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

Had this exact situation on my narrowboat last spring — BMV-712 sitting at 78% whilst the BMS on my Fogstar Drift 200Ah was confidently reading 85%. Took me a good few weeks to work out what was happening.

The short answer: neither is perfectly right, but for different reasons.

The BMV works on coulomb counting — it tracks every amp flowing in and out. Sounds ideal, but small measurement errors compound over time, and it drifts unless it gets a proper full charge to reset the 100% sync point. If your system rarely hits absorption/float fully, it'll gradually read lower than reality.

The BMS, on the other hand, is estimating SOC primarily from cell voltages, which on lithium is notoriously flat across the middle range. It can be wildly optimistic between roughly 20–80%.

What I settled on was this:

• Trust the BMV for day-to-day consumption tracking — it's better at showing how fast you're draining
• Trust the BMS for low-SOC warnings — cell-level visibility matters when you're near empty
• Let a confirmed full charge (proper absorption completion) re-sync the BMV to 100% regularly

Worth checking your BMV's charged voltage and tail current settings in VictronConnect — mine were factory defaults and completely wrong for lithium chemistry.

Anyone else running a dual-monitoring setup? Curious whether others have found a smarter way to reconcile the two readings, particularly on boats where partial state of charge living is just the reality.

@OffGridTerry the BMS is measuring cell voltage → SOC lookup table, which is essentially guesswork on a partially-loaded circuit, whereas the BMV-712 is doing proper coulomb counting — trust the Victron, full stop.

Key thing to check: has your BMV-712 ever hit a proper sync point? It only resets to 100% when it sees the tail current drop below your configured threshold at absorption voltage. If your Fogstar Drift hasn't had a genuine full charge cycle since you set it up, the BMV will drift pessimistically over time.

• Check your charged voltage setting (~14.2V for LiFePO4)
• Set tail current to around 4% of capacity (8A for a 200Ah)
• Let it do one proper full cycle and they'll likely converge

My cabin setup had the same 7% discrepancy until I sorted the sync parameters — sorted overnight.

@FogstarFan makes a fair point about the lookup table limitations, but I'm curious — does the accuracy shift depending on load conditions at the time you're reading it?

On my setup I've noticed the BMS figure looks more believable when the battery's been resting for 20+ minutes with no charge or discharge happening. Under load the voltage sag throws the lookup table right off.

The BMV-712 does its coulomb counting continuously, but it needs a proper sync point to stay accurate — has anyone checked whether their synchronisation settings are configured correctly? If the "charged voltage" threshold is set too low, mine was never triggering a full sync and the drift just accumulated over weeks.

Would be interested to know what charge voltage @OffGridTerry is running the Fogstar Drift to — that could be a factor here.

@George1972 yes, temperature absolutely hammers LFP voltage-SOC curve accuracy. A cold battery (say 5°C) will show a notably lower resting voltage than the same SOC at 20°C — so the BMS lookup table will under-read in winter and potentially over-read when warm.

The BMV-712 isn't immune either — its coulomb counting drifts if you haven't properly configured the Peukert exponent and charge efficiency factor in VictronConnect. Out of the box defaults are set for lead-acid assumptions.

My setup in the static caravan uses a Fogstar 200Ah with a BMV-712. I trust the BMV long-term once it's had several full sync cycles, but after a patchy week of partial cycling I'll cross-reference both and lean toward whichever is more pessimistic. Treat the BMS voltage reading as a sanity check rather than gospel.

@BayJason that's a really useful point about temperature — does that mean the BMV-712 is actually more reliable in cold conditions then, since it's tracking coulombs rather than interpreting voltage curves?

Asking because I'm mid-build on my van and expecting to park up in Scotland through winter. I've got a Fogstar Drift 100Ah on order and trying to decide whether to bother with the BMV-712 alongside the built-in BMS, or if that's overkill at this scale.

One thing I'm unclear on — how does the BMV handle self-discharge over a few days of the van sitting unused? Presumably the coulomb counting drifts if it's not syncing back to 100% regularly via a proper full charge?

@VanLee not exactly more reliable, but potentially more consistent in cold conditions — there's a difference. The BMV-712 is coulomb counting regardless of temperature, so it doesn't care whether it's 5°C or 25°C for its core calculation. However, it does have a Peukert compensation setting and a temperature coefficient input if you've got the temp sensor connected. Worth checking those are actually configured correctly for LFP — the defaults are set for lead-acid and will throw your readings off considerably. The BMS voltage-based reading meanwhile becomes increasingly meaningless in the cold as @BayJason touched on. That said, coulomb counting drift is still real — both methods have their weaknesses. Honestly the best approach on a narrowboat is using them together and letting a proper synchronisation at full charge periodically reset the BMV. That's what sorted Terry's situation if I recall correctly.