Battery going offline when min SOC reached

Been having a similar head-scratcher with my setup lately and wondering if anyone else has come across this.

Running a Victron MultiPlus-II with a pair of Fogstar Drift 48V batteries and a Cerbo GX. Since I updated the firmware across everything last month, when the system hits the minimum SOC threshold the batteries aren't just limiting discharge — they're actually dropping off the BMS comms entirely for a short period. Inverter throws a warning, then everything reconnects and recovers, but it's doing my head in.

My suspicion is it's something in how the BMS is handshaking with the GX after the update. The batteries themselves seem fine — voltage looks healthy, no cell-level alarms. It's almost like the BMS is doing a hard "nope" rather than a graceful reduction in discharge current.

A few things I'm trying to work out:

• Is this a deliberate protection behaviour in newer firmware, or a genuine fault?
• Has anyone adjusted the DVCC settings in VRM to smooth this over?
• Could it be the minimum SOC % being set too close to the actual BMS low-voltage cutoff, leaving no buffer?

Currently got the minimum SOC sitting at 20% which on previous firmware was completely fine. Tiny house setup so I'm not pulling massive loads, but it's still annoying to have the inverter wobbling around at night.

Would be curious whether anyone running Pylontech has seen the same — imagine the BMS behaviour is broadly similar. Might be worth a thread comparing notes across different battery brands to see if it's a Victron-side change.

@Thommo had this exact thing with my shepherd's hut setup.

Worth checking if the Fogstar's BMS is actually disconnecting at low SOC or if it's the MultiPlus itself triggering a low voltage cutoff — two different problems with similar symptoms.

In VE.Configure, check your DC input low shut-down voltage — mine was set slightly too high and kept tripping before the BMS needed to act. Also look at the ESS minimum SOC setting in the Cerbo if you're running ESS mode.

The Fogstar Drift BMS can be a bit aggressive on the low end too. What voltages are you actually seeing in VRM when it drops off?

@AnneWatson's point about BMS disconnection is worth pursuing — is your MultiPlus set to ESS or standalone inverter mode?

I had something similar where my min SOC setting in VRM was causing the MultiPlus to essentially "fight" the BMS when they both tried to act at low voltage simultaneously. The Cerbo was seeing a sudden drop to 0% SOC right at the cutoff point rather than a graceful shutdown.

Worth checking your DVCC settings in the Cerbo — specifically whether "Limit charge current" is enabled and whether the BMS is actually being given control authority. If the Fogstar's BMS cuts before the MultiPlus expects it, the GX can lose comms with the battery entirely and just reports it as offline.

Are you using the Fogstar's CANbus connection to the Cerbo, or just voltage-based monitoring?

@Thommo what does your DVCC settings look like in the Cerbo? Had something similar on my motorhome rig — when the BMS trips the battery offline, the Cerbo can sometimes lose communication and the MultiPlus gets confused about what state it's actually in.

Worth checking System → DVCC and making sure "Limit charge current" is being controlled properly. Also, is your minimum SOC set in both the Fogstar app and ESS settings? If they don't align you can get a race condition where the BMS disconnects before the MultiPlus has chance to gracefully shut down load.

What does VRM show in the logs right at the point it drops out — is it a sudden voltage collapse or a gradual bleed to cutoff?

Good shout from @WonkyRigger on the DVCC angle. One thing nobody's mentioned yet — check your ESS minimum SOC setting versus what the Fogstar BMS is actually set to cut off at. If your ESS floor is set higher than the BMS cutoff, the MultiPlus should back off gracefully, but I've seen setups where the BMS trips first if there's any lag in the communication.

On my boat I had a similar ghost-disconnect issue. Turned out the VE.Can termination resistor on one end wasn't seated properly — BMS comms were flaky enough that the Cerbo occasionally lost the battery and the MultiPlus went into a protective shutdown. Worth checking your CANBUS connections if you're using that rather than VE.Bus.

What firmware are you running on the Cerbo? There were some known quirks pre-v3.00 around BMS timeout handling.