Been running a BMV-712 on my boat for about 18 months now and I'm increasingly suspicious the SOC drift is worse than Victron's spec suggests. My setup is a 200Ah Fogstar Drift LiFePO4 with a 500A shunt wired per the manual — single negative tail, all loads and charging through the shunt, nothing bypassing it. By the book, supposedly.
The issue is that after roughly 3–4 weeks of mixed use (solar top-ups, occasional shore power, engine alternator), the BMV reckons I'm at 94% when I put a proper clamp meter on the individual circuits and do the maths, the actual consumption figures don't tally. It's not massive — maybe 4–6% off — but on a boat where I'm relying on this for emergency backup decisions it matters. I've already tightened the Peukert exponent and charge efficiency factor per Victron's guidance (currently 1.05 and 99% respectively for LiFePO4), and I do let it hit 100% sync via absorption at least once a week.
Has anyone run a parallel logging setup — a second shunt, a Cerbo, anything — to actually validate BMV accuracy over weeks rather than days? Wondering whether the tail current threshold for "charged" detection is doing something funny, or whether there's genuine cumulative error I should be chasing down. Concrete numbers from anyone who's done this properly would be gold.