Error 35 on SMA

[DanPhillips]

Hello

We installed a PV system approx. 18 months ago which has recently started showing an event number 35 in wet weather conditions which according to the manual is a ground fault.

The system is a floating class 2 array so presumably we cannot have a ground fault.

We have tested the system during dry operating conditions and again whilst the event is displayed and the results were as follows


DRY conditions, no event
Voc, String 1 - 320.1V String 2 - 320V
Isc, String 1 - 1.8A String 2 - 1.8A
Sun 86W/m2


WET conditions, event number 35
Voc, String 1 - 342.2V String 2 - 342.1V
Isc, String 1 - 1.6A String 2 - 2.5A
Sun 88.8W/m2

Is the fault likely to be as simple as water ingress into one of the connections? We would like to be sure before we go to the expense of getting back onto the roof and remaking all connections.

Any help or advice would be much appreciated.

Cheers

 

[SolarCity]

I would say that water ingress into one of the connection is almost certainly the issue.

I'm surprised that the inverter picks this fault up though if the array is 'floating'. Is it a timber construction roof?

 

[DanPhillips]

Hello SolarCity

Yes the roof is timber construction which is a bit puzzling!!

The manual also mentions the thermal varistors as a possible cause for the fault, is this likely?

Cheers

 

[pvman]

I second water in a connector giving this fault.

Tom

 

[Earthstore]

I would doubt that it is a fault with the internals of the inverter as it is fine when it is dry, it could be a chaffed cable, but as others have said most likely a joint.

 

[DanPhillips]

After some discussion with SMA they have told me that I can increase the value of resistance measured between DC and potential earth which triggers this error. This is at default set to 1000Khm.

Surely increasing this value only hides the fault rather than repairing it?

How could you determine a safe value to increase it to?

 

[SolarCity]

Hmm... seems like a bad idea to me.

What's their solution to the oil light showing on the car dashboard? Disconnect the oil light?

 

[brightasaspark]

If I am reading this right - I am assuming the Inverter is checking leakage to earth which in the case of the wet conditions suggested would either be via a wet connector or a chaffed cable?

The 1000Kohm threshold (1Megohm) at 320 volts say would only mean some 0.3 Millamps is leaking to ground. Hardly worth worrying about so SMA suggestion to raise the threshold point seems sound.

 

[Gavin A]

Hello

We installed a PV system approx. 18 months ago which has recently started showing an event number 35 in wet weather conditions which according to the manual is a ground fault.

The system is a floating class 2 array so presumably we cannot have a ground fault.

We have tested the system during dry operating conditions and again whilst the event is displayed and the results were as follows


DRY conditions, no event
Voc, String 1 - 320.1V String 2 - 320V
Isc, String 1 - 1.8A String 2 - 1.8A
Sun 86W/m2


WET conditions, event number 35
Voc, String 1 - 342.2V String 2 - 342.1V
Isc, String 1 - 1.6A String 2 - 2.5A
Sun 88.8W/m2

Is the fault likely to be as simple as water ingress into one of the connections? We would like to be sure before we go to the expense of getting back onto the roof and remaking all connections.

Any help or advice would be much appreciated.

Cheers

I'm a little concerned that you view this as being the test results that would matter in this situation.

The manual clearly details the tests you should be carrying out in this situation.

Check the strings for ground faults, as described in section 11.1 ”Checking the PV Generator for a Ground Fault” (page 79).

If you look on page 79 of the SMA 4000TL installation manual there is a detailed description of the method to be used to check for a ground fault on the array, and to even locate which part of the array the ground fault is located at.

Even without this though, surely if you knew it was a ground fault then you should have at least been carrying out insulation resistance tests to earth (and if you've bonded the frame to an earth spike as many of us were doing 18 months ago on a TL inverter, then also test to the frame bonding). Or did you carry this test out but not post the results up?

SMA will presumably have been assuming you've followed these instructions and found no fault before they recommend you to adjust the setting at the inverter.

This fault could be water ingress to one of the connectors you've fitted, in which case you want to sort it out before it rusts and burns out like one we got called out to a while back (not installed by us).

Alternatively, this fault does crop up regularly on these inverters in wet conditions due to capacitance issues that are much worse on TL inverters, and are worse in damp conditions.

If you want to batter your head with the technical ins and outs of capacitance and TL inverters SMA have a handy guide to it here, but in general as long as you carry out the full tests detailed above to rule out an actual proper earth fault, then as SMA suggest there should be no problem with adjusting the level to prevent nuisance tripping of the inverter from a few mA of capacitive leakage current.

Sorry, I've pulled half the muscles in my back at the moment so I'm a bit grumpy, but I do think that anyone installing these systems professionally ought to be at least carrying out the tests that are detailed in the manual when presented with a fault code like this. Hopefully this helps anyway even if it comes across that I'm a bit narked. Hope you get it sorted, but please do test it properly before assuming it's just a capacitance issue.

 

[Gavin A]

If I am reading this right - I am assuming the Inverter is checking leakage to earth which in the case of the wet conditions suggested would either be via a wet connector or a chaffed cable?

The 1000Kohm threshold (1Megohm) at 320 volts say would only mean some 0.3 Millamps is leaking to ground. Hardly worth worrying about so SMA suggestion to raise the threshold point seems sound.

It depends what the cause of that fault is.

If the earth leakages is occuring because the connector is allowing water in, then the leakage path to earth would be via the water on the roof and walls (or if the frame is bonded, then via the water to the frame) which is a high resistance path, so would only show up as a very small leakage to ground, but the problem it indicates could be possible is well worth picking up before it rusts the connector and causes it to burn out.

Or it could just be a capacitance issue, in which case there's no problem, but best to rule out the more problematic potential cause of the fault code before assuming it's not a problem.