Recently I have buddied up with another Electrician for solar PV work. The other Electrician is insisting on the use of a 30 mA RCD for fault protection on the AC side.

The consequence is that the contractor is reviewing his past and certified installations where a BS 60898 16A B-type CB has been used - which is extensive.

The preferred inverter of choice is SMA, for example Sunny Boy 1100, which provides electrical separation utilising 'Galvanic Isolation'. My point is that although the manufacturer 'recommends' the use of a RCD, there is no requirement under BS 7671 712.411.3.2.1.2 to provide one under these circumstances.

Can somebody please confirm my thinking as I'm going to nip this in the bud before the honest contractor gets his knickers in a twist.

- Wango.

-------------------------
'They are, in fact, men apart. Every man an emperor'

Field Marshall The Viscount Montgomery

It looks as though the Electrician is concerned with 522.6.7.

Mark

Ah fair point, but not relevent in these cases as cables are either clipped direct or run in trunking/conduit.

Thanks.

- Wango.

-------------------------
'They are, in fact, men apart. Every man an emperor'

Field Marshall The Viscount Montgomery

Don't want any knicker twisting going on, especially so early on in the week .

Have you read the following before commencing de-knicker-twisting weekly:

Capacitive Discharge Currents - Information on the system layout for transformerless inverters

Especially with the following:

The determination of the residual current is only possible to a certain limit and becomes more difficult with increasing discharge currents. Starting with approx. 50 mA, random fluctuations in the discharge current are so large that they can be interpreted as suddenly occurring residual currents of above 30 mA. In such a case, the inverter disconnects automatically from the grid as a preventative measure.

and

As previously described, discharge currents of above 50 mA should be avoided in order to ensure the functionality of the residual current monitoring.

Now that covers any residuals between the PV array and the inverter.

Now if you believe that at least "simple separation" exists (which the sunny boy range do) between the a.c. side and the d.c. side and all cable runs are such that you do not invoke the wrath of the RCD regs, then I guess you can say no to 30mA RCD if you want.


Does s/he have a justifiable reason for this insistence? It won't do any harm, but that is not the same as saying it is an absolute must in all circumstances.

Just as a wee side note - your 16 A mcb might be a wee tad excessive for the SB 1100 - or maybe its wishful thinking

Regards

-------------------------
E & OE

how about 134.1.1 and or 510.2

I suppose it's a situation of not wanting to re-open the case on completed installs... But for future work there seems little point in arguing the point - an RCD is a trivial cost compared to a pv install, and arguing that the inverter manufacturers only recommend rather than require doesn't seem to be worth the hassle - or am I missing something?

Of course the RCD should be at the consumer unit end of the cable run... any danger from say nailing into the wire comes from the mains side, not the inverter... The inverter shuts down as soon as the mains frequency or voltage drops out of spec...

Dee

The SMA link above appears to be concerned with the inverter disconnecting unnecessarily, and is aimed at system designers.

712.411.3.2.1.2 concerns inverters which may be able to feed d.c. fault current into the a.c. system, hence the type of RCD specified. Wiring Matters 2009_31 summer mentions this.

Mark

The RCD mentioned in 712.411 seems to be at the inverter end according to Fig 712.1, although for 522.6.7, I agree, you'd clearly want (another?) one on the CU end.

Anyone got any idea how long the inverter would normally take to disconnect once the mains supply disappears ... and what effect that might have on ADS disconnection times? (never mind RCD testing )

- Andy.

Thanks for the replies.

This enquiry relates to past work - time and expense.

I don't think the issue is straight forward though. As Andy alludes to, there are in some cases the mention of 2 RCDs. I may be wrong here, but a RCD to BS EN 61008/9 would be of AC type so would only be sensative to AC faults. Anything coming from the DC side wouldn't register so a A or B type RCD would have to be used in that case. So the RCD within the CCU must be to protect the inverter and DC side from AC derived faults coming from within the property (although SMA inverters have built in protection).

- Wango.

-------------------------
'They are, in fact, men apart. Every man an emperor'

Field Marshall The Viscount Montgomery

Nowt wrong with a 16 A CB. We utilise 16 A's on all our G83 installs. 1.5mm2 or 2.5mm2 cables from inverter to total generation meter to CCU. Anyway, we aren't interested in overload protection in this case, only fault protection.

- Wango

-------------------------
'They are, in fact, men apart. Every man an emperor'

Field Marshall The Viscount Montgomery