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

Been through exactly this with my own setup. Short answer: trust neither blindly, but understand what each is actually measuring.

The BMV-712 uses coulomb counting — it tracks charge in and charge out with a shunt, then applies your configured efficiency factor. It's only as good as your initial calibration and how well you've set the charge efficiency factor (typically 99% for lithium). If it's never seen a proper full charge cycle to re-sync, the drift compounds over time.

Your BMS is doing something different entirely — it's usually making a voltage-based SOC estimate, which on LFP is notoriously unreliable across the flat middle of the discharge curve. Some BMS units (particularly cheaper ones bundled with Fogstar or similar cells) have very crude lookup tables.

What I actually do:

• Use the BMV-712 as the primary SOC indicator during normal operation
• Use the BMS as a protection layer, not a gauge
• Ensure the BMV syncs to 100% reliably at each full absorption cycle — this is critical, check your "charged voltage" and "tail current" settings in VictronConnect

The gap between them is usually either a calibration issue on the BMV or the BMS interpreting resting voltage incorrectly after a load transient.

What's the actual percentage difference you're seeing, and at what point in the cycle? Also, what BMS are you running? Some have configurable SOC algorithms that can be tuned to be far more accurate.

Worth checking whether anyone else here has the same BMS pairing — there might be a known offset issue.

CopperWarden | 847 posts

Great summary @LazySocket. One thing worth adding — the BMV-712's accuracy lives and dies by two settings people often neglect: the Peukert exponent and the charge efficiency factor. Get those dialled in for your specific battery chemistry and the coulomb counting becomes dramatically more reliable.

Also worth checking your synchronisation trigger. If the BMV isn't hitting its "fully charged" threshold regularly, small errors compound over weeks and the drift becomes significant. I've seen setups where the absorption voltage was set slightly low, so the BMV never synchronised properly and was showing 20% out after a fortnight.

What battery chemistry are you running? That changes which figure I'd lean on as the sanity check.

Good point from @CopperWarden. On the boat I've found the BMV drifts noticeably over a few weeks if the synchronisation point isn't set right. The key thing most people miss is the tail current setting — if it's too high, the BMV thinks the battery is full before it actually is, and the whole count is off from that point.

With my Fogstar lithiums I had to drop the tail current right down and tighten the voltage threshold before the readings made any sense. BMS cell-level data is useful for spotting a rogue cell, but for day-to-day SOC I lean on the BMV once it's properly calibrated.

Nothing like two gadgets having an argument while your fridge quietly murders your battery — coulomb counting drift is the silent killer, and I learned that the hard way when my Fogstar 280Ah cells were "at 60%" according to the BMV right before everything went dark at 2am.

My rule now: BMS protects, BMV informs — treat the BMS like a bouncer and the BMV like a slightly unreliable mate who mostly knows what's going on.

Also worth checking your shunt placement — had mine wired wrong for three months wondering why the numbers were mental, turned out half my loads were bypassing it entirely. Classic ExBrickie moment right there.

SophieFisher | 203 posts

Something nobody's mentioned yet: the two devices are often disagreeing because they're measuring different things at different points in the circuit.

On my narrowboat I discovered my BMV shunt was positioned after the DC distribution bar, meaning certain loads weren't being counted at all. The BMS, naturally, sees everything happening at the cell level. That's not drift — that's a fundamental wiring error quietly compounding itself.

Before assuming it's a calibration or algorithm problem, I'd trace exactly where your shunt sits relative to all loads and charge sources. Renogy and Victron both cover this in their documentation, though it's easy to skim past.

Once I corrected the positioning, the two readings converged considerably. They'll never be identical — they're not designed to be — but persistent large discrepancies usually point to something structural rather than a settings tweak.

Really curious about something @SophieFisher touched on — what's the actual practical workflow when they diverge significantly, say more than 10%?

My Fogstar Drift 100Ah with its built-in BMS is reading about 8% higher than my BMV-712 right now and I've been just... ignoring it, which probably isn't ideal.

Do people typically:

• Force a BMV sync when the battery hits the BMS's 100% threshold?
• Adjust the Peukert exponent and charge efficiency factor in the BMV settings?
• Just use the BMS as the floor for low-voltage cutoff and trust the BMV for day-to-day tracking?

I've read conflicting things about whether tweaking the BMV settings actually narrows the gap meaningfully over time or just creates a different kind of inaccuracy. Anyone with LiFePO4 specifically had success dialling these in properly?