We are operating a Victron MultiPlus II GX with the ESS Assistant, configured using VEConfigure. We also have some customsetpoint logic using Node-RED.
System configuration:
-
ESS Assistant enabled in VEConfigure
-
We do not make use of BatteryLife
-
Battery type configured in VEConfigure
-
Voltage limits are properly configured for both Grid and Inverter
-
Battery connected via CAN-bus
-
Confirmed with the battery manufacturer that the battery does not shut down intermittently; DC supply remains continuously available
Our observed behavior (intermittent):
-
The issue occurs seemingly irregularly, roughly once every 2–3 weeks, sometimes each couple of days.
-
We observed it after the system charged normally until the battery reported 100% SoC.
-
Charging then stops as expected once the battery is full.
-
The system remains in this state until sunset.
-
At sunset, our external control software (Node-RED) requests discharge by setting the appropriate ESS setpoint.
-
At that moment, the inverter state is “Off”, not “On”.
-
As a result, discharge does not start.
-
The inverter must be manually switched back to “On” via the Remote Console’s software switch to resume normal operation.
The core problem:
-
The inverter appears to switch from “On” to “Off” autonomously, without user interaction.
-
We cannot identify what triggers this transition to “Off”.
-
Our Node-RED software does not issue any command to switch the inverter off; this behavior is not implemented on our side.
Additional observation (in VRM Portal):
-
Shortly before the inverter goes to “Off”, the VE.Bus State frequently toggles between Bulk and Absorption multiple times.
-
After this phase, the inverter ends up in the Off state.
Relevant screenshots from the VRM Portal are attached, showing the VE.Bus State behavior prior to the inverter switching off.
Question:
Under which conditions can a MultiPlus II GX (maybe because of the ESS assistant?) switch itself from On to Off without user interaction?
Are there known ESS behaviors, protections, CAN-battery interactions, or firmware-related mechanisms that could explain this?
Any guidance on where to look next would be appreciated.
