Unusual DESS behavior after upgrading Cerbo from version 3.66 to 3.75

Bit of a weird one and wondering if anyone else has hit this.

Upgraded my Cerbo to 3.75 last week — no other changes to the system — and DESS has been acting oddly ever since. Running a Victron setup with Fogstar Drift LiFePO4 cells and a custom BMS, DESS in trading mode targeting overnight cheap rate charging on Agile.

Before the update it was rock solid. Now it's occasionally ignoring the scheduled charge windows or cutting discharge short for no obvious reason. SOC targets look correct in the portal, ESS assistant settings haven't shifted, nothing in the VRM logs that jumps out immediately.

I've not touched charge/discharge limits or anything in the CCGX settings. Just the firmware bump.

A few things I'm wondering:

• Has 3.75 changed how DESS reads BMS allowable current flags?
• Could it be a MQTT/Modbus timing issue between the Cerbo and BMS?
• Anyone on Pylontech US5000s or similar seen the same? Pylontech's CANBUS comms can be a bit fussy with firmware changes historically

Rolling back to 3.66 feels like the sensible move while I dig into it, but curious whether anyone's already found a fix or a workaround. The Victron community forum has a thread but it's a bit all over the place.

Worth checking your DESS behaviour if you've updated recently — don't just assume it's fine because no alarms fired.

@ShortCircuit — what's DESS in this context? Genuinely asking as I've not come across that acronym before in relation to the Cerbo.

I've got a Cerbo GX running 3.75 on my cabin backup system and haven't noticed anything obviously wrong, but I'm not sure I'd recognise DESS-related behaviour if I don't know what I'm looking for. Could you describe what "acting oddly" actually means in practice — is it a charging issue, an alarm triggering incorrectly, something in the VRM dashboard?

Keen to understand because I'd rather know if there's something I should be monitoring on my own setup before it becomes a problem.

@GemmaStewart — Dynamic ESS. It's Victron's newer algorithm that replaced the older ESS logic, uses grid pricing data and solar forecasts to decide when to charge from grid, discharge to grid, or hold. Requires a Cerbo GX and is configured via VRM portal rather than locally.

@ShortCircuit — worth checking the DESS settings in VRM after the upgrade. There were some reports on the Victron Community forum of 3.74/3.75 resetting certain DESS parameters back to defaults, particularly the minimum SOC and grid setpoint values. Also double-check your grid code is still set correctly — seen that shift after firmware updates before.

What's it actually doing that seems off? Discharging when it shouldn't, not charging from solar as expected, something else? Narrows it down considerably.

@T6Solar beat me to it on explaining DESS but worth adding — it's still fairly "new" in Victron terms which basically means it's in that glorious phase where firmware updates either fix three things or break five different ones. Classic lottery.

@ShortCircuit what's your grid tariff setup — Agile? Because I've seen people report that

@GemmaStewart good question and don't worry — the acronym soup in Victron setups is genuinely baffling at first. You'll pick it up quickly here.

Back to the actual issue though — @ShortCircuit, I'm curious about a few specifics before jumping to conclusions:

• What grid pricing feed are you using? (Octopus Agile, N2EX, something else?)
• Is DESS making unexpected decisions, or just slower/different ones?
• Have you checked whether your MQTT or VRM portal connection changed behaviour post-update?

I'm running a Cerbo on my shepherd's hut setup but haven't pushed 3.75 yet specifically because I like to let others stress-test firmware first. This thread is exactly why. Does the Victron changelog for 3.75 mention anything touching the DESS scheduling logic?