PV systems

[JohnD]

If all of current 'supplying' the fault (which, don't forget, only needs to be a '30mA fault' to operate an RCD) was coming from the PV, then there would not be any reason for the 'mains side RCD' to operate, would there?

You mean, if the load plus the fault current, and the current available from the PV, were so precisely balanced that there was nothing flowing either from or to the network?

 

[ban-all-sheds]

I got 99 changes a second.

Ooh no - they change direction far more often than that.

If you're talking about their average drift speed at a macro level, in 1/100th of a second they will have moved a few tens of μm, so won't have got very far down a length of wire, or through an RCD.

 

[JohnW2]

If all of current 'supplying' the fault (which, don't forget, only needs to be a '30mA fault' to operate an RCD) was coming from the PV, then there would not be any reason for the 'mains side RCD' to operate, would there?

You mean, if the load plus the fault current, and the current available from the PV, were so precisely balanced that there was nothing flowing either from or to the network?

I need to think about that question (it's quite easy to get one's thinking tied in knots with this issue!) - watch this space

Kind Regards, John

 

[ericmark]

You have raised so interesting points. MCB may not mind which way around but RCD does.

if an RCD is comparing current flowing in two wires, how does it know which way the electrons are travelling? Especially when they reverse direction 50 times a second?

I got 99 changes a second.

It may be true that some RCD's can be connected either way around but many have electronics built into them and these will be either taking power from oncoming or outgoing terminals. With a passive it may not matter but it certainly would with an active device. I would guess even with a passive the electronics should be powered up at all times or disconnection times when resetting under fault conditions may be exceeded.

Even with an MCB having incoming power on the left (bottom) will mean when it operates all the coils and heaters become dead. Power it from the right (top) and there are more live components within the unit when open.

With the RCD even if no electronics the test button will fail to work if wired the wrong way around.

When there is power both sides then clearly the button will work but as the contacts start to open then power will be lost to button and it could chatter slightly.

There are capacitors shown in that circuit and there must be a charge time. I would think there will be a recommended type we has AC, A, and B which is required would depend on if any DC is superimposed on the AC under some conditions and I don't know the answer on which one should be used. It does state incoming 1 and outgoing 2.

 

[JohnW2]

If all of current 'supplying' the fault (which, don't forget, only needs to be a '30mA fault' to operate an RCD) was coming from the PV,...

Er could you explain the circumstances when that could happen when the PV output and the mains are in parallel?

As I said, I'm working from a position of relative ignorance about how these things work - which doesn't help! Are you saying that some of the current flowing through the installation's loads (including any small faults) will be grid-derived, even when the output of the PV is more than adequate to service those local loads?

Kind Regards, John

 

[JohnW2]

It may be true that some RCD's can be connected either way around but many have electronics built into them and these will be either taking power from oncoming or outgoing terminals.

True, but that obviously doesn't matter unless the device is in an 'operated' state, since there is negligible impedance between the incoming and outgoing terminals.

With a passive it may not matter but it certainly would with an active device. I would guess even with a passive the electronics should be powered up at all times or disconnection times when resetting under fault conditions may be exceeded.

I'd never really thought about it but, yes, a situation in which one is attempting to reset an RCD whilst a fault is still present is the one situation in which it might make a (small) difference 'which way round' one wires it.

Kind Regards, John

 

[westie101]

Are you saying that some of the current flowing through the installation's loads (including any small faults) will be grid-derived, even when the output of the PV is more than adequate to service those local loads?

That is my contention, yes owing to the higher availability of the grid to supply current whatever the circumstances.

However I am now wondering where the idea that the neutral & earth are not connected together in the inverter?
Each and every generator that I have come across (either owned by us, hired by us or privately owned and operated) that can operate in parallel with the mains sees either one end (single phase) or the star point (three phase) or the winding connected to earth, so why should an electronic device be different?
The cases are metal so an earth is connected to the case so is available.

As with others I've never pulled one apart so cannot say conclusively

 

[mikeyd]

.....if the load(s) on the circuit which developed a fault were being 'totally supplied' by the PV if that is possible), then there is no obvious reason (well, not one obvious to me!) why the installation's existing RCD should operate. In that scenario, only a separate RCD in the feed from the PV could react to the fault

The way I understand it, a sperate RCD in the feed from the PV would not see such a fault. The current through the PV's line and neutral would remain equal. The existing installation's RCD will still pick up the fault though.
Got to go but I can try to explain better later.

 

[JohnW2]

Are you saying that some of the current flowing through the installation's loads (including any small faults) will be grid-derived, even when the output of the PV is more than adequate to service those local loads?

That is my contention, yes owing to the higher availability of the grid to supply current whatever the circumstances.

Fair enough. As I said, I don't really understand these things at all, but what you say makes sense. So you are presumably saying that, in the circumstances I described, the current flowing through local loads will have contributions from both PV and grid, even though the 'excess' capacity of the PV is feeding into the grid - is that correct?

