Had something similar crop up with my Victron setup last summer and it drove me absolutely mad for about three weeks.
What I think happens — and someone please correct me if I'm wrong — is that once the battery hits absorption/float and the BMS signals "full," the MPPT essentially throttles back to maintenance levels. If your loads then kick in, the system defaults to grid rather than ramping the solar back up to cover them. It's a prioritisation/logic issue rather than a hardware fault.
A few things worth checking:
- ESS assistant settings — specifically your "minimum SOC" and "feed-in excess solar" config. These interact in weird ways.
- BMS communication — if you're running a Fogstar or similar lithium bank, check whether the DVCC settings in VRM are actually talking to the BMS properly. A dodgy CANbus connection caused mine to behave almost identically to what you're describing.
- Grid setpoint — is yours set to 0W or slightly negative? If it's sitting at +50W or something, it'll always pull a little from grid regardless.
Mine turned out to be a firmware update that had quietly reset a couple of ESS parameters back to defaults. Logged into VRM one morning and half my settings had reverted. Victron's changelog mentioned it but easy to miss.
Has anyone else noticed this happening more since the v3.x firmware rollouts? Feels like there's been a few quirks introduced. Would be interested to know if this is isolated or more widespread across different battery setups.