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

Been through exactly this rabbit hole with my Fogstar Drift 200Ah LiFePO4 last winter.

Short answer: neither is gospel, but they're measuring different things.

The BMV-712 is doing coulomb counting — it's tracking electrons in and out, starting from whenever it last hit a sync point (usually 100% SOC). Drift a bit on the shunt calibration, leave a small parasitic load uncounted, or fail to reach a proper full charge for a few cycles, and it quietly wanders off reality.

Your BMS, meanwhile, is estimating SOC from cell voltages. That sounds more "honest" but LiFePO4's flat discharge curve means a tiny voltage difference mid-range represents a massive swing in capacity. It's genuinely terrible at mid-SOC precision.

What I actually do:

• Trust the BMV-712 for day-to-day numbers, because it's smoother and more responsive
• Use the BMS as a sanity check at the extremes — near full or near empty is where voltage-based readings get meaningful again
• Make sure the BMV syncs properly at 100% by letting the battery sit at absorption for long enough (I had mine syncing too early)

The gap tends to close once you nail the sync settings. Mine were 2-3% apart constantly until I tightened up the tail current and charged voltage thresholds in the BMV configuration.

What's the spread you're seeing, and what chemistry is your pack? Makes a difference to which way I'd lean.

@VanJim had this exact argument with my narrowboat setup — BMV drifts like a drunk on a towpath if you never hit 100% to let it resync, whereas the BMS is just watching cell voltages and making educated guesses at the top and bottom of the curve where LiFePO4 is famously flat as a pancake.

Basically you've got two different flavours of wrong competing for your attention.

I just trust the BMV during discharge and let the BMS handle protection duties — that's what they're actually designed for respectively. If they disagree wildly, it usually means your BMV hasn't had a proper full-charge resync in ages.

Stick a Victron scheduled charge in once a week to let it resync and the drift sorts itself out sharpish.

@VanJim @DodgySocket my Fogstar + BMV combo disagreed by about 12% until I let it hit a proper full charge and synchronised the BMV — now they shake hands like old mates at a pub.

The BMS is basically a bouncer watching voltage thresholds, the BMV is an accountant tracking every electron in and out — neither's lying, they've just got different jobs.

Practical tip: if your Victron's efficiency factor and Peukert settings are still factory default, that's your drift culprit right there — LiFePO4 needs different values than the lead-acid assumptions baked in from the factory.

@VanJim nailed the fundamentals, but there's a practical calibration point worth adding: your BMV-712's charge efficiency factor and Peukert exponent settings matter enormously here. Factory defaults are often wrong for LiFePO4.

On my static caravan setup with 400Ah of Fogstar cells, I had a persistent 8-10% drift until I set the charge efficiency to 99% (LiFePO4 is nearly lossless compared to lead-acid) and dropped the Peukert exponent to 1.05.

Also — what's your tail current set to for the sync trigger? If it's too low, the BMV never sees a valid 100% sync point and just drifts indefinitely regardless of how many full charges you do.

BMS percentage is typically just a voltage lookup table anyway. Useful as a sanity check, not a precision instrument.

Had this exact tension with my garden office setup — Victron BMV-712 against the BMS on a Fogstar Drift 100Ah.

What I'd add: the BMS SOC is often just voltage-based, which is notoriously unreliable mid-charge on LiFePO4 because of that flat discharge curve. The BMV's coulomb counting is generally more accurate between syncs.

Practical fix that sorted mine:

• Set BMV charge efficiency to 99% for LiFePO4 (not the default)
• Make sure your tail current and charged voltage thresholds are dialled in properly — Victron's defaults assume lead-acid
• Let it hit a genuine full charge to reset drift, as @LazyRanger mentioned

Once calibrated correctly, I'd trust the BMV day-to-day over the BMS readout. BMS is really there for protection, not accurate monitoring.

One thing nobody's mentioned yet — check your Peukert exponent in the BMV settings. Default is 1.25 which is fine for lead acid but LiFePO4 is much flatter, closer to 1.05. Wrong value means the coulomb counting drifts faster under heavy loads.

Sorted mine after the BMV kept reading pessimistically when running my inverter hard. Small tweak, big difference in accuracy.

@WhatsAFuse65 charge efficiency factor matters too obviously, but I'd fix Peukert first — it compounds the error more noticeably on LiFePO4 chemistry.