MultiPlus-II + Pylontech RT12100G31 + DVCC: no Float at 100% SOC

Right, so this has been niggling at me for a few weeks now and I finally want to get some other eyes on it.

My tiny house setup runs a MultiPlus-II 48/3000 paired with a Pylontech battery bank — different model to some I've seen discussed, but the topology is essentially identical. Cerbo GX doing the orchestration, DVCC enabled, comms over VE.Can.

Here's what's bothering me. Once the bank hits 100% SOC, I'd expect the MultiPlus to drop into Float like a sensible, well-behaved inverter-charger. Instead, it just... sits in Bulk. Indefinitely. The BMS is clearly talking to the Cerbo — I can see the charge voltage limits being pushed through correctly — but the transition never happens.

My charger profile is set to Lithium on the MultiPlus side, and DVCC is handling the CVL/CCL hand-off from the Pylontech BMS. In theory, when the BMS signals it's full, the charge current should taper and Float should kick in naturally.

Has anyone else seen this with a Pylontech + DVCC combination specifically? I've read a few conflicting things about whether Float is even meaningful when DVCC is fully in control — some argue the BMS is effectively managing Float implicitly through current limiting. But the Victron documentation isn't exactly crystal clear on this, and I'd rather not just shrug and assume everything's fine.

Curious whether anyone's dug into the VE.Bus assistant settings or tweaked anything on the Cerbo side to resolve this. What does your charge cycle actually look like at the top end?

Hey @DriftGal, welcome to the rabbit hole! 😄

Quick question before anything else – have you checked what CCL (Charge Current Limit) your Pylontech is actually broadcasting via CAN when it hits 100%? With DVCC active, the MultiPlus should be handing control over to the BMS, so if Float isn't kicking in, it's often because the battery is still requesting a non-zero charge current rather than signalling "I'm done."

Pop into VRM or connect via VictronConnect and watch the "BMS controlled" charge current in real time as you approach full. The RT12100G31 can be a bit slow to drop its CCL to zero, which keeps the charger in Bulk/Absorption longer than you'd expect.

Also worth confirming your ESS assistant is configured correctly if you're running grid-tied. That can interfere with the normal charge cycle progression.

With DVCC enabled, the MultiPlus-II hands charge control entirely to the Pylontech BMS via the CAN bus — if the BMS isn't sending a Float CCL (typically dropping to ~2A tail current signal), Venus OS will just park in Bulk/Absorption indefinitely rather than transitioning to Float, which is perfectly normal behaviour for a BMS-controlled system and not actually a fault worth losing sleep over.

Check your Venus OS DVCC page: if SVS (Shared Voltage Sense) and SCS (Shared Current Sense) are both active, your Pylontech RT12100G31 is calling the shots — Float is essentially redundant when the BMS is managing the upper knee properly.

Worth confirming your firmware: Pylontech RT series occasionally needs a flash to fix CCL reporting bugs, and Victron's v3.x Venus OS tightened up DVCC compliance considerably.

@DriftGal — to add to what @OldSailor is getting at, the RT12100G31 specifically uses a CANBUS profile that doesn't always map cleanly to Victron's Float stage. What you'll often see instead is the BMS requesting 0A charge current at 100% SOC rather than dropping to a Float voltage — so the MultiPlus appears "stuck" in Bulk/Absorption rather than transitioning to Float. This is by design, not a fault.

Check your VictronConnect → MultiPlus → DVCC page and confirm SVS (Shared Voltage Sense) and SCS are both active. Also worth checking the BMS assistant version — Pylontech pushed a firmware update late 2023 that adjusted the CCL curve near full SOC. I saw identical behaviour on my narrowboat bank before updating.

@DriftGal one thing worth checking that nobody's mentioned yet — the Venus OS DVCC log on your GX device. Navigate to Settings > DVCC and watch the live CVL/CCL/DCL values being pushed from the Pylontech BMS over CAN.

Specifically, if your CVL is sitting below the Float voltage threshold the MultiPlus-II expects, it'll never transition to Float — it just holds Absorption indefinitely because the BMS hasn't signalled permission to reduce current. I've seen this exact behaviour on my motorhome setup with a Fogstar-adjacent BMS; turned out the CAN profile had a CVL mismatch of about 0.4V per cell.

Also worth confirming which CAN profile is selected in Venus OS — CANBUS BMS (500 kbit/s) versus the Pylontech-specific profile behaves quite differently in practice.