Change in Standards

Dear all,

A quick “public service” announcement.

The new “PV standard” SANS10142-1-2(2021) has been published and is now law.
All new installations and any alterations on old systems will need to comply.

Nothing too drastic at least slipped in.
The following is the main points:

    1. Labelling requirements have been updated dramatically and I would recommend budgeting significantly more for labelling as everything needs to be labelled twice at least.
    1. Requirement for a Disconnector has been brought in, with the following statement: “This disconnector shall be lockable and accessible to the supplier, which is typically located within or near the boundary of the owner’s property.”
    1. Single Line Diagrams are now to be displayed at points of isolation/control (e.g. at main DB)
    1. The installer shall ensure that a Planned Maintenance and Operations manual for the specific system is handed over
    1. The new Test Sheet in 10142-1-2 shall form part of the CoC and shall be done by the installation electrician.
    1. You need metal conduits if DC cables are longer than 5meters.

Make sure that your local electrician friends doing installations are aware of this, if they are not then they will get into trouble.

I find it strange that the PV industry bodies have not made public announcements. Or perhaps I am just not important enough to get informed.

Cheers and no, I do not know if Pr. Eng/Tech sign-offs will now be a thing of the past. CoCT always said that should the PV standard be accepted that they will re-evaluate the Pr. Eng/Tech. requirement.

1 Like

Does that include PV panel cabling?

Is there an announcement somewhere? Where did you get this information from?

Not sure of announcement, but it’s on the SANS website now, wasn’t a month ago:

Yes: The DC system where applicable shall be protected against the effect of induced voltage surges by, implementing of the following wiring measures:

a) keeping the DC cable as short as practically possible;

b) installing or running the positive and negative cables alongside each other, to prevent the formation of induction loops;

c) the bonding conductors used to bond metallic parts in the array e.g. PV module frames, PV module supports and wireways/trunking shall be run as close as possible alongside the live conductors of the array; and

d) using screened/armoured cables, or to installing the cable in earthed, metal conduit/trunking where the DC cables are greater than 5.

1 Like

Thanks Karel, will implement on my current installations in Capetown.

In our conversation a few weeks back we did speculate this will follow sooner than later. I wish they postponed it to the new year though…

Man, most of the armoured cables lately only have a 600volt rating, not the 1000volt it had a few years back. My current installation will exceed the 600volt otherwise I could have used normal 4x4 armoured cable. It will be cheaper than using PV cables in Steel conduit, and where on earth will we find Hickeys and pipe benders these days.

Erm… I quoted verbatim above, but now I’m not sure… 5 meters or in a bundle of 5 or even mm2?

Seems there’s a cost advantage to interpreting this as a bundle instead of a length, but it doesn’t actually say which.

FWIW, this is why you don’t sneak new things into a standard when publishing v1; you’re supposed to just rename the final draft. Alas.

One would think that if length were implied “greater” would be replaced by “longer”.

I so hope its rather quantity. 2 Strings and an earth… That would be fun, and no change at all.

1 Like

Changing the conduits on an existing installation, if for example you wanted to apply for feedback, would be nuts.

If yours were signed off before the regulation changed, you would not need to make changes.

Well, you already need to bring a place up to current standards when you sell it as part of the COC process, so it’s not unheard of.

And I’m guessing this will be required retroactively since it was in draft so long. Another reason why I’m annoyed this conduit thing slipped in now.

But definitely a reason to not touch a signed-off installation, at least until you try to sell the house.

I’m keen to apply to allow feedback on my installation, but it seems it won’t be worth it if I have to change conduits. Labelling etc. is ok, but conduits nyet.

You can use Bosal’s conduit system but I reckon you get better value with Cabstrut P8000 trunking.
Are you allowed to mix AC and DC wiring in these?

So the greater than 5 thing is confusing as said earlier. Is it perhaps more than 5 wires in a single condiut? I.e 4 wires is okay but 6 a no no.

Quantity 5?

Is an earth wire counted as a DC cable?

If yes, ok so two pairs and an earth?

If no, how can dc cables have an odd number unless they mean pairs in which case we can do up to 5 pairs before needing conduit?

Honestly, how the fsck do these guys put something as ambiguous as this into effect without the stakeholders being aware? AREP had no idea. PVGreenCard told us nothing. SAPVIA said nothing.

This has massive legal and practical implications for installers and we find out because it’s suddenly available on the SABS website?

No, no and please do not! SANS10142-1 should also be read in conjunction. It is never good practice to mix AC and DC. Also your AC and DC does have different insulation levels on the cables. The “spirit” behind the clause in SANS is to avoid different insulation levels. So strictly speaking this would apply to Battery Cables (100V rated) and PV string cables (1500V rated).

So the way I read it is 5m, so length terms. Why am I reading this as such? Due to having talked to people on the committee that has been part of the development. I have requested clarification as well.

The first draft stated the following: “Where conduit is used for DC cables, it shall be metal conduit. Conduit shall be bonded at all junctions to ensure electrical continuity of the conduit. The metal conduit shall be bonded to earth at a suitable location.” It was then changed to state a limit. I was told the limit was set to 50m, so I am surprised that it now states 5. I think this might have been a drafting error in the document (note the document is rife with cross-references not being there and various other errors.

Note that everyone on the list here was on the drafting team. So they are aware of this, not sure how the final version came to be. If you have direct access to Carel Ballack please ask him as well. I think they were blindsided as well. This standard was planned for mid 2022, so how it came to be this year is strange.

Yes, this is very unfortunate and I find it strage that there was no fanfare!

1 Like

I think if you assume point B then you cannot get an odd number, and if you add point C then you get your odd number, yip we got this nailed lol

Yes I’ve been speaking to Carel all afternoon. He’s aware and investigating.