lv 6548 (in split-phase mode) grounding in bypass mode


Solar Addict
Jan 31, 2021
Since all this discussion about the growatt 5000es has started, I was curious if there was a definite answer on the LV 6548 grounding.
This video from Dan Fitzpatrick kinda shows the issue that I assumed, switching to/from utility power seems to be a bit funky. I've read other threads where they say there is a ground fault when the inverter clicks over modes, but it clears within a minute or two.

I am mostly curious/confirming if:

1.) Sub/CL panel should *not* have neutral/ground bond (as expect) in bypass mode
2.) Sub/CL panel *needs* a N/G bond in non-bypass mode, as seems to be indicated by Dan's video
3.) I've read the inverter should automatically do #2 with an internal contactor?



Photon Sorcerer
Jun 21, 2020
The video does not give much to work with.

All the branch circuits from the aux box should have nothing passing current to branch's ground return wire so there should be no ground current on grounding wire to aux box. If he forgot to disconnect the aux box neutral bus bar ground strap to case of aux box then he could get current on the ground wire.

Since most breaker boxes are used for main breaker panel, they commonly come default assembled with a strap from the isolated plastic mounted neutral bus bar to case. It is small and not obvious unless you know to look for it. This ground strap from neutral bus bar to box case should be disconnected if panel is used for an aux panel. Neutral bus in aux panel should be connected to inverter neutral then to main panel neutral with no disconnect switches along the way.

By leaving the aux box neutral ground strap connected he is essentially splitting neutral line current between real neutral wire and ground wire. This would be why he is seeing current on ground wire. Remember neutral is connected to ground at main panel


Photon Sorcerer
Jan 3, 2021
Dan, who is a licensed electrician, posted about it in this thread.

The solution was he wired the ground directly from the inverter to the main service panel. Previously, he had input ground from a junction box above the inverters that contained a busbar and it was not bonded.

Now, as to why this occurred. My theory is this, the inverter will bond N-G when in inverter mode. As you mention, when switching to bypass mode, there appears to be a ground fault on the ground wire that clears in a short time. This is an important clue as N-G bond at the inverter when in battery mode is accomplished by a relay. My theory is the unit is not switching the relay immediately. This could be a programming issue or it could even a capacitor bank discharging when the switch to bypass occurs but I wouldn't see that taking a minute. It also depends on the type of relay used. It could also be the unit checks for some type of ground thru a EGC when shutting the inverter down. It's not like the inverter completely discharges all residual power in 1 ms.

I really like this diagram and it is close to how Dan originally installed his. While the diagram doesn't show the ground wiring, you do not run ground thru the Power panel shown. Instead run it either thru the transfer switch (not switched, only on a busbar by itself) or directly from the main panel.

I will be wiring mine similar, Dan installed a subpanel after the inverters and ran critical loads off it while running his main panel thru the grid with some circuits still grid powered. Instead, I will install a new main panel that will only have 2 breakers in it. One will go to the inverters, the other to the transfer switch. This panel would be bonded in a normal install but I have an acreage and the N-G bond occurs at the meter pole with the breaker panel installed there. This panel however will need a grounding conductor to the grounding bus bar as this is a multiwire branch circuit. My original main panel in the house will then become fed thru the inverters. It too will be unbonded. As the distance is short, I will run a EGC directly from the inverters to the new main panel and not thru the transfer switch to the inverters.
Last edited: