Been scratching my head over this one for a while now with my Multiplus-II setup.
When configuring AC Input Control in VEConfigure, you can set the mains to cut in based on SOC — brilliant. But when it comes to the disconnect threshold, you're limited to voltage only. No SOC option. Anyone else find this oddly asymmetrical?
My suspicion is it's a deliberate safety choice on Victron's part. Voltage is a near-instantaneous measurement, so under heavy load your battery voltage can sag significantly below the "true" SOC level. If disconnect were SOC-driven, there's a risk of over-discharging before the BMS even gets a chance to intervene — particularly with older lead-acid banks where SOC estimation is less reliable anyway.
That said, I'm running Fogstar Drift LiFePO4 with a proper Victron-compatible BMS feeding accurate SOC data over VE.Can. In that scenario, SOC-based disconnect feels like it should be trustworthy enough.
The workaround I've landed on is setting a fairly conservative voltage disconnect threshold — accounting for load-induced sag — combined with an ESS assistant rule in Node-RED that monitors SOC and adjusts the input current limit dynamically. Not elegant, but it works.
Has anyone managed a cleaner solution? Particularly curious whether anyone's used DVCC settings or Grid Setpoint tuning to effectively mimic SOC-based disconnect behaviour without going full Node-RED rabbit hole.
Would love to know if this is something Victron have addressed in newer firmware — I'm on v497 currently. Worth raising on the Victron Community forum directly, or has someone already done that?