What have I done wrong here (Sol-Ark 15k)

[hpeyerl]

Discovered this the unpleasant way, while up on my roof.

In short (pun not intended) I discovered 240vdc present between the ground bar in my Sol-Ark 15k, and the MPPT + terminals, if the 'DC' switch is turned 'ON'.

This has the unfortunate "side effect" of making my entire roof be 240vdc positive if someone touches any of the PV positive lines even if my RSD is active (ie: 0v PV).

In the photo, my meter positive is resting on one of the bolts on my GND bus bar, and the meter negative is resting on the MPPT3+ screw. The panels are disconnected but the ground that comes from the roof is still connected (far left bolt on the ground bus bar). If I turn off the 'DC' switch on the side of the inverter, it all goes to 0vdc.

So I'm asking myself whether it is expected that the inverter is putting 240vdc on the ground (and since the ground is bonded to neutral, there's also 240vdc on the neutral).

 

[Skypower]

Do you have the panel frames grounded?

 

[hpeyerl]

Do you have the panel frames grounded?

Yes. Lots of grounding everywhere. Trough, pull boxes, panel(s), disconnect, inverter, etc.

If it was 240Vac I'd be looking there but it's 240VDC which has me confused.

Edit: I just realized you meant the solar panels, not electrical panels. I am dumb.

Answer: yes, the panel frames are grounded to my Kinetic racking which is grounded to my combiner box which is grounded to my inverter which is grounded to everything else.

 

[Trappy]

stupid comment alert but I think it might be related to how you have the grounds connected specifically the bare copper one on leftmost.

 

[hpeyerl]

stupid comment alert but I think it might be related to how you have the grounds connected specifically the bare copper one on leftmost.

I like possibly stupid comments because often they lead to discoveries.

Where should the racking (and thus panel frames) be grounded? I assumed to the system ground.

I suppose it's a worth disconnecting that ground from the bus bar and confirming that the Sol-Ark is still putting 240v DC on the ground. I assume I must be the only person who has encountered this because it was a shocking discovery when I made it (pun intended).

 

[Trappy]

I think the ground from the combiner box should be going to an earth ground rod and not into the Sol Ark.

 

[hpeyerl]

I think the ground from the combiner box should be going to an earth ground rod and not into the Sol Ark.

well, in my system, the Sol-Ark is connected to the Earth Ground through 24" of #6 copper and a 1/4-20 bolt.

 

[BentleyJ]

I discovered 240vdc present between the ground bar in my Sol-Ark 15k, and the MPPT + terminals, if the 'DC' switch is turned 'ON'.

Yes, there is supposed to be a potential difference between the PV panel output wiring and ground, PV's are a source of power! Neither PV+ or PV- is supposed to be grounded. Only the PV panel frames and racking are to be grounded. Therefore when the switch is ON it simply means the PV's are producing voltage as they are supposed to. When the switch is turned OFF, PV positive at the MPPT input is disconnected thus 0 potential difference between the two.

This has the unfortunate "side effect" of making my entire roof be 240vdc positive if someone touches any of the PV positive lines even if my RSD is active (ie: 0v PV).

Is your roof metal? Then it should be grounded. How is the PV array leaking potential to the roof? As stated above neither the PV+ or PV- wiring should be attached to anything except the MPPT input terminals (unless there a combiner box in the system) The entire PV DC system should be insulated all the way from the box on the back of the PV's to the MPPT inputs.

So I'm asking myself whether it is expected that the inverter is putting 240vdc on the ground (and since the ground is bonded to neutral, there's also 240vdc on the neutral).

The inverter is NOT putting 240VDC on the ground. The meter is measuring the expected difference in potential between the 2. If you measure the AC voltage between L1 or L2 to ground it should be 120V, yes. Does that mean 120V is being put to ground? no just the opposite. Meter is measuring the difference between 2 independent wires that are not shorted out.

You stated. "my entire roof to be 240vdc positive if someone touches any of the PV positive lines". What?? the PV positive wires are insulated and why would you be "touching" any of the PV wires or bare terminal connections when the system is live during the day producing dangerous voltage.