Mind you, even if that is the case, it presumably means that the fault current will be shared between the PV and grid supplies - so that, at the least, a 30mA L-E fault current on a final circuit would not (necessarily - thresholds vary!) result in a 30mA 'mains-side' RCD operating.

However I am now wondering where the idea that the neutral & earth are not connected together in the inverter?

It was not an idea that came from me. I observed that an RCD in the feed to/from the PV inverter could never operate in response to a fault on the downstream side of the RCD (i.e. opposite side to inverter) unless the neutral of the PV inverter was connected to earth (since there would otherwise be no alternative path to the inverter N which circumnavigated the RCD), and therefore asked if that was what was done. IIRC, at least one person said 'no'.

Kind Regards, John

 

[JohnW2]

The way I understand it, a sperate RCD in the feed from the PV would not see such a fault. The current through the PV's line and neutral would remain equal.

Indeed, that's the point I'm been making all along.

The existing installation's RCD will still pick up the fault though. ... Got to go but I can try to explain better later.

That's the point westie has been making, and around which I'm currently trying to get my head! Any help you can offer would be welcome!

Mind you, as I've just written to westie, if the load (and hence fault) current is shared between the PV and grid supplies, then a 30mA fault on a final circuit would probably not usually result in the "existing installation's RCD" (assuming 30mA) operating - an appreciably larger total fault current than that would usually be necessary for there to be a 30mA L-N imbalance in that RCD to cause it to operate.

Kind Regards, John

 

[westie101]

Here's some nice reading

https://www.ofgem.gov.uk/ofgem-publ...09-07-12-inc-ofgem-comments-clean-version.pdf


Top of page 20

5.7.2 Inverter Connected Generation Inverter connected SSEGs generally have small short circuit fault contributions.
However DNOs need to understand the contribution that they do make to system fault levels in order to determine that they can continue to safely operate without exceeding design fault levels for switchgear and other circuit components.
As the output from an Inverter reduces to zero when a short circuit is applied to its terminals, a short circuit test does not represent the worst case scenario; in most cases the voltage will not collapse to zero for a network fault

 

[mikeyd]

.. if the load (and hence fault) current is shared between the PV and grid supplies, then a 30mA fault on a final circuit would probably not usually result in the "existing installation's RCD" (assuming 30mA) operating

I'm having a hard time trying to think how to explain it, but if we start with the above quote, why do you think that? If there is a 30mA fault then the RCD will see a 30mA difference between line and neutral. It makes no difference if current is being imported or exported

As an example, suppose the load needed by the house was exactly the same as the PV was supplying. So no current flowing through the rcd either way. Then lets suppose there is a 30mA fault between Neutral & earth, does it not make sense that there will suddenly be a 30mA difference in the neutral current at the RCD?

 

[plugwash]

The existing installation's RCD will still pick up the fault though. ... Got to go but I can try to explain better later.

That's the point westie has been making, and around which I'm currently trying to get my head! Any help you can offer would be welcome!

The key thing to remember is that current flows in loops. So when a "fault to earth" happens current flows from the supply, "LINE" through the fault, through the "earth"*, through the connection between neutral and earth and back to the neutral of the supply.

If the connection between neutral and earth is on the upstream side of the RCD and the fault is on the downstream side of the RCD then the fault will create a corresponding imbalance in the RCD regardless of which side of the RCD the supply is on.

*"earth" in this context may include the earthing system, the general mass of the earth and anything else that is intentionally or unintentionally connected to them.

 

[DetlefSchmitz]

My current theory is that if the cable (inverter to house supply) needs an RCD, then it needs one at both ends. Supplying to the house CU via a way which is NOT RCD protected, deals with the house end.

The question then, is whether the generator (inverter) end RCD needs the neutral to be earthed on the inverter side. I am leaning towards the view that it does not. As it is at earth potential anyway, it would still be the case that live to earth current in the 'middle' would cause an imbalance.

 

[JohnW2]

.. if the load (and hence fault) current is shared between the PV and grid supplies, then a 30mA fault on a final circuit would probably not usually result in the "existing installation's RCD" (assuming 30mA) operating

I'm having a hard time trying to think how to explain it, but if we start with the above quote, why do you think that? If there is a 30mA fault then the RCD will see a 30mA difference between line and neutral. It makes no difference if current is being imported or exported

Let me try an example, then. Say there is a 10A load and, per westie's proposition, say that 5A of that is coming from the PV (connected to final circuit downstream of the house's RCD, hence 'avoiding' that RCD) and 5A from the grid, via the house's RCD. There will be 5A in both L and N of that RCD, hence balanced. Now add a 30mA L to E fault on the loaded circuit. The total 'load' increases to 10.030A, hence 5.015A from each source. 5.015A therefore goes through the RCD's L, but only the 5.000A due to the 'proper load' goes through the RCD's N. The L-N imbalance in that RCD is therefore only 15mA, although the fault itself is 30mA. With those figures, one would need a 60mA fault on the circuit to achieve a 30mA imbalance in the house's RCD. ... or am I still missing things or getting things wrong?

Kind Regards, John