Cu & Pv Configuration

[trevor_clark]

thanks to all the guys that gave advice, found it most helpful, will be going down the split meter tails & mini CU route.

regards

Trevor

 

[the doctor]

Just to add to the above, it's because when someone gets a shock and the RCD trips, the Solar PV inverter will continue providing power for up to 5 seconds before it detects loss of grid connection and shuts down. By which time your poor victim is probably dead.

Sadly it seems a lot of Solar installers just don't understand this.

Hi Prodave,

This is most interesting and I understand your point.  Is it a regulation though? I dont see it in BS7671 except a half point at it 551.7.2 (v)  In NI we do quite a few TT jobs, though we usually end up with a new consumer unit on its own...

 

[Andy™]

712.411.3.2.1.1 - must be on supply side of any protective device for ADS of other circuits (short text of it)

about as clear as you can possibly get it, yet the monkeys who install it still ignore it...

 

[steptoe]

Some inverters CANNOT be used on TT at all.

 

[the doctor]

712.411.3.2.1.1 - must be on supply side of any protective device for ADS of other circuits (short text of it)

about as clear as you can possibly get it, yet the monkeys who install it still ignore it...

This is one of these moments in life, when one has what you would term an epiphany. With Andy's and Prodave's comment and explanation I now fully understand the problem and regulation.  If you know why something is written then you fully understand it.  

Worse still, no-one has ever explained it to me though it is now as plain as the nose on my face.  After consulting the ECA's 'guide to the installation of photovoltaic systems' I can see it does not mention this now obvious fact.  G83 requires a 5 second cut off while BS7671 requires far less. I now understand why the regulation is written can report that I had never thought about it that way before.  I cannot say for certain if I have or have not done any PV jobs 'wrongly' in the past.

Just goes to show you..... thanks for the lesson.  The story will be in my next 17th edition course if I ever get to do one again... :facepalm:

 

[Andy™]

dont forrget the 5 seconds is after the ADS time... so if you have an RCD that must trip within 300mS, then the total time could be 5.3s... a long time to be holding onto a live wire...

 

[Sidewinder]

According to the scams though it's a non problem!

SERIOUSLY!!!

 

[the doctor]

Im humbled here........ never saw it.... now i do....

So it would be an issue on a consumer unit with an rcd main switch...  a normal consumer unit would not be an issue, though it would break 712.411.3.2.1.1

 

[Sidewinder]

The issue is that IF the PV feeds in down stream of any final circuit protective device.

The protective device for the final circuit could trip, to protect the innocent, whether that device be MCB, or RCD.

Then the Micro-generation can shut down 5s after loss of supply, so for a "large" final circuit, that could be 10s after the fault occurred.

For an RCD protected circuit with someone holding on to the metal nail they have just hit into the live conductor of the flat twin cable, the RCD say trips in 300ms, the solar micro-generation can keep back feeding the circuit for up to 5s after the loss of supply.

It, may, shut down sooner, but, the product standard only requires shut down in 5s.

 

[the doctor]

Hi Sidewinder,

In Northern Ireland it is common to fit panels on a shed, which is fed via a sub main from the house.  And that sub could be backed up by an RCD.  Effectively then this would be a no no... yes?

 

[binky]

on decent inverters they tend to shut down pretty pronto, but I doubt its milliseconds, however many companies fit cheap Chinese units, which I doubt are so responsive - not really tried these so can't comment.

 

[Andy™]

Effectively then this would be a no no... yes?

Correct

 

[Phoenix]

Hi Sidewinder,

In Northern Ireland it is common to fit panels on a shed, which is fed via a sub main from the house.  And that sub could be backed up by an RCD.  Effectively then this would be a no no... yes?