Please describe how you got shocked.

 

[hpeyerl]

Yes, there is supposed to be a potential difference between the PV panel output wiring and ground, PV's are a source of power! Neither PV+ or PV- is supposed to be grounded. Only the PV panel frames and racking are to be grounded. Therefore when the switch is ON it simply means the PV's are producing voltage as they are supposed to. When the switch is turned OFF, PV positive at the MPPT input is disconnected thus 0 potential difference between the two.

PV+/- are not grounded. Only the frames/racking etc are grounded. I had my Tigo CCA RSD active meaning the panels were not producing power which I confirmed by hooking my meter to PV+/- in my combiner box.

You stated. "my entire roof to be 240vdc positive if someone touches any of the PV positive lines". What?? the PV positive wires are insulated and why would you be "touching" any of the PV wires or bare terminal connections when the system is live during the day producing dangerous voltage.

Please describe how you got shocked.

Because it is a rooftop setup, I have Tigo TS4-A-O's, one on each panel and a Tigo CCA that allows the panels to produce power. When I disable the CCA, the panels stop producing power (via the TS4-A-O's).

So I was up on the roof, inside my combiner box (this is a new system), and had exposed one of the PV leads. So while touching the metal roofing (which is grounded to my rails and thus to my Sol-Ark), I inadvertently touched the exposed PV+ lead, which I'll remind you, was not producing power by way of the Tigo optimizer/rsd combo units, and received a shock.

My first thought was the Tigo's weren't doing their job, so I went down to the Sol-Ark and completely disconnected the PV+/PV- leads from the Sol-Ark (which you can see in the photo above) so now the only thing connecting my combiner box to the Sol-Ark is the ground wire from my racking. There was 240VDC potential between the MPPT+ leads in the Sol-Ark and the Ground bus-bar in the Sol-Ark. Panels were not implicated.

This potential goes away when I turn off the DC switch on the side of the Sol-Ark

My assumption is that the 240VDC is just rectified AC and that what I did wrong was not disabling the DC before poking around inside my combiner box and that the full root of the problem is that my RSD button is wired to the Tigo CCA and not the Sol-Ark RSD pins which would shut down the Sol-Ark DC and AC.

 

[Quattrohead]

Grab yourself a couple of 120 volt light bulbs and put them in series and connect to the point where you were getting a shock and see if the bulbs light up. If they do you have an extremely serious problem but if they don't you just have a leakage or induced voltage. How or why I don't know but try that first.

 

[Skypower]

Many MPPT (PV inputs) terminals output voltage when the inverter is on (no panels connected). If you should short the PV inputs or bring them to ground you could hurt the PV section of the inverter. It’s a good idea when connecting to them that the inverter be off and if possible, a PV disconnect opened.

 

[hpeyerl]

Grab yourself a couple of 120 volt light bulbs and put them in series and connect to the point where you were getting a shock and see if the bulbs light up. If they do you have an extremely serious problem but if they don't you just have a leakage or induced voltage. How or why I don't know but try that first.

Good Thinking. 100w incandescent bulb hooked up and no voltage nor light.

So it looks like an induced voltage. Nice clean 240vdc though (without the bulb connected). no real noise I can see.



So I guess this isn't to be concerned about.

 

[hpeyerl]

Many MPPT (PV inputs) terminals output voltage when the inverter is on (no panels connected). If you should short the PV inputs or bring them to ground you could hurt the PV section of the inverter. It’s a good idea when connecting to them that the inverter be off and if possible, a PV disconnect opened.

Yeah, so I guess this is my takeaway from all this.

 

[Quattrohead]

Bottom line is, assume every electrical connection could bite you no matter what it's known state appears to be.

 

[timselectric]

Never work on electrical systems, until.......
1. You turn all possible sources off.
And
2. You test and verify that it's off.

This is the best way to avoid mistakes.

Luckily, you didn't get seriously hurt. And you learned more about your system.