Parasitic Capacitance.

allinthehead

On some PV installations there can be intermittent tripping of the rcd/rcbo when it rains or when condensation levels are high. This occurrence is usually only with single phase, transformerless inverters.

The pv panels themselves become conductive in the moisture and half of the grid amplitude can arrive at the panels.

You have the panel mounted on a grounded frame and when the voltage arrives on the conductive panel it can store charge like a capacitor. The alternating voltage changes the state of charge of the capacitance and this gives rise to a displacement current proportional to the capacitance and the applied voltage.

This "leakage" current is seen as an imbalance at the inverter and beyond at the rcbo causing the nuisance tripping.

This is my understanding of the phenomenon after reading up a little on it. I believe the fix is to add a 150 nf compensation capacitor to the circuit via a 4 pole rcd/rcbo.

Just looking for a sanity check on this before proposing the fix to someone having the issue.

5d2846461b8bd.pdf

graememk

https://www.boards.ie/discussion/2058326679/parasitic-capacitance

The capacitor is this, is proving some "back" leakage current.

I wonder is it blinding the RCD a bit, making a 50ma RCD into a 75ma RCD? (As it's providing a back current )

allinthehead

https://www.boards.ie/discussion/comment/121489005#Comment_121489005

I'm not sure I understand.

My take on it is that the leakage current is added to any residual current, meaning the 30ma rcd trips.

If it was blinded to this current then the nuisance tripping wouldn't happen.

10-10-20

https://www.boards.ie/discussion/comment/121490277#Comment_121490277

He means blinded by way of having a bypass current over the RCD so that the overall trip current is therefore greater as some bypasses the measuring coils within the RCD.

Anyway the answer in my opinion is yes it can as it's only offsetting for a parasitic current anyway, so the difference is negligible when the compensation current calculated correctly.

The relationship given is:

Differential current = leakage current + residual current - compensation current

• Differential current = the trip current observed by the RCD and identified to be 15mA to 30mA for a 30mA A-type RCD.
• Leakage current = the cumulative parasitic capacitance of the panels overall.
• Residual current = the circuit/inverter residual current loss.
• Compensation current = The calculated offset current to offset the leakage current.

If the compensation current is oversized then it's my understanding that it's hiding some of residual current and potential fault-current, so the RCD is less sensitive to a true fault condition.

So 150nF would be suitable for offsetting 11mA if the cumulative parasitic capacitance of the panels is 11mA but not if it's less.

That's my take on it.

10-10-20

Oh yeah, I just realised that the other factor here is that the parasitic capacitance is being observed and calculated for during times of rain, so outside of these times there is little or no cumulative parasitic capacitance of the panels and therefore the sensitivity of the RCD will be fully derated by the compensation current. So yes, again pick the bypass capacitor on the basis of the calculation, and make sure it's a class X2 capacitor (across the line capacitor) so that the failure of the capacitor causes the RCD to trip and not a class Y where the capacitor will go open-circuit.

https://www.allaboutcircuits.com/technical-articles/safety-capacitor-class-x-and-class-y-capacitors/

johndoe11

Not an electrician and not sure if this is related, but I had my installer out a few months ago to look at something like this. I was replacing an end cap on one of my roof mounted rails when i got a small tingle of it, it was a very wet day. They measured about 90V ac when checked with a multi meter between the ladder and the rail. The installer has no clue where the ac current was coming from. After a few visits, they eventually grounded the mounting frame and I have no issues since. I didn't get any inverter trips or faults before, only really noticed as I was replacing the end cap.

allinthehead

https://www.boards.ie/discussion/comment/121490383#Comment_121490383

Yes I understand what he meant now. Thanks for the breakdown.

https://www.boards.ie/discussion/comment/121490438#Comment_121490438

That's a crucial point you make, a very real danger of rendering the rcd ineffective in the event of a real fault.

Thanks for the advice on the capacitors too.

allinthehead

https://www.boards.ie/discussion/comment/121490530#Comment_121490530

May well have been the same issue but not enough current to trip the rcbo.

