DESS | trade mode | unexpected pulling from grid

Been thinking about this after reading a few threads on DESS and trade mode configurations lately.

So the concept is sound enough — charge the battery during peak solar, sell excess overnight when grid export rates are better. On paper it makes sense, especially with the kind of time-of-use tariffs some of us are now able to access. But I'm genuinely sceptical about how reliably Victron's DESS actually executes this without ghosting in grid imports you didn't ask for.

My setup is a GX device with a Multiplus-II and a Fogstar Drift 48V pack on the boat, and I've seen weird behaviour when letting any automated scheduling touch the charge/discharge curves. Nothing as dramatic as unintended grid pulling, but enough to make me wary of handing full control over to an algorithm I can't fully interrogate.

Has anyone here actually run trade mode long-term on a UK tariff — Octopus Agile or similar — and seen it pull from the grid unexpectedly during what should've been a discharge window? I'm wondering whether it's a DESS forecasting issue (overestimating overnight load, maybe?), a SOC floor conflict with the BMS, or something in how the schedule priority is resolved when conditions don't match the prediction.

Would be particularly interested if anyone's running this on a cabin or static setup with more predictable load profiles — might behave differently to a boat where hotel loads jump around constantly.

Anyone dug into the VRM logs to correlate the grid pull timestamps with what DESS thought it was doing at the time?

@RelayNomad what's the actual trigger threshold you've got set for grid pull? Because on my static caravan setup I had something similar — Victron MultiPlus kept dipping into grid during what should've been pure battery discharge.

Turned out my minimum SOC was set too conservatively and the BMS was sending a early warning signal the DESS algorithm was interpreting as "grid backup needed."

Few questions:

• What BMS are you running?
• Is this happening at a predictable time of night or random?
• Have you checked the ESS assistant logs in VRM?

The "trade mode" label gets thrown around loosely too — are you actually on a proper Octopus Agile/Go tariff with the scheduler set up, or just manually timed?

Because the grid pull behaviour is completely different depending on which mode you're actually in under the hood.

@RelayNomad the bit that catches people out with DESS trade mode is the minimum SOC buffer interacting with the discharge floor setting — they're not the same parameter but the GX device can end up referencing both depending on grid conditions.

On my setup I found the Cerbo was pulling from grid at around 78% SOC because the scheduled charge window hadn't fully closed before DESS reasserted control. Ended up with a 15-minute overlap causing exactly the behaviour you're describing.

Worth checking in VRM whether your ESS mode transitions are logging correctly — the assistant state changes are visible in the advanced graphs if you enable them. Also verify your feed-in excess solar is actually enabled under the ESS assistant, not just the inverter settings, as they're separate toggles that confuse everything.

What GX firmware are you running? There were some DESS scheduling fixes in recent builds.

@Bazza60 makes a fair point about the SOC buffer — but has anyone actually mapped out how the discharge floor interacts with the inverter's own low-battery cutoff? On my garden office setup with a Victron Multiplus and Fogstar Drift cells, I found the BMS protection threshold and the DESS minimum SOC were essentially fighting each other during overnight discharge cycles. Result was the system pulling from grid earlier than expected and more aggressively than I'd configured.

What's the actual priority stack when both thresholds are active simultaneously? Does DESS override the inverter settings or does it defer to whichever cutoff triggers first? I've not found a straight answer in the Victron documentation and it's driving me mad.

@DodgyMechanic yeah I mapped this out on my van conversion setup last winter — the discharge floor and the DESS target SOC can genuinely fight each other if you've not locked down the priority order in the ESS assistant settings.

On my Victron kit, I found the inverter minimum SOC was being overridden intermittently when grid export was triggered. Took ages to nail down.

Worth checking your VEConfigure ESS tab — specifically whether "sustain voltage" is doing anything unexpected overnight. Mine was pulling from grid at 2am for no obvious reason until I realised the sustain threshold was set higher than my discharge floor.

Fogstar cells held up fine through all the testing at least 😄

@WonkyMender the bit nobody mentions is that when DESS calculates its overnight discharge schedule, if the grid import tariff drops below your configured sell threshold mid-cycle (Octopus Agile, anyone?), the ESS assistant essentially throws a tantrum and starts pulling from grid to protect the very SOC buffer that was supposed to be funding the discharge — seen it on my static caravan Victron setup, Cerbo logs full of confused state transitions at 3am while I'm asleep blissfully unaware I'm buying units at 28p when I should be selling at 15p. The fix is tightening your minimum SOC for sustain in the ESS assistant settings and making sure your dynamic grid setpoint isn't fighting your scheduled charge window — two separate parameters that Victron's documentation treats like they're unrelated, which they absolutely are not.