Solar PV insulation fault ABB aurora inverter.

[broombox]

Hi guys, I am new to the forum so be nice!

I used to install solar PV systems back when the tariffs were still good, and I've never had to go back to one of my old installs because of a fault, so I'm not too clued up with regards to fault finding on PV.

A customer has recently called me out to look at a PV system not working (installed by a firm that's now bust).

Everything looks all present and correct, but i cannot gain access to the panels without scaffolding. No visible obstructions from standing at ground level looking up at the array, and I can see up underneath where the wires are plugging into one another. I've seen worse installs, DC cables could have a few more clips in them but generally everything looks OK.

Now, the inverter (ABB aurora power one) looked like it had frozen when i got to it to have a look. switched off and rebooted, now coming up with a RISO fault, 0.00 M ohms, so that's obviously why the PV inverter isn't starting up. Tested the voltage on the strings + to - at 480 volts. This is for 14 panels on one string, so i think that seems about right. Now at this point, I do not have access to the inverter manual, it's a sunday so cannot phone anyone for technical advice, and it's about 30 degrees outside so i don't think i can megger the wiring with the voltage present without frying my meter.

I've now had chance to do a bit of research, obtained the inverter manual online, and it's recommending in the event of a ground fault, to test with a multimeter from + to earth and - to earth, and if there is a stable voltage, then there is an earth fault. The thing that's confusing me here though, is that there is no earth to the PV array - the panels are floating, and the rail etc is not bonded. The only earth to test against is the inverter earth, which connects back to the AC supply. So I don't see what testing to this earth, with the dc disconnected, is actually proving? As it is separate from the array on the roof. How can the inverter detect an insulation fault to earth, with no earth connected to the panels? surely what it is detecting is a + to - short circuit, and if so what is the correct way to test to prove this?

I contacted ABB technical, and they told me that a RISO fault as low as 0.00 M ohms is usually a faulty inverter. But that totally contradicts the advice in the inverter manual, and I need to prove 100% if the fault is on the array or not before I instruct the customer to have scaffolding put up.

Any help would be much appreciated!!

 

[marconi]

Do you have an any LED lamps lit? Is GFI lit?

Try a controlled power down and then power up.

Has it rained recently over the array and any of the dc wiring?

Are the panels 'amorphous' types? Is the 'amorphous' mode selected?

Are any over-voltage surge protectors fitted to the dc array wiring?

Take a good look at all the dc connections at the array.

Any signs of moisture inside any of the panels?

With dc array disconnected from inverter input do this check page 42 of manual:

https://library.e.abb.com/public/e57212c407344a16b4644cee73492b39/PVI-3.0_3.6_4.2-TL-OUTD-Product manual EN-RevB(M000016BG).pdf

Checking of leakage to ground of the photovoltaic generator Measure the voltage present between positive and negative pole of each string with respect to ground. If a voltage is measured between an input pole and ground, it may be that there is a low insulation resistance of the photovoltaic generator and the installer will have to carry out a check to solve the problem. The ground in question is the cpc to the inverter. Check the cpc is indeed connected to the MET.

[automerge]1597302192[/automerge]

[automerge]1597302298[/automerge]
Sorry - study the 3rd and 4th images not the 1st and 2nd which I am unable delete because of the daft 20 minutes rule.

 

[whinmoor]

It would have been much better at the outset to wire the array up with 2 strings. Then you can try running the inverter with just one string and swap to the other. If it works with String A, but fails with String B, you can be pretty confident there's a fault with String B.

You could try plugging the string into mppt 2 but I doubt that will change the outcome.

The system may be underperforming in bright conditions as there may be a power limit on the mppt.

If it's the 3.0 version, it'll be capping at 2 kW.
If it's the 3.6 version, it'll be capping at 3 kW.

You could try taking the inverter to another property and testing it on a system which is working well. Equally, you could take a good (or new) inverter to that house and see whether that works or not.

If there's a problem on the roof, it's most likely a connector. Probably one on an extension cable made up by the installer.

Also, pay attention to whether the problem occurs in wet conditions. The system may work when it's dry but fail when wet.

 

[broombox]

Do you have an any LED lamps lit? Is GFI lit?

Try a controlled power down and then power up.

Has it rained recently over the array and any of the dc wiring?

Are the panels 'amorphous' types? Is the 'amorphous' mode selected?

Are any over-voltage surge protectors fitted to the dc array wiring?

Take a good look at all the dc connections at the array.

Any signs of moisture inside any of the panels?

With dc array disconnected from inverter input do this check page 42 of manual:

https://library.e.abb.com/public/e57212c407344a16b4644cee73492b39/PVI-3.0_3.6_4.2-TL-OUTD-Product manual EN-RevB(M000016BG).pdf

Checking of leakage to ground of the photovoltaic generator Measure the voltage present between positive and negative pole of each string with respect to ground. If a voltage is measured between an input pole and ground, it may be that there is a low insulation resistance of the photovoltaic generator and the installer will have to carry out a check to solve the problem. The ground in question is the cpc to the inverter. Check the cpc is indeed connected to the MET.View attachment 60075View attachment 60076
[automerge]1597302192[/automerge]
View attachment 60077View attachment 60078
[automerge]1597302298[/automerge]
Sorry - study the 3rd and 4th images not the 1st and 2nd which I am unable delete because of the daft 20 minutes rule.

Hi, thanks for your reply.

I have tried a controlled power off/on, several times, and this does not clear the fault. The red GFI lights up and the message RISO 0.0 M ohms is displayed on the screen, after the unit first measures the voltage on the AC side. After this is displayed, it will not do anything else until you reset again.

