The definitive answer r.e. disconnection times and final circuits on PV inverters

[kme]

OK - this comes from the man who wrote the NICEIC PV training course (nice guy, too).

The fact that the inverter will continue feeding for five seconds isn`t an issue to your other final circuits; providing the upstream RCD is double pole.

The reason is that, when the RCD operates, it breaks the neutral path out to the star point of the transformer, and therefore breaks the earth path.

The inverter is producing its voltage between phase and neutral; neither of which will now be referenced to earth.

So there is no fault path between live and earth - so there is no risk of that shock voltage continuing whilst the PV "over-runs"

I think I`m happy with that. It may not be ideal, but apparently it IS compliant.

p.s. Albert: sorry

 

[ProDave]

So what about if the shock you are receiving is because you have one hand touching live, and the other hand touching neutral? Are YOU happy for that to continue for another 5 seconds before the inverter shuts down?

It seems to me that EVERYTHING about wiring regs assumes all faults are from L to E and gives no consideration for a L to N fault.

 

[steptoe]

no worries KME, so how come if you disconnect the neutral only then you can still get a reading from the inverter to earth?

try it, put you voltage tester across N-E on an inverter supply, and disconnect the N, the inverter doesnt shut down, its still getting a path back, which is a very likely scenario under fault conditions, imo

obviously this would be very dependent on the resistance of the circuit across N-E.

---------- Post Auto-Merged at 18:38 ---------- Previous post was made at 18:37 ----------

pushy shovey Dave.!

 

[kme]

Fair points guys - I`d simply repeated what was said on the `phone, without thinking it through too much.

I was about to say that a P-N shock could still continue under "normal" conditions.....the MCB isn`t going to save you; is it? Unless sufficient current is diverted to earth AS WELL.

Albert: I`ll put that to him (he gave me his email addy - doesn`t know what he`s let himself in for ] ) :slap , and see what he says.

 

[steptoe]

to be fair, Ive been trawling this past couple of hours and I cant find a definitive answer anywhere,

but,

see my other post, :_|

---------- Post Auto-Merged at 21:53 ---------- Previous post was made at 20:03 ----------

right,

from my extensive studies so far on this,

unless otherwise instructed to do so, the best advice I could offer anyone doing a PIR on a property with solar PV installed is this,

extent of installation covered = >xx% of standard installation, [or whatever you normally write]

limitations = solar PV installation not tested or inspected, cct x isolated and locked off for the duration of the testing and inspection.

this is going to be my standard procedure on installations with solar PV from now on.

 

[steptoe]

people dont want to be told that their PV install doesnt comply, which means the other few hundred properties that the same muppets installed are probably also non-compliant as well,

plus the fact that no-one (in authority?) seems to want to bite the bullet and commit to saying a 5s disconnection time circuit installed on the load side of a 0.4s disconnection RCD compromises the installation.

it all comes down to everything only being a recommendation once again, and different interpetations , makes you wonder why we have regs at all TBH.

 

[kme]

I`m amazed - reply back already

Under normal circumstances, when a fault occurs there will be a loss of the line conductor only on a circuit protected by a single pole device and both line and neutral via a double pole device such as an RCD and the disconnection times in BS 7671 are applied to this condition. We are not talking about just a losing the neutral only

 

[kme]

Yup - me too

 

[steptoe]

Im still unsure as to how you couldnt get a shock L-E though between the time the RCD trips and the inverter shuts down.

that is, a faulty DIY fitting where (as we have recently seen) line is connected to the metal casing, mrs brown is changing the 'bulb' and touches the casing, gets a shock and the RCD trips due to the imbalance to E,

now, Mrs Brown is still referencing Line to E to why wouldnt the inverter keep juicing here for the next 5s?

or as Dave asked, L-N, the inverter doesnt care where the current goes surely, it still takes 5s to realise the mains has been switched off,

still, good work KME,

at least someone in there cares and isnt just rolling out the usual company line.

 

[steptoe]

d'you know the one really pi ssing off thing about this whole install?

it was a H-I board with spare ways before the RCDs,,,,,,,,,,,,,,,,

 

[kme]

Im still unsure as to how you couldnt get a shock L-E though between the time the RCD trips and the inverter shuts down.that is, a faulty DIY fitting where (as we have recently seen) line is connected to the metal casing, mrs brown is changing the 'bulb' and touches the casing, gets a shock and the RCD trips due to the imbalance to E,

now, Mrs Brown is still referencing Line to E to why wouldnt the inverter keep juicing here for the next 5s?

or as Dave asked, L-N, the inverter doesnt care where the current goes surely, it still takes 5s to realise the mains has been switched off,

still, good work KME,

at least someone in there cares and isnt just rolling out the usual company line.

