Grid tie ground fault isolation problem.

[qrper]

I know it's not related to Midnite solar, and that being said, I'll end up installing some new breaker boxes to help solve this issue with ground isolation.

I have 24-250W modules wired as two groups of 12 modules. This is a solaredge grid tie, so each module has one of their optimizers.

It seems every time we get a wet/cloudy/foggy day I get a isolation fault, and the inverter shuts down. Rats!

So standing in the mud and cold, I disconnected one of the strings, the one I suspect, and after re-pairing the optimizers, the inverter started.

The way it's wire, I have to open the, box, pull out the wires and re-close everything. I plan on installing a MNPV6 with fuse holders (the dc input is 480ish volts)

I don't have a Megger and gosh are they expensive, so anyone have some tips and tricks to locate a ground fault in a grid tie (0r battery based dc system for that matter) ?

I'm mike, wb8vge

[boB]

Can you isolate one of the two strings to the one that has the fault ?

As long as it's not a fault with the inverter itself

[qrper]

Bob... and the group....

If it wasn't for bad luck as they say.

Yes, I was able to separate the two strings. One of them used brand new factory fresh panels, the other string is comprised of used modules I purchased from eBay. they were from a decommissioned solar farm.

I don't have a megger but do have several good Fluke meters. According to the trouble shooting guide by Solaredge, a panel should test 40MΩ with both leads connected together and measured from ground.

The new panels tested fine. According to the guide line, anything under 40mΩ is considered bad. My Fluke read 519K on the first used panel. Ah poop!
Switching over to volts, and measured from + PV to ground I saw several VOLTS.
It appears that all the used panels have some type of leakage that is adding up to the point the inverter shuts down.

I have no issue with the Classic 150 tripping a ground fault using the same modules in my 48V dc system.

My fix is going to be adding a MNPV-6 with fuse holders because of the 400 V dc and when the weather gets wet and humid, pull the fuse so that part of the array won't shut down the inverter.

The only fix that I can see, is to remove the panels and replace them with new ones. Looks like I'll have 24â€"250W modules I won't be able to use.

It just seems that every time I try and save a buck, something comes up and bites me in the @^#.

Mike

[boB]

What happens if you just don't ground the module frames ?   

These GT inverters are really sensitive !

Merry Christmas !

[qrper]

What happens if you just don't ground the module frames ?   

These GT inverters are really sensitive !

Merry Christmas !

That’s the rub! Neither lead of the POV panel is connected to the frame. But!! If you measure from PV+ to ground, you read voltage even through the neg lead is dangling in the air.

Mike

[boB]

That’s the rub! Neither lead of the POV panel is connected to the frame. But!! If you measure from PV+ to ground, you read voltage even through the neg lead is dangling in the air.

Mike

Were both PV leads hanging out or was one of them connected to the inverter ?  I am assuming they were both hanging loose.  Remember that the inverter itself, if non-isolated has a connection to ground.

Was that AC or DC voltage ?   There will always be AC voltage picked up just by 60Hz in the air.

If it was DC, you can probably trace it to one particular module but connecting one PV line to GND and measuring V drops.  Maybe even shorting the array (careful of arcs !) and measruing DCV alone the string to GND/frames.

The question then is, how much possible current is behind that voltage (AC or DC) ?  If you take a 100 Meg-Ohm resistance and hook it between the two points, you might measure nothing, in which case there is no current possible and therefore, safe.

I guess that's what the pre-turn on tests do in the new inverters.  But they understand this.  So I'm just not sure.

[qrper]

That’s the rub! Neither lead of the POV panel is connected to the frame. But!! If you measure from PV+ to ground, you read voltage even through the neg lead is dangling in the air.

Mike

Were both PV leads hanging out or was one of them connected to the inverter ?  I am assuming they were both hanging loose.  Remember that the inverter itself, if non-isolated has a connection to ground.

Was that AC or DC voltage ?   There will always be AC voltage picked up just by 60Hz in the air.

If it was DC, you can probably trace it to one particular module but connecting one PV line to GND and measuring V drops.  Maybe even shorting the array (careful of arcs !) and measruing DCV alone the string to GND/frames.

The question then is, how much possible current is behind that voltage (AC or DC) ?  If you take a 100 Meg-Ohm resistance and hook it between the two points, you might measure nothing, in which case there is no current possible and therefore, safe.

I guess that's what the pre-turn on tests do in the new inverters.  But they understand this.  So I'm just not sure.

bob,

I have the pv panels removed from the inverter. Their leads (pv) are dangling in mid air. I have them disconnected from the optomizers, too.
According to SolarEdge, a typical value of resistance between both the pv + and pv- *(shorted) should be nominal 80MΩ.
I don't have a Megger, so I can't read that high of resistance, however my Fluke 77 shows 517KΩ to ground.

OUt of the 12 panels on that subarray, 8 of them show resistance to ground. If I connected the +lead of the pv panel to my meter and the neg lead of the meter to the frame, with the neg lead of the pv dangling, I read anywhere from 5 to 27 V.
Of course there's hardly any current flow because of the +500K resistance between the PV+ and the frame, but enough to trip the inverter down.

So.... I decided to check the panels I used on my 48 Vdc system. They are the same brand from the same supplier. Guess what? I have dc flow from the PV+ to the frame! Yup. Apparently there's not enough current flow to make the Classic 150 think there is a ground fault.

It never ends...

I'm mike, wb8vge

[boB]

The current required to trip the Classic's gnd fault is at least 1/2 amp so WAY more than you are going to get with this.

I think the deal is that with these non-isolated inverters is that they treat the leakage kind of like a bathroom GFCI or
a few milliamps ?   They are VERY sensitive !   A bit too sensitive maybe ?

Maybe you could cheat by adding a high value resistor to the frames from the opposite polarity and  being a common mode cancellation signal could balance it out ?  Naaahhhh....   Better not.  Probably wouldn't really work anyway

boB

[qrper]

Bob and the group,

I purchased a Klein MEGGER. It's not a $900 one, but it's also not a cheap Chinese meter either.

What an eye opener.

According to Solaredge, a good PV panel should show at least 80MΩ of resistance when both leads are connected together and measured to the panel's frame.

I tested on the of the new Hyundai panels and the meter shows 4000MΩ

One of the used Trina panels metered at 1.2MΩ
Another one, would not test because the meter saw voltage on the leads to ground.

I need to test all the used panels, however the weather is going into the pooper for the next few days. I guess it really doesn't matter, because if the majority of them are bad, they needed replaced.

Apparently, they will work okay in a battery based system. So far, my Classic 150 hasn't had any issue with them, and there is 7Kw worth of then connected together for a 48 V battery system.

I'm mike, wb8vge

[boB]

YaY !!!   Good deal, Mike !

Glad you found something !

It might even be that you can possibly find the bad part of the module and even fix it ?

Place the tape and glue in the right spot ?   Maybe if it is obvious.

Happy new year !