RCD Protection on a PVinstall

[MorganPVI]

We've spent this week thrashing out the RCD issue and after an internal technical meeting this afternoon - based on test results and information supplied by inverter manufacturers, RCD manufacturers, electrical and Electronic engineers we will now no longer be fitting Type A 100ma RCDs on TL inverters. We will probably now fit 100ma Type B RCDs and ensure our designs do not require 30ma RCD protection.

If anyone wants our stock of Type A's please feel free to PM me.

 

[Worcester]

@edexlab,

There is a difference between 'the inverter' and 'the installation'

Whilst the inverter indeed may well have a built in RCMU that is not an RCD, in addition, many inverters are indeed designed to not inject DC into the supply under fault conditions in the inverter.

The problem comes when you have, over a period of time a DC leakage current slowly building up due to faults in the Array (even double insulated) then the RCMU in the inverter will be 'blind' to it.

The DC currents then saturate the coil in a Type A RCCB rendering them useless.

Type B RCCB's are a completely different design from Type A's, and they are aware and will monitor (amongst other things) DC leakage, and are designed to provide protection under these types of conditions.

The point is that the inverter manufacturers are correct, as far as their statements go. These statements DO not say that the INSTALLATION will not require a Type B.

The key thing is, now that everyone reading and contributing to this thread has been made aware, they cannot ignore what they have read.

Everyone should make their OWN judgement - and you need to be educated and competent to do that. You cannot rely just on what the manufactures (of RCD's or Inverters say) you should do your own due diligence.

To ignore this will make you personally liable, in the same was as not installing a 30mA RCD where required would.

 

[Scrag]

Agree to a point. But correct me if I am wrong. I we not to accept manufacturers recommendations above any regulations current. Thereby passing the buck to them. Just another paper trail exercise. Not that I agree with it of coarse.

 

[MorganPVI]

I tried to argue the point with our contracts guy, he basically said that argument wouldn't stand up up in court and our liabilities for not fitting the correct protection would be unacceptable. Then went on to tell me that Germany has set a precedented by requiring a Type B for TL inverters. Still the sales guy wants to spin it as a USP and make sure every customer wants one with their shiny 4000tl.

 

[JulianC]

Going back to the point of 30mA RCDs fitted for additional protection with buried cables on a TN system, when using transformered inverters (apologies if this has already been covered elsewhere) - if G83 allows up to 5secs inverter shutdown time in the event of loss of mains, isn't the RCD effectively useless? It may trip within the allocated time in the event of an earth fault, but the inverter will still be keeping the circuit live for considerably longer. Therefore, no buried twin and earth for PV. Or am I missing something?

 

[Markc]

Going back to the point of 30mA RCDs fitted for additional protection with buried cables on a TN system, when using transformered inverters (apologies if this has already been covered elsewhere) - if G83 allows up to 5secs inverter shutdown time in the event of loss of mains, isn't the RCD effectively useless? It may trip within the allocated time in the event of an earth fault, but the inverter will still be keeping the circuit live for considerably longer. Therefore, no buried twin and earth for PV. Or am I missing something?

This was exactly what was raised on Friday for our annual MCS inspection. Only applies where the incverter is connected to say a overall RCD device of 30mA where furthe rout going circuits require 0.4 disconnection time. Unless you can show via the inverter manufactures the voltage degregation on shut down is less then 0.4 then its none compliant to BS7671 and putting users at risk on say a shower circuit.

Where the outgoing circuits do not require 0.4 sec disconenction time it is fine to fit the inverter circuit to any avaliabel way even if shared with a overall RCD.

For example: A 17th edition consumer unit in a domestic property. All useable ways propted via 2 x 30mA RCD's. No none RCD protected ways. Socket outlet circuits, cooker, hob, immersion, lighting and shower all split over the two RCD's. Inless it is proven and documented that the inverter being installed will reduce in voltage on shut down in less then 0.4s then there are NO ways avaliable to connect it into. Only option is to install a ONE way standalone C/U on to the tails. RCD protected if needed to meets BS7671 for embedded cables or alike but NO possability of additional circuits being added to it in the future that require 0.4s.

 

[JulianC]

Only option is to install a ONE way standalone C/U on to the tails. RCD protected if needed to meets BS7671 for embedded cables

That's my point I guess. RCD protecting embedded cables to your inverter is pointless if your inverter is still kicking out 230v a few seconds after Johnny DIY'er has drilled through the cable and the RCD has tripped.

So, no embedded cables that would require 30mA RCD protection to be used to connect an inverter.

 

[Markc]

That's my point I guess. RCD protecting embedded cables to your inverter is pointless if your inverter is still kicking out 230v a few seconds after Johnny DIY'er has drilled through the cable and the RCD has tripped.

So, no embedded cables that would require 30mA RCD protection to be used to connect an inverter.

SWA would be option 2 then no RCD is required at all.

 

[BruceB]

or put an rcd on the output of the inverter (if it will allow one as low as 30mA)?

 

[dpelec]

sma tl inverters do not require a type (b) dc sensitive rcd to be installed as they provide by circuit design prevention from dc faults occuring onto the ac side.

files.sma.de/dl/7418/RCD-UEN110430.pdf take note on section 4

Also You do not need to bond the array frame metal work on a pme system if the array frame is not within the equipotential zone.

 

[Worcester]

@dpelec

Many thanks for that, good to see a manufacturer spelling it out clearly, with the technical reasons why. Interesting to see that they recomend that you install one in all circumstances, and as said before its 100mA per inverter.

 

[BruceB]

a blast from the past!

 

[dpelec]

Yeh hope that helps everyone as this issue has been left in the air since pv really got going back in april 2010