RCD's with Photovoltaic's?

[sharpener]

In this case the RCD is purely there as a DP switch as we haven't been able to get a non RCD board.

Screwfix have a Crabtree main switch unit stock no. 77250 which is what I used, it gives a spare non-RCD way for the freezer into the bargain.

 

[SolarCity]

Thank you for your view on this matter, i will be contacting megger to see if they can shed some light on how my megger works!

Let us know what they say. Should be interesting.

 

[dougalthedog]

Hi,

I've done some testing too and have always found disconnection times comply with regs. However, you need to consider how much the inverter is generating vs. how much the other circuits on the RCD side are drawing. If you are drawing far more than you are generating, you are effectively short circuiting the inverter so you would expect it to shut down pretty quickly. On the other hand if you are generating far more than you are drawing the inverter's mains loss circuitry will probably react very quickly to turn off. It would be interesting to see what happens when the two are closely matched.

The point is that the inverter spec. allows 5 seconds for disconnect in a loss of mains situation. It may well be that in most scenarios it turns off far faster - quickly enough to pass an RCD test but this is no guarantee that it will do so in every situation. The only way to guarantee safety is to have the PV system protected seperately.

Can anyone see a problem with fitting a 30mA RCBO on the PV circuit (providing inverter can work with 30mA RCD) - thus you would have the main RCD protecting circuits from the mains supply and RCBO protecting other circuits from the PV system. I can't think of any reason why this wouldn't work.

 

[SolarCity]

Can anyone see a problem with fitting a 30mA RCBO on the PV circuit (providing inverter can work with 30mA RCD) - thus you would have the main RCD protecting circuits from the mains supply and RCBO protecting other circuits from the PV system. I can't think of any reason why this wouldn't work.

The problem with this is that if the RCD protecting the non-PV circuits was to operate then you are still left with the potential for the inverter to give a longer disconnection time.

I have a couple of installs that I have carried out personally that I will be looking to alter having considered the problems with RCDs and disconnection times.

 

[dougalthedog]

Yes, but the RCBO on the PV circuit would / should also trip surely as there would still be an imbalance (earth leakage). However I'm not sure whether the trip time would be twice as long as there are 2 RCDs or if it would work as normal.

 

[SolarCity]

The PV RCD would only trip if the fault was on the PV side, would it not?

 

[moggy1968]

there is actually no guarantee the RCD will trip at all, so it's a false sense of safety.
the only way to guarantee correct tripping as I see it is a type B RCD

 

[Markc]

Mr NICEIC (on ANNUAL inspection) is who brought this to my attention as I have said. My test equipment has shown a <40mS time on RCD tests of inverter's in generation (Metrel test multifunction test kit) Metrel make Meggar's as well so I would see them both working the same in theory.

G83 only requires a disconnection time of 5 seconds. Yes some may drop touch voltage in a shorter time scale but there is no guarantee.

PV on RCBO is not in question as we are looking at communal RCD protection offering supplementary protection to further circuits requiring it. RCBO on a RCD protected way is pointless as the supply is still on a communal RCD so the problem still exists.

The question for any 30mA RCD device is also more a concern where cables are embedded in a wall at <50mm with no earthed mechanical protection. BS7671 requires a disconnection time of <40mS and this isn't going to be guaranteed with any 30mA RCD/RCBO for what I can see. Yes it might trip but the voltage from the inverter may still be present over save touch voltage for >40mS.

As for the training centers, MCS inspectors and certifying bodies. I think this proves that the PV industry is still very new. Not all the in's and out's have been fully understood and things such as this will crop up. I for one have no problem admitting that until last October I was not to unhappy about using any spare way available whether RCD protected or not but now I certainly will not fit to a communal RCD and will only use a none RCD way. I still have not come up with reasonable way of offering supplementary protection where required.

 

[VMan]

Understandable but i'm on a TT system so need the RCD.

Think im going to change to all RCBO's soon.

 

[moggy1968]

you don't need a 30ma RCD for a TT system. the only requirement is the circuit disconnects in 0.4secs. The reason for an RCD is because of the high ZE a MCB wouldn't disconnect in 0.4secs. As such you could use a 100ma type B RCD on it's own sub board.
As an example. when I rewired my folks house which is a TT system I fitted the supply to the garage (fed by SWA) with a 100ma time delayed RCD. As the supply is in the loft (overhead cable) I wanted to introduce some discrimination to the tripping when my dad tripped the garage supply (as he almost certainly will!) so he didn't have to go back up into the loft every time it went. This would also mean the freezers in the garage wouldn't trip if one of the other circuits tripped (all the garage circuits were on 30ma RCBOs). If I had fed the garage with a 30ma RCD both the ones in the garage and the one upstairs would probably trip.
That assumes of course an RCD isn't required for the other reasons detailed in 1771

 

[VMan]

granted, but its already got a 30mA RCD as main switch, plus i didn't wire the house and im sure the other circuits are not buried more then 50mm or in safe zones.
I was under the impression a type B RCD was only required if there wasn't electrical separation i.e a TL inverter?

 

[moggy1968]

you only need an RCD if there isn't electrical seperation, although even that is debateable as some inverters have inbuilt earth fault protection which satisfies that requirement (so I believe!)
Unfortunately for you the reason you need an RCD is because your on a TT, and if you need an RCD for PV it has to be a type B.
main switch RCDs are a pain because if they trip everything goes off. RCBOs is the gold standard option but it can be tight getting them in a board on every way. twin RCDs is the more common solution. A high integrity board gives you the best compromise IMHO as it gives you twin RCDs plus a couple of spare ways for circuits you may want on ther own RCBOs (freezers for example)

 

[dougalthedog]

Think we're mixing up 2 different things here. This thread started off discussing the possibilty of compromising the RCD functionality on an RCD protected section of a split load board. The issue of potential DC leakage from an inverter is another topic in itself but not what was initially raised.

The point is if you have an RCD protecting multiple circuits (such as a bathroom or a socket circuit buried < 50mm) and you connect a PV installation to one of the MCBs on this board, when the RCD trips, the PV cicuit is still linked (via the busbar in the CU) to the other circuits including the one which the end user just hammered a nail into, for example!

If the inverter takes the 5 seconds which it's allowed to to shut down then your customer may very well be fried, whereas he would be entitled to expect only a brief shock of the few mS required for the RCD to trip.

However, if the inverter was fed by an RCBO instead of an MCB then I would expect this to trip as well as the main CU RCD thus cutting the inverter off from the rest of the circuits (there would still be earth leakage which it would detect) although it is possible this would take twice as long as usual to trip (2 RCDs needing to detect >30mA leakage not exactly easy to work out the current flows)

The best solution as has been said before is to connect your PV system to a separate garage CU (with an RCD if required due to cable mounting method or without if specified by inverter manufacturer!). This also allows for easier measurent of power being generated and in use should the customer require it.

 

[BruceB]

You have it a bit the wrong way around Moggy.
712.411.3.2.1.1 is the relevant reg

That reg does not require an rcd.
But it says if there is an rcd for ADS (eg TT)
and if no simple separation (ie transformerless)
then use a type b rcd
unless the inverter meets the exception at the bottom.

There is a longer debate around this subject that you will find on here, but that is what the reg says.

 

[dougalthedog]

Re. Biggs' point that the RCBO would only trip if the fault was on the inverter side of the RCBO, I'm not entirely sure. I can't see any reason why an RCD/BO should be directional and once the main RCD trips, there will still be an imbalance between the line / neutral currents so it should trip. I probably wouldn't want to bet my life on it so will stick to installing a separate CU.