As to the RCD issue, I am with the majority regarding the fact that it should not be read from the same RCD as other circuits. (however there is a bit of a saving grace in that when the RCD trips * that the suppliers N-E link is gone and the inverter is basically a separated IT system until it shuts down)

*if it does... as has been alluded to the solar PV is more than capable of de-sensitising or completely locking the RCD

As to the submains, I suppose you could look at the actual disconnection time - likely to be 0.1sec if MCB, look at the inverter specs, find out the actually shut down time (which might be significantly below the 5s the standard requires) add these together and if you are still less than 5s required for the submain circuit  then there isn't a really a problem.

The other think to bear in mind is say a fault trips the 32A breaker feeding the submain, its not true to say that any touch voltages will remain on the metalwork until the inverters control logic detects the loss of grid and shuts it down, the output of the inverter would be dragged down to just about nothing when presented with a fault of low enough impedance to have tripped an MCB... theres just not enough energy there to maintain the output in the face of the fault.

 

[Andy™]

As to the submains, I suppose you could look at the actual disconnection time - likely to be 0.1sec if MCB, look at the inverter specs, find out the actually shut down time (which might be significantly below the 5s the standard requires) add these together and if you are still less than 5s required for the submain circuit  then there isn't a really a problem.

even if it will trip within 5s, it still doesnt comply with 712.411.3.2.1.1

 

[the doctor]

Andy, Sidewinder et al....So at home I have a 40A mcb feeding my barn,  with a 16mm 3 core armoured running out there.  At the far end, there is a consumer unit with a 30mA RCD main switch.  It feeds a number of breakers including one to the pv unit, which is not fitted yet.  Reading 712.411.3.2.1.1 if I feed the PV breaker off the incoming side of the RCD, would I be off the hook here as the regulation refers to final circuits and I am on a distribution leg? 

 

[Sidewinder]

You need to look at the required disconnection times, and see if the circuit protection on the DNO supply, and the micro generation supply both meet the requirements for BS7671, and compare these with those documented for the inverter.

Also, if you are going for FIT, you need to consider the requirements for that also.

I don't know the scenario you have to say from here.

 

[binky]

why not swop main switch RCD for a main swith and add RCBOs for circuits you need 30mA protection on, or meter tail blocks and Mini CU for solar.

what's your house supply on earthing wise

 

[the doctor]

As to the RCD issue, I am with the majority regarding the fact that it should not be read from the same RCD as other circuits. (however there is a bit of a saving grace in that when the RCD trips * that the suppliers N-E link is gone and the inverter is basically a separated IT system until it shuts down)

Ker pow Phoenix is correct.  If you drive this imaginary nail into a wire and the RCD trips, the inverter then becomes electrically separate so you cannot get a shock to earth.  I wonder is why this regulation 712.411.3.2.1.1 is effectively ignored by the industry perhaps,,, :innocent

why not swop main switch RCD for a main swith and add RCBOs for circuits you need 30mA protection on, or meter tail blocks and Mini CU for solar.

what's your house supply on earthing wise

The house, in Ireland,  is a *******ised TNCS which they call a ' neutralized supply'  with an installation earth rod fitted.  Don't fancy a mini cu and I cannot afford RCBOs :facepalm:

 

[the doctor]

You need to look at the required disconnection times, and see if the circuit protection on the DNO supply, and the micro generation supply both meet the requirements for BS7671, and compare these with those documented for the inverter.

Also, if you are going for FIT, you need to consider the requirements for that also.

I don't know the scenario you have to say from here.

There is no FIT here, just the free energy.  It is standard to Irish which is more or less the same as the UK

 

[steptoe]

Ker pow Phoenix is correct.  If you drive this imaginary nail into a wire and the RCD trips, the inverter then becomes electrically separate so you cannot get a shock to earth.  I wonder is why this regulation 712.411.3.2.1.1 is effectively ignored by the industry perhaps,,, :innocent

The house, in Ireland,  is a *******ised TNCS which they call a ' neutralized supply'  with an installation earth rod fitted.  Don't fancy a mini cu and I cannot afford RCBOs :facepalm:

thats a big if

I see loads of these in council properties where the RCD simply doesnt trip while the PV is energised.