10-10-20

https://www.boards.ie/discussion/comment/121490530#Comment_121490530

That's part of the parasitic current. It's the C2 part in figure 1 on that PDF attached.

Anyway the frame is supposed to be grounded so the potential is removed (the 'potential' is the voltage difference between the metal frame and the earth) and not having it earthed would mean that it was "floating" and that could possibly cause "other issues" on the solar inverter.

In the past I heard installers say: "sure the mountings don't need to be earthed as they won't be touched". Well they do.

johndoe11

Thanks, just read that now and I recalled that they were also getting a voltage between the wet slate roof and earth!

graememk

https://www.boards.ie/discussion/comment/121491662#Comment_121491662

Then what would the difference be between doing that and going with a higher trip current RCD? Although that would be 100ma?

allinthehead

https://www.boards.ie/discussion/comment/121492151#Comment_121492151

That would work too, problem is you can't use a 100ma rcd, regs won't allow it.

graememk

https://www.boards.ie/discussion/comment/121492261#Comment_121492261

Totally defer to you on that then.

then technically using a capacitor would be circumventing the RCD a bit too.. well into the grey area there!

Would love if you could actually test the trip time/current on a RCD with a capacitor Vs no capacitor.

allinthehead

https://www.boards.ie/discussion/comment/121492311#Comment_121492311

Agreed, testing is the way to go here. So long as it trips when 30ma is applied there shouldn't be an issue. Safe Electric have confirmed that 100ma can't be used, no solution forthcoming on their part.

Mr Q

I had a similar issue, fitted a higher mA unit and hasn't tripped since then.

I never checked what the safe electric view was, but it mentioned in the inverter manual that a 100mA or higher unit may be needed.

TerraSolis

https://www.boards.ie/discussion/comment/121491882#Comment_121491882

Whilst I absolutely agree RE Earthing, the funny thing is IS:10101 does not require earthing of metallic array components unless they're in contact with metal components that are connected to the building earth (e.g. metal roof and drain pipes, structural steel etc.). Even many metal shed roofs get away with it according to IS:10101 as many aren't earthed...

10-10-20

https://www.boards.ie/discussion/comment/121513166#Comment_121513166

As far as I know panels which are attached to a class-II inverter (galvanically isolated transformer, typically) aren't extraneous parts, so I don't think IS:10101 applies to the DC side at this time.

S'

I have noticed this intermittent tripping of the RCCB when it starts raining since I swapped out a Solis inverter for a Deye/Sunsynk. Would the type of inverter be causing this?

BTW, I have a Hager 63A 30mA RCCB

allinthehead

https://www.boards.ie/discussion/comment/121549646#Comment_121549646

The Sunsynk is a transformerless inverter, which are affected more. Is the solis?

whizbang

The issue is very similar to non-earthed 2pin mains connector Switchmode power supplies. With very close coupling of the high frequency tranformers, the output would tend to float up to 300v dc value of the mains input.

The 2 pin reversible connection means the manufacturer cannot designate either of the pins as Neutral and then so discharge the output leakage to neutral. So what they do is create a rough midpoint voltage with capacitors between L and N, which usually measures about 90-110v, and link the output to that. So its a lot better than 300v.

TV,Video equipment is notorious for this, and lots of othe small mains powered devices, usually ITE. Ever notice all laptops now have a 3pin plug..

Theres nothing to stop users from adding a bigger capacitor to neutral, and thats effectivly what the extra capacitor does for your solar system. It does bring in some other issues tho..!

10-10-20

https://www.boards.ie/discussion/comment/121981546#Comment_121981546

Interesting detail.

Maybe I have this wrong but if a solar inverter has designated L and N, why then do the manufacturers not have sufficient capacitance at N then to overcome the issue anyway? Or does the additional capacitance just become a nuisance load during normal operation?

And is it N to E which you're suggesting, so class Y rated? How would that offset the RCD though if it's not bridging the RCD terminals?