Growatt 3000 24v Reversed ground/Hot reading

[timselectric]

Until the inverter is tested with all wires disconnected from AC input and AC output for continuity between N-G with the inverter off, there is no clear answer.

Incorrect.
You can take voltage readings with the system running, to verify if a bond exists.
If a bond is present, you can't tell where it's coming from unless you disconnect everything.
But, because we know that there's no bond present. No further testing is needed.
For the OP, we are already passed this part. And now we just need to create a bond in the proper way.

 

[Zwy]

Tim, did you look at the Paneltronics wiring diagram? https://www.paneltronics.com/docs/wiring/2305.pdf?v=20240219042206

There is a N-G bond wired into this panel. It is switched when on shore power and N-G bond is provided from shore power.

The only way 45V can be present on G with that N-G bond provided by the panel or shore power is a parallel path.

 

[ShockinglyAverage]

Just for the record, THIS is my panels diagram not the one you keep sharing.

 

[timselectric]

There is a N-G bond wired into this panel. It is switched when on shore power and N-G bond is provided from shore power.

The AIO is providing the shore power to this panel. So I would assume that the N/G bond is switched off. (Or possibly was removed at a previous time)

The only way 45V can be present on G with that N-G bond provided by the panel or shore power is a parallel path.

The 45v reading shows that the ground is floating.
This is just a reading of potential difference. As there is no 45v source.
Whether there are 1 or 10 (just an example) N/G bonds, the reading will be the same as the source.
It only differs when the system is floating.

 

[Zwy]

Just for the record, THIS is my panels diagram not the one you keep sharing. View attachment 197028

That certainly helps. Thank you. Most helpful.

I'm looking at your photo comparing the wiring diagram above to the photo and I see this. Note G appears to be touching stud on N busbar and quite possible the threads have cut thru the insulation. This would create a parallel path. Makes sense too when under shore power.



You will still need to determine if the inverter dynamically bonds N-G. That determines how N-G bond is handled on inverter power.

 

[ShockinglyAverage]

That certainly helps. Thank you. Most helpful.

I'm looking at your photo comparing the wiring diagram above to the photo and I see this. Note G appears to be touching stud on N busbar and quite possible the threads have cut thru the insulation. This would create a parallel path. Makes sense too when under shore power.

View attachment 197040

You will still need to determine if the inverter dynamically bonds N-G. That determines how N-G bond is handled on inverter power.

It’s not touching it’s just the perspective

 

[ShockinglyAverage]

Also I have checked the voltage and resistances. You’re just wrong. I know that’s painful to admit sometimes, I’ve been there. But the panel is not bonded and the inverter is not dynamically binding anything. When I ground it at the plug, the voltages read proper, from both the inverter and on the panels bus bars. When I leave the inverter to try and do it on its own I get ghost voltages.

 

[Zwy]

Also I have checked the voltage and resistances. You’re just wrong. I know that’s painful to admit sometimes, I’ve been there. But the panel is not bonded and the inverter is not dynamically binding anything. When I ground it at the plug, the voltages read proper, from both the inverter and on the panels bus bars. When I leave the inverter to try and do it on its own I get ghost voltages.

Something is wrong though, you get the "Rev Hot/Ground" and have 45V on G.

Have you tried disconnecting the Paneltronics panel and wire up a cord to inverter AC outlet with a receptacle to test the inverter standalone? Then add a N-G bonding jumper to see the result?

It is possible there is a fault in the inverter so testing will help support any warranty claim. Testing standalone eliminates any possibility of downstream faults.

It’s not touching it’s just the perspective

I wondered and stated it as a possibility. That is good, that possibility is eliminated.