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

Had exactly this argument with my own setup last winter — spoiler: they were both lying to me.

The BMV-712 is only as good as your shunt calibration and the Peukert settings you've dialled in, so if you've not spent a proper afternoon fiddling with those, it's essentially making educated guesses. Meanwhile most budget BMS units have the SOC accuracy of a broken dartboard.

My Fogstar Drift 12V BMS was reading 67% whilst the BMV was showing 54% — turned out my charge efficiency factor in the Victron was set to the default (which assumes lead acid behaviour, not lithium). Knocked it up to 99% and suddenly they were singing from the same hymn sheet.

General rule of thumb I've landed on:

• Trust the BMS for high/low voltage cutoffs — it's watching individual cell voltages and will save your pack
• Trust the BMV-712 for day-to-day SOC tracking once it's properly calibrated — it's doing the coulomb counting properly
• Trust neither when you've just done a partial charge on a cloudy Tuesday in November 😅

The real tell is a full charge cycle — both should reset to 100% at the top, and if they diverge significantly over a discharge, your Peukert exponent or charge efficiency is off.

Anyone else running a dual-setup like this? Curious whether those with Pylontech or BYD batteries find the BMS comms via DVCC makes the BMV redundant entirely.

AngliaNomad | 847 posts | Midlands

@ExBrickie94 Ha, brilliant — "both lying" is exactly the right framing. I'd add that the BMS SOC reading is often just voltage-based under the hood, which is notoriously wobbly in the middle range of LiFePO4 cells where the discharge curve is essentially flat. The BMV wins on methodology (coulomb counting is genuinely more accurate day-to-day) but it accumulates drift over time if your charged-full detection threshold isn't set correctly.

My approach: treat the BMV as your working gauge but use periodic full-charge cycles to let it resync. If the two readings diverge by more than about 10%, that's your cue to investigate rather than just pick the more optimistic one — which, human nature being what it is, we're all tempted to do at 11pm on a cloudy day! 😄

OldSailor | 1,203 posts | Array

@ExBrickie94 The BMS protects the cells but couldn't care less about accurate SOC — it's a bouncer, not an accountant; the BMV-712 is the accountant but only if you've set your Peukert exponent, charged efficiency factor, and — critically — done a proper synchronisation point at 100% after a full absorption cycle. My Fogstar Drift 200Ah and Victron BMV-712 agree to within 2% consistently, but that took three afternoons of faffing with settings I'm not ashamed to admit I initially ignored entirely. Trust the BMV-712 long-term provided your shunt is on the negative terminal with nothing else bypassing it — any rogue earth connection and your coulomb counting is fiction dressed as data.

CaddyProject | 312 posts | Array

Worth adding the practical workflow I use:

• BMV-712 = my day-to-day SOC reference (Peukert set correctly, tail current dialled in tight)
• BMS = my protection layer, full stop

The BMV drifts if you never hit a proper full charge to let it sync. Running lithium on a boat means I sometimes float for days without hitting absorption properly — BMV goes wonky fast.

Key fix: set your charged voltage and tail current in the BMV to match what your BMS actually reports as full. Mine's set to 27.6v / 2% tail on a 24v Fogstar pack. Once those align, the two readings converge much better.

@OldSailor is right that the BMS doesn't really do SOC — it just throws the handbrake on.

DriftKing | 634 posts | South West

Great thread. One thing nobody's mentioned yet — sync events are your best friend here. I let the BMV-712 auto-sync at 100% only when I'm genuinely sure the bank is full: tail current drops to under 1% of capacity AND voltage has held at absorption for a solid 20 minutes. If you're letting it sync too eagerly, your reference point drifts and the gap between BMV and BMS widens over time.

Also worth checking your shunt connection — any resistance from a dodgy crimp or loose bolt will skew BMV readings progressively. Had mine reading 8% optimistic for months before I found a slightly loose terminal.

@CaddyProject's workflow is solid. I'd just add: periodically cross-reference both against a full charge cycle to catch long-term drift before it catches you out.