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

Had this exact headache with my garden office setup last year. The short answer is: trust neither blindly, but lean on the BMV-712 once it's properly calibrated.

Here's my thinking:

• Your BMS is typically measuring cell voltages and making a rough SOC guess from that. Voltage-based SOC on lithium is notoriously flat in the middle range — tiny voltage differences = massive SOC swings on paper
• The BMV-712 does proper coulomb counting (tracking actual current in/out), which is far more accurate over time provided it gets a clean sync point to reset from

The common culprit for drift between the two is the charge efficiency factor and Peukert exponent settings in the BMV. Most people leave these at defaults, which aren't always right for lithium. Worth checking those first.

Also — does your BMV get a regular full charge cycle to re-sync? If your system's sitting at 80% most of the time (as mine does to preserve cycle life), the BMV can wander without that reference point. I ended up setting a weekly scheduled full charge just to keep things honest.

What BMS are you running? Some of the cheaper units are genuinely terrible at SOC reporting. My old one was reading 60% when I was nearly flat — not ideal when it's powering a work setup!

Would be curious what the actual percentage gap is between the two readings — a few percent is normal, 20%+ suggests something needs attention.

Great thread @BoxerProject! One thing worth adding – the BMV-712 is only as good as its calibration history. If you've had any unmonitored current flowing (perhaps through a secondary load wired before the shunt, which is a very common wiring mistake), the BMV will have drifted without knowing it.

Worth checking: does your BMV regularly hit 100% SOC and sync? If it's been weeks since a full synchronisation, accumulated errors stack up surprisingly quickly, especially with LiFePO4 where the voltage curve is so flat.

My setup took about a fortnight of tweaking the charged voltage threshold and tail current settings before the BMV and my JK BMS were consistently agreeing within 2-3%. Now they track each other beautifully. Patience is definitely required! What battery chemistry are you running? That changes the calibration approach quite significantly.

@BoxerProject nails it, and this hits close to home for me.

On my boat I run a Victron BMV-712 alongside a Daly BMS on a 200Ah Fogstar Drift LiFePO4 bank. For the first few weeks the two were regularly 8-12% apart — maddening.

The thing that sorted it was tail current calibration. My charged voltage was set correctly but the tail current threshold was far too low, so the BMV never actually registered a proper full charge sync. Once I bumped that to 2% of capacity it started agreeing with the BMS far more consistently.

@LindaCross66 is right about calibration history — a BMV that's been half-configured is worse than useless because it gives you false confidence.

The BMS SOC on most budget units is essentially voltage-based guesswork anyway. Coulomb counting from the BMV wins every time, if it's been allowed to sync properly.

Really interesting thread, and something I've been wrestling with at my cabin setup too.

One thing nobody's mentioned yet — does your BMS actually measure current accurately, or is it just estimating SOC from voltage? Most cheaper BMS units (looking at you, JBD/Overkill-based ones) are pretty rough on coulomb counting.

The BMV-712's shunt is doing proper current measurement at the battery terminals, which is why I ended up trusting it more for my EV charging scheduling. I use it to trigger my Victron MPPT to stop dumping charge once I hit a certain SOC threshold.

Worth checking:

• Is your shunt only measuring battery current, with nothing else on the chassis side?
• Have you set the Peukert exponent correctly for your chemistry?

If the BMV is wired sloppily, all bets are off regardless of which number you prefer.

Really solid points all round here.

One thing worth adding for anyone using a Fogstar Drift or similar lithium with a "smart" BMS that reports SOC over Bluetooth — that figure is almost always calculated from resting voltage curves, which are notoriously flat in the middle of LiFePO4's discharge profile. The BMS genuinely cannot distinguish 40% from 60% SOC in that plateau region without coulomb counting.

The BMV-712's coulomb counter accumulates error over time, yes, but it self-corrects every time the battery reaches your configured "charged" threshold (100% sync point). On my narrowboat setup I've got this set to 27.6V absorption with a 2A tail current — once that triggers, the BMV resets cleanly and stays remarkably accurate until the next full cycle.

@FormerMariner36's Daly comment is relevant here too — Daly's SOC reporting is particularly unreliable under load due to poor internal resistance compensation.

Good thread. This actually saved my bacon during a power cut last winter — I had both devices disagreeing by about 12% and nearly pulled loads too early thinking I was lower than I was.

The thing that sorted it for me was doing a proper full charge cycle to let the BMV-712 synchronise to 100%, then running it down slowly under a known load. After that, the drift between it and my BMS shrank to under 3%.

Worth noting: the BMV-712's Peukert correction and tail current settings matter enormously with lithium. Most people leave these on the lead-acid defaults out of the box, which will absolutely throw your readings off.

@DevonDweller's point about Fogstar cells is interesting — I'd be curious whether their built-in BMS reports via a different protocol, because that could explain some of the discrepancies people are seeing.