,
STP 20000TLHE-10
The possibilities of faults were examined without taking the integrated residual-current monitoring unit (RCMU)
into account. When examining these faults in terms of the currently valid installation standards, no danger in
combination with a type A upstream residual-current device can occur. Accordingly, faults that would otherwise
require the use of a type B residual-current device due to the inverter can be excluded.


The integrated all-pole sensitive residual-current monitoring unit (RCMU) results in additional safety. For
inverters with grounding conductor monitoring, this must be activated. These statements also apply to versions
of the listed devices with deviating power

[source = SMA]