A shared array is causing insulation resistance error? - Help diagnosing a weird project.

[b34rb34r]

I was a local sollar installer and I was called by a home owner for assistance when their solar system wasn't working and their installer couldn't solve it. This happens a lot around here.

So, they have installed an SMA 8kW sunnyboy inverter with 20 longi 455 panels in two strings (10 modules each). The inverter is showing error 3501, which is low insulation resistance error or grounding fault.

What baffled me is that, the installer took 6 modules from one string, connected them with Y connectors and is running a battery charger MPPT with them while all 20 modules are also connected to the SMA inverter. The charge controller seems to be working fine and charging the battery system well.

The wires coming from the 6 modules are directly connected to the MPPT charge controller too. No isolators, no grounding provided.

Now, I have never done anything like this simply because intuitively I assume that's not possible parallely, but if I were to do that, I would have made a switch so the owner can manually choose between On grid and off grid during load shedding.

But, is that the reason why the inverter is showing the grounding fault? Any thoughts are welcome, this is more educational for me than solving the actual problem to be honest.

 

[littleharbor2]

Never heard of such a thing. MPPT is trying to track the Maximum power point and we have a separate controller trying to pull power from the same string. The installer took a set of parallel connectors and made a splitter out of them. Time to get those parallel connectors out and give the controller its own panels.

 

[RCinFLA]

A PV panel can sometimes delaminate the plastic backing seal plastic at the edges of panel. Moisture can intrude the delamination and create a conductive electroylsis leakage path from the cells or their connections to aluminum frame of panel.

The top of panel typically have a couple of conductive straps that run close to the edge of glass so they are prime location for leakage to edge frame.

Since the SMA is a transformer-less inverter it has no AC grid to PV panel isolation and has ground fault detection to check for too much leakage from the PV lines to grid ground.

There is likely leakage from one or more of the panels from the PV circuit to the aluminum frame of panel which is grounded causing the ground fault error.

On each panel, you can check the resistance with an ohm meter from pos and neg leads of panel to the panel's aluminum frame. The resistance should be greater than 200k ohms.

 

[Solar Guppy]

What baffled me is that, the installer took 6 modules from one string, connected them with Y connectors and is running a battery charger MPPT with them while all 20 modules are also connected to the SMA inverter. The charge controller seems to be working fine and charging the battery system well.

A huge NO, the owner would be lucky equipement hasn't failed. Solar panels cannot be "shared", and with almost all current AIO inverters the are not isolated this would end instantly in a shower of sparks and flames. Restore the array ( remove the "Y" connectors splitters ) and hope for the owner it's not all dead.

 

[b34rb34r]

A PV panel can sometimes delaminate the plastic backing seal plastic on the edge of panel. Moisture can intrude the delamination and create a conductive electroylsis leakage path from the cells or their connections to aluminum frame of panel.

The top of panel typically have a couple of conductive straps that run close to the edge of glass so they are prime location for leakage to edge frame.

Since the SMA is a transformer-less inverter it has no AC grid to PV panel isolation and has ground fault detection to check for too much leakage from the PV lines to grid ground.

There is likely leakage from one or more of the panels from the PV circuit to the aluminum frame of panel which is grounded causing the ground fault error.

On each panel, you can check the resistance with an ohm meter from pos and neg leads of panel to the panel's frame. The resistance should be greater than 200k ohms.

Okay, I am aware of that. But haven't checked yet.

But does the current going into the battery system make the inverter assume there is a current leakage?

 

[b34rb34r]

A huge NO, the owner would be lucky equipement hasn't failed. Solar panels cannot be "shared", and with almost all current AIO inverters the are not isolated this would end instantly in a shower of sparks and flames. Restore the array ( remove the "Y" connectors splitters ) and hope for the owner it's not all dead.

I know right? It felt so stupid to see that and the audacity of the installer to do this.

 

[RCinFLA]

Okay, I am aware of that. But haven't checked yet.

But does the current going into the battery system make the inverter assume there is a current leakage?

A direct to battery charge controller is more tolerant of PV line leakage to ground. On a direct to battery charge controller you can usually even ground the negative PV lead. 100k ohms of leakage to ground is enough to trip a GFI on transformer-less inverter.

On a transformer-less inverter the PV pos and neg lines are riding on top of the AC line voltage. If you are standing on ground and touch one of the PV pos or neg connections you will get a nasty shock from the AC voltage the PV is riding on top of.

Because the GT inverter AC output is not electrically isolated you cannot parallel MPPT inputs on two inverters running in parallel or series stacked for 120/240v operation. It can blow out the inverters.

You also do not want to run same PV array to separate direct to battery MPPT charge controllers as they will interfere with their ability to find MPPT point.

 

[Solar Guppy]

I know right? It felt so stupid to see that and the audacity of the installer to do this.

Are building permits and engineering required at this location? if not, anyone can call themselves an "installer"