The array is up on a roof, and i cannot gain access to check it or any of the DC connections without scaffolding. This is why I am trying to ascertain, with no doubt, that the fault is definitely on the array, and not a problem with the inverter, before I instruct the customer to go ahead and get a company in to put up scaffolding in the street. The panels do not look damaged (from street level) and I can't see any obstructions such as birds nest or dirt that may cause moisture to gather. Also, the customer has reported the system has not been working for several months, and when I have attended the call out, it has been the hottest week of the year - temperatures in the 30's and no rain for over a week. fault is still present.

I am not sure what you mean by "amorphous type." Would this affect an insulation fault? bear in mind the system had been working for 6 years prior to this so I'm not sure this would cause it to suddenly fail?

I don't think there is any surge device fitted, other than what may be built into the inverter. Are you thinking that a surge, possibly a lightning strike could have caused a surge which has in turn damaged the inverter?

I have carried out the testing mentioned in the manual. I isolated the dc from the solar array and tested from + to the earth at the inverter, and then the same with the negative. Both showed 0 volts (tested in day time with a voltage > 400v positive to negative). Whilst this would suggest there is no ground fault, I cannot see how this would show a voltage in any case, as there is no earth whatsoever bonding the array on the roof back to the inverter, so testing from the DC connections to the earth on the AC side surely would not show a fault, even if there was one, because the earth from the AC side is not common with the exposed conductive parts of the array. This is why I am struggling to see how the inverter is detecting an insulation fault, unless the fault is a short between positive and negative, and my question is, is there any way I can actually test this, without gaining access to the array itself?
[automerge]1597499592[/automerge]

It would have been much better at the outset to wire the array up with 2 strings. Then you can try running the inverter with just one string and swap to the other. If it works with String A, but fails with String B, you can be pretty confident there's a fault with String B.

You could try plugging the string into mppt 2 but I doubt that will change the outcome.

The system may be underperforming in bright conditions as there may be a power limit on the mppt.

If it's the 3.0 version, it'll be capping at 2 kW.
If it's the 3.6 version, it'll be capping at 3 kW.

You could try taking the inverter to another property and testing it on a system which is working well. Equally, you could take a good (or new) inverter to that house and see whether that works or not.

If there's a problem on the roof, it's most likely a connector. Probably one on an extension cable made up by the installer.

Also, pay attention to whether the problem occurs in wet conditions. The system may work when it's dry but fail when wet.

Hi, thanks for your response.

I share your thoughts on the system having been installed on only one string. I did not install the system, I have just been called upon to try and get it working again. All of the systems I installed myself were done with 2 strings for exactly the reasons you have mentioned. It is interesting to consider that the inverter might be capping the power. This is a 3.5KW system, so if the power is being capped at 3KW then it might be worth getting access to the roof to change it, even if the fault does lie with the inverter, if it means the customer will then be getting back more return from the panels. I will contact the manufacturer again to check this. I have tried swapping the string inputs on the inverter to string 2, but as you've said it makes no difference.

The fault is present in dry conditions. Temperatures have been in the 30's all week with no rain. This has made me wonder if the detected short is actually within the inverter itself, although there are no visible signs from what I can see. I guess if water has got into a panel in the past though, creating a short, the damage might already be done.

I think getting a replacement inverter to try might be the only solution, provided I can return it free of charge. I just really want to prove what is causing this before wasting time/money going with the wrong solution.

 

[marconi]

My diagrams were an attempt to explain how voltage readings (V1 or V2) can be produced which are other than zero if there is an insulation fault on a panel(s) and their wiring to surrounding 'stuff' which is in turn 'connected' to earth.

I suppose - don't know for certain - that the way the ABB measures IR is to measure V1 and the current I1 which flows, and then V2 and the current that flows I2, and then its computer calculates V1/I1 and V2/I2 to estimate the leakage resistance, and it displays the lower value of the two.

Re Amorphous - look at error code EO37 Red LED in:

ABB Inverter Fault Codes - Empower Energy - https://empowerenergy.co.uk/domestic-solar/abb-inverter-fault-codes/

I just wondered if a grounding kit was fitted to the inverter, the panels were amorphous types and the amorphous mode had been selected.

Have you tried disconnecting the ac mains to the inverter and then powering up? See the link above and error code EO18 Red Led. Ac leakage currents through the array can be troublesome too because of capacitive leakage currents.

I suppose you have thought of it already but in case you have not try using only one array at a time and also swapping them over and doing the same trial again.

Another test I would do is to buy two 10MOhm resistors (or high value MOhm) which can operate at the array voltage and then wire them in series with the positive and the negative feeds from the array into the inverter - see my diagram. The array is then connected to the inverter but when the RISO test is done it is the two 10MOhms resistors which it 'sees' - which don't indicate an earth leakage fault - so the computer may not generate the RISO fault code and light the GFI LED red.

Resistors | Rapid Online - https://www.rapidonline.com/Catalogue/Search?Tier=Resistors&Size=20&Attributes={%22Resistance%22:[%2210M%26amp%3BOmega%3B%22],%22Working%20Voltage%22:[%22350V%22]}

This might help isolate the fault to the array if during the trial the inverter powers up as if it was night time.

 

[anotherbrownbear]

Good day

I have the same fault coming up on an inverter but different circumstance. Customer for some reason enabled Amorphous mode while not having a grounding kit. Now I am unable to get in to inverter to change setting back as it keeps coming up with Riso Low error. Does anyone have a solution for this?