I am posting here as I use HomeAutomation to control my fox system through Modbus.
I noticed recently as I was setting up a new PyScript system to postpone charging till later in the day that even when the system was placed in Feed-In Priority the battery would still slowly charge through the day when the sun was shining and the house usage was low.
For example here you can see that I set feed-in at around 8am after an early morning solar charge and the SOC rose until the dryer was turned on when it fell but continued to rise when the dryer finished. During all this time feed-in was set.
I was initially suspicious of my Pyscript programming and thought it might be continually switching briefly back to self-use but I can find no evidence for this. And you can see that during all this time the battery charge line is flat and 0kW.
I have since disabled the Modbus integration and set feed-in in the app with the same results - actually the later part of the graph above is with Modbus disconnected.
Is this 'leakage' of battery charge normal when feed-in is set?
If your PV generation is greater than your house load, and your export limit, then the overspill will charge the battery if there is spare room.
This avoids PV clipping, though some of your details are not shown on your picture to get the proof.
This avoids PV clipping, though some of your details are not shown on your picture to get the proof.
Thanks,
That is a fair reason. But I am suspicious it's not what was causing it today - it wasn't that sunny.
How do I check what my export limit is set to? Is it in the app or on the inverter? Are these the settings of which you speak?
If they are then the export (purple) comes nowhere near the limit....
That is a fair reason. But I am suspicious it's not what was causing it today - it wasn't that sunny.
How do I check what my export limit is set to? Is it in the app or on the inverter? Are these the settings of which you speak?
If they are then the export (purple) comes nowhere near the limit....
There's a lot of good information of how things work here:
https://github.com/TonyM1958/HA-FoxESS-Modbus/wiki
Look at Inverter Work Modes in the 11 additional pages.
I have no idea about ModBus, but I'm still very grateful to Tony M for this.
'Note: users report current KH and H1-G2 inverters leak charge to the battery in feedin mode, causing the battery to charge to 100%. The only known work around is to set the max charge current to 0A during the day.'
https://github.com/TonyM1958/HA-FoxESS-Modbus/wiki
Look at Inverter Work Modes in the 11 additional pages.
I have no idea about ModBus, but I'm still very grateful to Tony M for this.
'Note: users report current KH and H1-G2 inverters leak charge to the battery in feedin mode, causing the battery to charge to 100%. The only known work around is to set the max charge current to 0A during the day.'
Mine is a KH so could be the known bug.
Thanks for the link - I will read through them later. I think modbus exposes charge current so I should be able to work with that for now.
Thanks for the link - I will read through them later. I think modbus exposes charge current so I should be able to work with that for now.
entdgc wrote: ↑Tue May 05, 2026 5:02 pm Thanks,
That is a fair reason. But I am suspicious it's not what was causing it today - it wasn't that sunny.
How do I check what my export limit is set to? Is it in the app or on the inverter? Are these the settings of which you speak?
Screenshot 2026-05-05 at 16.56.01.png
If they are then the export (purple) comes nowhere near the limit....
Yeah you weren't going above 13kW for this to divert some power into the battery.
What I think occurs is the BMS tracks the cell temperature and seems to increase the SoC.
Here is an earlier post I made, and you can see the SoC is increasing... but like you there was not an excess of PV to account for it (apart from the time you see I have selected) Either the stats are terrible with Fox, and it IS putting some PV into the battery, or as I think happens, as it heats up, the BMS recalculates a higher SoC.
viewtopic.php?p=14977#p14977
A picture I posted somewhere on the forum, where the SoC climbs up (matching your findings)
ETA: I am not fully sure it is related to Cell Temp Low, as I can see in HA logs, the SoC increasing whilst the Temp is dropping... maybe PV is being passed to the battery, but failing to log correctly.
I added a line to my PyScript to set the charge current to 0 whenever feed-in is selected and so far there is no sign of any leakage into the battery. So it seems it is likely the firmware bug mentioned by Dave/Tony. Thanks to the great knowledge of this forum! 


There I was thinking I had a nice nuclear powered battery
I had a theory about the temperature affecting the SoC, but looking at the logs it was increasing as temps dropped.
Logically it had to be some PV being 'leaked' into the battery.
Great forum indeed, and some really useful information out there for sure.
I had a theory about the temperature affecting the SoC, but looking at the logs it was increasing as temps dropped.
Logically it had to be some PV being 'leaked' into the battery.
Great forum indeed, and some really useful information out there for sure.