Connecting a PV systems the right way

[Barx]

We have all heard and most have seen how NOT to connect PV. I know a few things about PV but here I sit starting to doubt myself. All the PV we have installed to date, we have done a fashion that does not require RCD protection (quite easy to achieve retro-fit). My thoughts, and the regs thoughts, say to use a B Type RCD..... Have you tried to get hold one?

Wholsalers....... Nope, never sold them

T'internet......... Struggle to find anyone dealing in them.

So, my question for this thread is, how do people fit PV properly when RCD is needed? i,e new build, re-wire, etc.

I have seen many installs using garage boards and AC type RCD, to me this is wrong. Wylex seem to offer some PV switchgear and they state that the unit can come with an DP RCBO but nowhere does it say it is a B Type and searching the t'internet return nothing for wylex b type DP rcbo.

Now, I have seen some inverter manufacturers state that an A type RCD will suffice as DC can only leak from +ve OR -ve giving the pulsing DC not constant. A lot of RCBO's are A Type, but I wouldn't like to hold this up in court.

So, without bashing me sensless with a stick, What configurations have you seen in real life that you deem right?

Answers on a postcard (or just reply to the thread)

Ta

 

[Essex1]

Surely the B type you need is for the OPCD and not the RCD.  A B Type RCBO yes but a B type RCD.???

 

[Barx]

It is a B Type RCD Essex, to deal with DC currents.

 

[Rob.]

Residual current devices may also be classified as Type a.c.,Type A and Type B as follows:

Type a.c. Ensures tripping for residual a.c. currents, whether suddenly applied or slowly rising.

Type A Ensures tripping for residual a.c. currents and pulsating d.c. currents, whether suddenly applied or slowly rising.

Type B Ensures tripping for residual a.c. currents, pulsating d.c. currents and smooth d.c. currents, whether suddenly applied or slowly rising.

@Essex1

 

[Essex1]

I knew I should have stayed out of this thread. 

This is what happens when you are bored and on a train into London. 

 

[Rob.]

I knew I should have stayed out of this thread. 

This is what happens when you are bored and on a train into London. 

I was just posting the information for reference that's all.

 

[Essex1]

I was just posting the information for reference that's all.

I know. Scoobed appropriately.

 

[Barx]

I knew I should have stayed out of this thread. 

This is what happens when you are bored and on a train into London. 

10 points for effort though pal :coat

 

[binky]

that's only about he RCD not tripping all the time as far as I'm concerned. I've always used 'normal' RCDs but not shared with any other circuit. All has been fine to date. The only inverter I have seen that was causing issues with an A type RCd was an Power One unit as mentioned in a previous posting whereby we were advised to ft a 300mA RCD (which did work incidentally), during which we concluded that DC faults can affect the AC side, and I know that array has DC issues with damp. Given the segragation within the inverter of AC and DC, I can't see how DC could ever be present on the AC circuit?

 

[Barx]

We had 1 install that had an AC RCD in a garage board. It's the only local private install we had done last year and the ONLY install we put an RCD in. No idea why the RCD was there as it wasn't required. Me thinking the lad that did it picked up the wrong garage board type and didn't take notice that it was an RCD (I have checked the others and all okm). Just happened to be the install I took MCS inspector to. Got an NC for it not being a B Type. Most TransformerLess inverters are deemed to not have seperation internally. Bloody typical...

Changed the RCD to main switch and NC cleared. Inspector reiterated the requirement for B Type when no seperation stated by manufacturer. BTW, this isn't the reason for this thread, inspection was a couple of months ago.

I have emailed a few manufacturer's asking if the built in RCD that most inverters seem to have cover faults AC side, none have replied. If it does then the way I see it is when the inverter is live and generating, the inverter RCD protects the circuit even if the Mains side one is locked in place with DC. Then when the inverter is dead, the mains side one works as required still protecting the circuit. There is no concrete evidence of this though and no-one is willing to accept it so all just say 'B Type required'. Like talking to robots.

Page one of the following doc shows the reason for needing it

http://www.doepke.co.uk/download/Techpub-04.pdf

 

[leesparkykent]

reg 712.411.3.2.1.2

 

[leesparkykent]

Does anyone have Februarys professional electrician magazine kicking about? I've chucked my copy but Elecsa have written an article about PV and RCD's IIRC page 37. 

 

[Barx]

Not seen it Lee, i don't visit wholesalers much these days to pick a copy up. Would be interesting to read that articla though. Can you remember the concensous of it?

 

[binky]

Interesting article from Doepke. I don't quite see how you get a DC earth leakage. Apart from the fact there is no earth connection on the DC side, the inverters self check as they start up, specifically looking for DC faults. Whether they maintain any form of monitoring after start-up I don't know, although we have earth leakage faults reported during daytime running caused by freak-weather events on SMA units. Given that daylight usually comes in gently, I would have thought a DC leakage would be detected before it reached more than 6mA? I know my Power One often reads 2mA on start up. But I suppose the argument is that on a functioning system, would the RCD do its job? I'm trying to think how the relatively high resistance of an Earth Rod would affect mA leakage, as this is about the only time I would fit an RCD.

