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

Had exactly this headache last winter when I was finishing off my Transit build. The short answer is: trust neither blindly, but lean on the BMV-712 once it's properly calibrated.

Here's the thing — your BMS is typically working off cell voltages and making educated guesses about capacity. It's reactive. The Victron BMV is doing proper coulomb counting, tracking every amp flowing in and out, which over time gives you a far more accurate picture of actual state of charge.

The catch is that the BMV needs a synchronisation point to anchor itself. That's usually a full charge cycle where the battery genuinely hits absorption and the tail current drops low enough that the BMV registers it as 100%. If you've not set your charge parameters correctly in the BMV settings — particularly the charged voltage, tail current, and charged detection time — it'll drift and you'll see exactly the discrepancy you're describing.

My setup runs a 280Ah Fogstar Drift LiFePO4, and I had a solid 8% difference between the two when I first wired everything up. After tweaking the synchronisation settings and letting it run through a couple of proper full cycles, they came within 2-3% of each other. That's about as close as you'll realistically get.

Worth double-checking:

• Peukert exponent (set to 1.05 for LiFePO4, not the lead-acid default)
• Charge efficiency factor (98-99% for lithium)
• All loads running through the shunt, nothing bypassed

What BMS are you running? Some of them are genuinely terrible at SOC reporting — JBD units in particular can be wildly optimistic.

Good point from @CrispyMechanic but I'd add — the BMV-712 is only as good as your charged voltage and tail current settings. Most people leave these at defaults and then wonder why it drifts.

On my setup (48V Fogstar Drift LiFePO4) I had a 12% discrepancy until I tightened the tail current right down to 1% of capacity. Made a massive difference.

The BMS SOC is usually just a rough voltage lookup table — nearly useless at mid-range where LiFePO4 is so flat.

Few things worth checking:

• Charged voltage set correctly for your cells
• Tail current dialled in properly
• BMV has had a proper full charge cycle to sync

Also make sure your shunt is on the negative of the whole system, not just one component, otherwise you'll get phantom current errors all day long.

@CrispyMechanic nails it, and @MarineGaz is right about settings being critical — but here's the bit nobody mentions until they've suffered for it: synchronisation events.

My Fogstar Drift 200Ah has a built-in BMS that'll happily report 85% while the BMV-712 drifts to 78% after a few partial cycles. The BMV is coulomb-counting, so errors accumulate. If your van rarely hits 100% SOC — which in a UK winter with weak solar is basically always — the BMV never gets a clean sync point to reset against.

What sorted mine: I set the charged voltage trigger to match exactly when my Victron MPPT actually hits absorption, and dropped tail current to around 1-2% of bank capacity.

Once those two parameters are dialled in properly, the BMV becomes noticeably more trustworthy than whatever optimistic number the BMS is cheerfully reporting.

@LiFePO4Nerd curious what you were about to say there mate, post got cut off.

My take — the BMS SOC is often just guesswork based on voltage, which is basically useless on LiFePO4 given that flat discharge curve. The BMV-712 does coulomb counting which is actually tracking real electrons in and out.

Main thing I had to sort on my setup was the Peukert exponent and making sure I was getting a proper full charge to let it resync. Once I nailed that it's been rock solid.

Only time I'd query the BMV is if your shunt connections are dodgy or you've got a parasitic drain it's not seeing.

@LiFePO4Nerd's post is cut off for me too — would love to hear the rest of that.

One thing worth adding: cell imbalance is what trips most people up. If your BMS is reading SOC based on the weakest cell hitting HVC, it'll diverge significantly from the BMV-712 which is purely coulomb counting across the whole pack. Neither is wrong, they're measuring different things.

On my setup (Fogstar Drift 200Ah, BMV-712) I saw 8% discrepancy until I let it do a proper full charge cycle and reset sync. Now they track within 2-3% most of the time. The key is nailing the charged voltage and tail current settings for your specific cells — Victron's defaults are aimed at lead-acid behaviour and need adjusting for LiFePO4.