Anyone else struggling to get accurate SOC readings with a cheap shunt on a lithium battery?

Picked up a 100A shunt monitor off Amazon for about £12 to keep an eye on my 200Ah LiFePO4 leisure battery in the van. Followed the wiring instructions, set the battery capacity to 200Ah and the charged voltage to 14.2V, but the state of charge reading seems to drift all over the place. After a full charge it'll show 100%, then by the next morning — even with no loads running — it's down to 87% without anything obvious pulling current.

I'm wondering if it's down to the shunt not being wired as the single negative path, because I've got the chassis earth also tied into the negative busbar near it. Could that be causing current to bypass the shunt entirely and throw the counting off? I did read somewhere that every single negative in the system needs to pass through the shunt for coulomb counting to work properly, but I wanted to check with people who've actually sorted this out in practice.

Has anyone managed to get one of these budget shunts working reliably, or is it worth just biting the bullet and getting a Victron BMV-712? I know it's £90-odd but if it actually works properly it's probably worth it compared to constantly second-guessing my battery state.

@Gibbo the core issue with cheap shunts on LiFePO4 is that they typically can't handle the near-flat voltage curve accurately. The SOC algorithm assumes a predictable voltage-to-capacity relationship, which works reasonably well for lead-acid but is essentially useless for lithium across 20–80% SOC where voltage barely moves.

What you actually need is coulomb counting done properly — tracking every amp-in and amp-out with high precision. The £12 units use low-resolution ADCs that accumulate significant counting errors over time.

I run a Victron SmartShunt 500A on my garden office battery bank and the difference versus the cheap unit I trialled first was immediately obvious. The Peukert exponent and charge efficiency factor settings matter enormously with LiFePO4.

Also worth checking: is your shunt rated for your actual peak current draw? Undersized shunts read incorrectly under load.

Great thread, this comes up a lot! One thing worth trying @Gibbo is making sure your shunt is properly synchronising its 100% reading. With LiFePO4 you really need a proper full charge cycle to complete — ideally with a genuine absorption phase at around 3.45-3.65V per cell — before the monitor will reset its counter accurately. Many cheap units also have a default self-discharge rate baked in that's calibrated for lead-acid, so dig into the settings menu and set that to zero or as low as possible. Also double-check your actual battery capacity matches what the manufacturer states — some budget cells are optimistically labelled! None of this fully solves what @PennineVanLifer mentioned about the flat voltage curve, but it can significantly tighten up the drift you're probably seeing between reality and the display.

Great point from @LizWalker about synchronisation. One thing I'd add specifically for LiFePO4 is that these cheap shunts often have quite poor coulomb counting accuracy — even small current measurement errors accumulate over time and your displayed SOC drifts further and further from reality. The fix that worked for me was setting a realistic "full charge" synchronisation trigger, so every time the battery hits absorption voltage and tail current drops off, the shunt resets to 100%. That way any drift gets corrected regularly rather than compounding indefinitely. Also worth double-checking your shunt's actual resistance matches what it's calibrated for — some of the cheaper ones are slightly off from the factory. A basic multimeter check against a known load can reveal surprising discrepancies. Not a perfect solution but it'll get you much closer to reliable readings without spending big on a Victron BMV.

Been through exactly this on the narrowboat. What finally made sense of it all for me was accepting that cheap shunts drift over time — tiny measurement errors compound across every charge/discharge cycle until the reading is basically fiction.

Ended up fitting a Victron BMV-712 and the difference was night and day. The self-discharge compensation and proper synchronisation logic just works without fiddling.

That said, if budget's the constraint, @WonkyDrifter and @LizWalker have covered the synchronisation angle well. One thing nobody's mentioned yet — check the shunt's rated current vs your actual peak draw. If you're occasionally pulling close to or over that 100A rating, the shunt heats up and resistance changes, throwing readings out further. A 100A shunt on a 200Ah LiFePO4 can get stressed quickly with inverter loads.

What @Jonno said rings true from my narrowboat days. The real killer with LiFePO4 specifically is that flat voltage curve — the battery sits at roughly the same voltage from about 20% to 80% SOC, so if your shunt drifts even slightly, there's nothing for it to "correct against" the way lead-acid monitors can cross-reference voltage. A £12 shunt simply doesn't have the precision components to track coulombs accurately over time.

Spent two frustrating seasons chasing phantom SOC readings before I caved and fitted a Victron BMV-712. Night and day difference. Yes it's £90-odd, but on a LiFePO4 bank you're essentially flying blind without something that can handle that flat discharge curve properly. Fogstar batteries even have a recommended pairing with the BMV in their setup guides for good reason.

Really good thread this. To add something slightly different - have you checked what current the shunt is actually seeing at rest, @Gibbo? These cheap units often have a small offset error that's almost invisible with lead acid but compounds badly over time with lithium. Even 0.1A phantom draw that the shunt misreads adds up to 2.4Ah per day of drift. Worth disconnecting everything except the shunt itself and checking it reads dead zero. If it doesn't, some have a zero-calibration function buried in the menus - took me ages to find it on mine. Also double-check your actual battery capacity matches what you've set; many budget 200Ah cells are genuinely closer to 180Ah when tested properly, which throws your percentage right out from the start.