 

[Barx]

There are loads of articles from others too @binky all saying same thing. Others also mention 'parasitic capacitance'. SMA explain it a bit here -> http://files.sma.de/dl/7418/Ableitstrom-TI-en-25.pdf

Here is another interesting point. Wylex do a 'PV Consumer Unit' (bottom of page 4 on this doc -> https://www.tlc-direct.co.uk/Technical/DataSheets/Wylex/WylexSolarPV.pdf) the interesting thing is they do an RCBO version. I have asked Wylex to confirm that the RCBO has B Type RCD parameters...... I just got set a datasheet that does not state the type. Just that the MCB part is 16A 6kA B Curve.

 

[leesparkykent]

ok found it.

Requirements of BS7671 connecting the PV a.c. mains.

At the connection to the a.c. mains, the PV supply cable must be connected to the supply side (i.e. upstream) of the OCPD's gor the final circuits in the electrical installation (regulation 712.411.3.2.11).

  therefore, the PV supply cable must be a dedicated circuit. It cannot be connected to an existing final circuit of the electrical installation.

When connected in this way, the operation of the protective devices of the other final circuits of the electrical installation is unaffected by the PV installation.

Fault protection

The PV supply cable on the a.c. side must be protected against fault current by an OCPD installed at the connection to the a.c. mains (regulation 712.434.1). A circuit breaker in a one way consumer unit fulfils this requirement.

General requirements

The addition of the PV installation must be selected and erected in compliance with the requirements of section 712 of BS7671 and other relevant requirements of BS7671.

Earthing & bonding

Regulation 712.312.2 states that earthing of one of the live conductors of the d.c. side is permitted, if there is at least simple separation between the a.c. side and the d.c. side. The note to regulation 712.312.2 states "any connections with the earth on the d.c. side should be electricaly connected so as to avoid corrosion (see BS 7361-1:1991)".

A risk assessment should be carried out to determine the presence of any exposed conductive parts or extraneous conductive parts.

where protective bonding conductors are installed, they must be parallel to and in close contact as possible with the d.c. cables and a.c. cables and accessories (regulation 712.54). Where pre-assembled cable harnesses are used, its important to ensure that the cable system complies with the requirements of BS7671.

When an RCD is required at the PV inverter end

An RCD may be required to provide either fault protection by ADS or additional protection in accordance with regulation 415.1. Where it has been determined that an RCD is required, it must be located at the PV inverter end of the circuit and the type of RCD selected must take the characteristics of the PV inverter in to account.

If the PV inverter, by its construction does not have simple separation between the d.c. side and a.c side ( this is some times called a transformerless inverter), a type B RCD to BS EN 62423 is required as it will detect a.c., pulsating d.c. and steady d.c. currents. Reference should be made to the manufacturer of the PV inverter for advice. Where a PV inverter has simple separation incorporated, the RCD does not have to be type B.

When an RCD is not required at the PV end

If the purpose of the RCD at the PV inverter end is solely to prove protection against impact, the RCD would not be required if the supply cable was run in a floor or ceiling void, or on the surface or in SWA cable.

Regardless of whether or not the PV inverter has simple separation incorporated, an RCD would not be required in such circumstances.

Shared RCDs are not permitted

At the a.c. mains end the PV supply cable must not be connected to the outgoing way of a final circuit which shares an RCD with other final circuits (see IET code of practise for grid connected PV systems page 88). The reason this is prohibited is that the other final circuits maybe supplied by the inverter, during the shutting down period. This would mean the required maximum disconnection times for these final circuits could be compromised. In addition, sharing an RCD may cause unwanted nuisance tripping due to accumulation of earth currents.

 

 

[Barx]

It is very interesting that they state 'at the inverter end'. I don't remember reading that in any other material I have gone through. What happens to additional protection when the inverter is dead? be it either no generation due to lack of sun or the built in RCD has detected a fault and tripped, this leaving a live AC cable burried in the wall for example with no RCD protection.

The mystery contnues...

p.s. thanks for digging it out @leesparkykent

 

[leesparkykent]

It is very interesting that they state 'at the inverter end'. I don't remember reading that in any other material I have gone through. What happens to additional protection when the inverter is dead? be it either no generation due to lack of sun or the built in RCD has detected a fault and tripped, this leaving a live AC cable burried in the wall for example with no RCD protection.

The mystery contnues...

p.s. thanks for digging it out @leesparkykent

No problems. The subject has been  much debated on another forum and after 100 posts it was still as clear as mud lol

 

[binky]

OK, 'at the inverter end' I'll have to double check data sheets but I'm fairly sure decent inverters have this in-built. SMAs have in-built DC string monitoring, so I'm assuming this is effectively RCD protection.

Moto:- avoid cheap Chinese inverters

 

[binky]

from SMA 3600TL data sheet:--

Protective devices
DC disconnect device
Ground fault monitoring / grid monitoring
DC reverse polarity protection / AC short-circuit current capability / galvanically isolated
All-pole-sensitive residual-current monitoring unit
Protection class (according to IEC 62103) / overvoltage category (according to IEC 60664-1)

I checked a few other makes of inverter, and all have some form of DC moitoring, but are a bit vague with regards to residual current. I would assuem they would monitor for DC leakage to prevent destruction of AC components? So I think answer to quesion is you need to check manufacturers instruction for each make of inverter.