Anyone else struggling to get accurate SOC readings from a cheap BMV-style shunt monitor?

Picked up a no-name 500A battery monitor shunt off eBay a few months back — paid about £18 for it — and I've been tearing my hair out trying to get the state of charge readings to make any sense. I've got a 200Ah 12V LiFePO4 bank (two 100Ah cells in parallel) and the monitor seems to think it's full when I'm clearly only at around 80%, then it dives off a cliff when I get down to the lower end. Set the capacity to 200Ah and the Peukert exponent to 1.05 as recommended for lithium, but it still drifts badly after a few charge/discharge cycles.

My main suspicion is the shunt itself isn't accurate — I've seen mentions of cheap units having resistance values that are off spec, which would throw the amp-hour counting out from the start. Has anyone actually measured their shunt with a proper meter to check? I don't have a good enough multimeter to measure milliohm-level resistance accurately, so I'm not sure how to verify it.

Also wondering if the synchronisation/reset point is the issue. I've got it set to reset to 100% at 14.2V, but with lithium that flat voltage curve I'm never quite sure if that's even firing correctly. Would love to hear how others have configured theirs, especially if you've gone down the cheap shunt rabbit hole before eventually splashing out on a Victron BMV-712 or similar.

SarahLamb56 | 847 posts

@Ben1973 Oh, I feel your pain! I had almost identical frustrations with a cheap shunt before I finally got mine behaving reasonably well.

The biggest thing that sorted mine out was nailing the battery capacity setting — most people enter their nameplate Ah rating, but you really want to derate it based on your actual battery type and typical discharge rate. I dropped mine by about 15% and suddenly the readings made far more sense.

Also, are you doing proper synchronisation cycles? The monitor needs to see a genuine full charge to reset its 100% reference point. If your charge cycle isn't actually reaching absorption properly, it'll drift progressively worse over time.

What battery chemistry are you running — lithium or lead-acid? Makes a big difference to how you'd approach calibrating it. Happy to share more specifics once I know what you're working with! 🔋

Wez1993 | 312 posts

@Ben1973 One thing nobody's mentioned yet — have you set your battery capacity correctly in the monitor settings, and more importantly, have you accounted for Peukert's exponent? Most cheap units default to 1.25 which is way off for lithium if that's what you're running. Also worth checking your shunt is actually installed on the negative side only, with absolutely everything running through it. I found a single earth return bypassing my shunt was throwing my readings out by 20% consistently. The £18 units can actually work reasonably well once properly configured — they're just not plug-and-play like a Victron. Give it a proper full charge cycle after sorting the settings and let it recalibrate from a known 100% SOC. That usually sorts things out.

DailyAdventure | 1,204 posts

@Ben1973 Mate, £18 is roughly the cost of a Victron BMV-712's packaging — your wallet made a choice and your sanity is paying the consequences.

PylontechFan | 2,156 posts

@Ben1973 One thing that's often overlooked with these cheapo shunts is the synchronisation/reset threshold — basically the voltage point at which the monitor decides "right, that's 100% charged." If it's set even slightly off for your battery chemistry, the drift compounds over every cycle and you'll be miles out within a fortnight. Worth digging into the settings menu if it has one. Also, make sure your shunt is genuinely the only path for current — any rogue earth connections bypassing it will absolutely murder your accuracy. A loose chassis ground is a classic culprit. None of this fixes the underlying quality issue mind you, but it might get you closer to usable readings while you save up for something decent! 😄

PaulMurray | 47 posts

Had almost identical grief with a cheap shunt on my tiny house build last year. The thing that finally sorted mine was making sure every single load and charge source was routed through the shunt — I had my Renogy MPPT going directly to the battery, bypassing it entirely, so the monitor never had a clue what was actually going in.

Worth double-checking your wiring diagram against what's actually installed. The shunt needs to sit between the battery negative and everything else — no exceptions.

Also, are you on lithium or lead acid? The default charge/discharge parameters on these cheapo units are usually set for lead, which throws the SOC calculation completely off for LiFePO4 if you haven't manually corrected the Peukert exponent and charge efficiency settings.

CrispyBuilder | 634 posts

@Ben1973 Something worth checking that nobody's mentioned yet — the current sense resistor on those cheap units is often wildly inaccurate from the factory. Grab a decent clamp meter and compare what the shunt is reporting against an independent reading. I found mine was out by nearly 8% which absolutely destroys SOC accuracy over time as the errors compound.

Also worth checking your battery's actual capacity versus what you've programmed in. If you're running used or reconditioned cells, the real usable capacity could be significantly lower than the nameplate figure. Programme in a conservative number and you'll get far more reliable readings day-to-day.

Neither fix costs a penny, so worth ruling those out before spending more money. If it still plays up after that, honestly @DailyAdventure isn't entirely wrong about the Victron — the 712 transforms battery monitoring.

HamishLee | 312 posts

@Ben1973 One thing I'd add to what the others have said — these cheap units often have absolutely terrible temperature compensation, or none whatsoever. If your batteries are in an unheated space (shed, garage, etc.), the internal resistance changes enough that your shunt is essentially lying to you about what's going in and coming out. Also worth double-checking your battery capacity setting matches actual usable capacity rather than the headline figure. I had mine set to nominal and was chasing my tail for weeks before I twigged. Sometimes the software side is the culprit rather than the hardware itself.