It took me a couple of minutes mate - but what he was getting at was this:

The PV is pushing out its voltage between phase and neutral - the inverter doesn`t bond the neutral down to earth.

The ONLY point where the neutral is earthed is at the star point of the Tx, or during the PME transmission (sake of argument - CNE)

If the RCD operates, and is double pole, the neutral bar isn`t connected to the incoming supply neutral any longer - therefore the 230V on the phase conductor can only be referenced to the neutral conductor, not the earth conductor, because the inverter doesn`t ground its neutral. The link between the neutral and earth has been broken by the neutral pole of the RCD.

Does that help??

---------- Post Auto-Merged at 00:05 ---------- Previous post was made at 00:05 ----------

d'you know the one really pi ssing off thing about this whole install?it was a H-I board with spare ways before the RCDs,,,,,,,,,,,,,,,,

The plot thickens........

Were they DP or SP?

Does it matter?

Is it a "proper" H-I board, full of RCBos? ; or a pretend HI board, with RCDs?

If its the first - the RCBo`s will provide disconnection time for the final ccts.

And anyway - it seems that 411.3.2.4 has to go out of the window - how the heck can you possibly comply with that??

I`m of a similar opinion - limit it out.

 

[steptoe]

the board was dual RCD with unprotected ways between the main switch and RCDs,

there is actually a pretty good drawing in one of the guides canoe as to why it must be DP protection,

I still dont like it,

its going to be a LIM for me from now on unless specifically asked for, as for what Im doing right now Im neither being paid for this amount of hassle, or am I in need of it.

 

[kme]

BPG - p13

 

[johninlondon1]

Maybe you lot should get together and do a guide on testing and inspecting PV.

 

[kme]

Yes....but the problems are myriad.

For instance - my own local AE came back to me this morning, with a couple of points.

1. it HAS been addressed in sec. 7 of the BGB - where they now stipulate that connection must NOT be made downstream of any protective devices serving other final circuits.

2. I repeated to him what the other AE said last night about breaking the N path back to the star point, and therefore the earth wasn`t referenced as far as the inverter was concerned - HE isn`t happy with that, either.

3. BUT - he couldn`t make a decision on, as part of an EICR, is it a C3 or a C2? Yes, it needs investigation (but does it? We KNOW what the problem is!) ; but is it likely to cause danger?

4. Albert - regarding limiting the EICR to not include the PV system - problem is that, when you inspect the CU, you can immediately see the problem, so you know it doesn`t comply. Therefore you would have to at least make a note of it on the cert.

KME

 

[steptoe]

1) ah, so if it has been installed post 1/1/12 then it doesnt comply?

2) we still have no answer

3) is C3 simply a cop out saying let someone else decide, Ive now brought it to your attention so its not my problem anymore

4) can we limit testing HI Bays yet we can still see the lid missing from the lighting trunking and just bits of tape here and there?

maybe no-one wants to stick their neck out,

I know for a fact a lot of the ones I will come across have been installed since January.

---------- Post Auto-Merged at 14:01 ---------- Previous post was made at 13:58 ----------

EDIT,

well it deffo has to be at least a C3 in every case, it no longer complies with the current version of the regs,

is it then down to the inverter manuf to supply some specs stating whether or not it constitutes a danger or not?

do you think the manuf will commit themselves in that way?

 

[Sidewinder]

I'll bet a

 

[steptoe]

that formed part of my argument SW,

without the installer supplying me with his installed specs then I have to apply the regs which IMO make the install non-compliant

and I cant in all honesty see any manuf taking responsibility and stating their inverter is safe installed in this manner, they wont want the liability.

 

[Phoenix]

I see the point about the double pole RCD breaking the neutral, but isn't there another issue to consider:

If we assume that (for simpilicity sake) that half the load of the installation is being met by the grid and the other half for the inverter, and the inverter is feeding in after the RCD, then if a leakage fault to earth occured (say someone picked up the end of a mowed over mower cable and touched the end) and 40ma flowed to earth, then as its a 50% split at current moment in time, then thats 20ma from the grid and 20ma from inveter, but inverter is feeding in after RCD, so only 20ma is passing through the RCD, so the 30mA rcd will not trip even though there is a leakage of 40ma going on.

I don't like it one bit, and I've never installed PV on the rcd section of a split board

 

[ProDave]

I predict in about 2016 when amendment 2 of 17th edition comes into force * , it will demand microgeneration facilities are fed in via their own dedicated CU.

Very useful debate and by the sounds of thing there are a LOT of non compliant installations (like the one I'm going to fix on Tuesday where they disconnected the immersion heater to feed in the PV to that mcb)

* = total guess at when amendment 2 may happen