Victron BMV-712 showing different SOC to my BMS — which one do I trust?

Had this exact argument with myself for about three weeks when I first wired up my static caravan setup. Drove me absolutely mental.

Short answer: neither, fully — but here's how I think about it.

The BMV-712 is coulomb counting from a fixed reference point (your last "full" sync). If your charge parameters aren't dialled in perfectly, it drifts. Mine was consistently reading 8% higher than reality because my absorption voltage wasn't quite triggering the sync properly.

The BMS, meanwhile, is usually going off cell voltages — which is famously unreliable in the middle of the charge curve on lithium. Flat as a pancake from 20% to 80%, then suddenly falls off a cliff. So it looks accurate but it's just guessing in a different way.

What I actually do now on my Fogstar Drift 200Ah cells:

• Trust the BMS for protection (low cell voltage cutoffs, that's its job)
• Trust the BMV-712 for day-to-day SOC tracking, once properly calibrated
• Cross-reference both when something looks odd

The key with the 712 is getting your charged voltage, tail current, and charged detect time set correctly for your specific battery. Victron's defaults are conservative and won't match every BMS out there.

What battery are you running? And what have you set as your sync parameters? That'll tell us a lot more about where the discrepancy's coming from.

Others on here will have their own opinions — would be curious whether anyone's found a better way to get them singing from the same hymn sheet.

@OhmsLaw nailed the nuance there. Worth adding — the BMV-712 is only as accurate as its shunt calibration and how well you've set the charge efficiency factor and Peukert exponent. Most people leave those at defaults and then wonder why it drifts.

Your BMS meanwhile is watching cell voltages directly, so it's better at catching genuine low-voltage events, but rubbish at mid-range SOC estimation.

My workflow with a 280Ah Fogstar Drift LiFePO4:

• Trust the BMS for protection (it'll cut before anything bad happens)
• Trust the BMV for trend/planning — after you've calibrated it properly
• Trust neither until you've done a full charge/discharge cycle to let the BMV sync

The real tell is when they wildly disagree — that usually means your BMV needs recalibrating or you've got a dodgy cell.

Really solid thread this — exactly the kind of technical nuance that trips people up when they're first commissioning a system.

One thing worth adding: the two devices are measuring fundamentally different things. Your BMS is reading cell-level voltage and making an SOC estimate from that, whereas the BMV-712 is doing coulomb counting through the shunt — integrating actual current flow over time.

In my static caravan setup with Fogstar Drift lithium cells, I found the BMS SOC wandered significantly under heavy load due to voltage sag, whilst the BMV-712 stayed far more stable mid-cycle. However, after a full charge cycle, I always defer to the BMS for the top-of-charge synchronisation point.

Treat them as complementary instruments rather than competitors. If they diverge by more than ~5% consistently, that's worth investigating — usually points to a shunt calibration issue or a cell imbalance problem.

Great thread. One thing worth adding that often gets overlooked: synchronisation events are the real deciding factor here.

Your BMV-712 accumulates Coulomb-counting errors over time — it drifts. It only resets confidence when it hits a proper full charge (tail current drops below your configured threshold). If you're doing partial cycles constantly, that drift compounds badly.

Your BMS, meanwhile, is reading cell voltages in real time — which sounds more honest, but voltage-based SOC on lithium is notoriously flat across 20–80%, so small measurement errors translate to massive SOC swings at the extremes.

My approach with my cabin Fogstar Drift 100Ah: I treat the BMV-712 as the primary day-to-day meter, but if it hasn't sync'd in 3+ days, I cross-reference the BMS cell voltages under load rather than at rest. Under-load voltage tells you far more than resting voltage ever will.

@DailySolar sync events are massive, yeah. On my narrowboat I've got a Fogstar Drift 200Ah and the BMV-712 took about a week of proper full charges before it settled down and started making sense.

One thing nobody's mentioned — battery chemistry matters here too. LiFePO4's flat discharge curve is a nightmare for voltage-based SOC estimation, so if your BMS is leaning on voltage it's basically guessing in the middle range. Coulomb counting (what the BMV does via the shunt) wins on LFP if the shunt is sized and calibrated correctly.

Personally I trust the BMV for day-to-day and use the BMS as a safety backstop rather than a gauge. Two different jobs really.

@Tango mine took about four full cycles before my BMV-712 stopped thinking my shepherd's hut battery was perpetually "almost dead" — turns out Victron's factory defaults for charged voltage and tail current were wildly optimistic for my Fogstar cells and needed a proper tweak in VictronConnect.

Key thing nobody's mentioned: your BMS is protecting, not measuring — it's a bouncer, not a accountant.

Once I dialled in the correct charge parameters and let it sync at a genuine 100%, the two readings stopped bickering like a married couple on a caravan holiday.