Anyone else struggled to get accurate SoC readings with a budget shunt on LiFePO4?

Picked up a cheap 500A shunt off Amazon a few weeks back — one of those no-name jobs that came with its own little LCD display. Wired it up on my 200Ah LiFePO4 bank (2 x 100Ah Lishen cells in parallel) and I've been getting some properly confusing readings. It'll show 80% state of charge, I'll run the kettle for ten minutes and suddenly it's jumping down to 61%. Doesn't seem right at all.

I've set the battery capacity to 200Ah in the settings and double-checked the shunt is on the negative side between the battery and the busbar. The wiring looks tidy enough — 6mm² cable, decent crimps. I did wonder if the issue is with how LiFePO4 voltage curves are so flat compared to lead acid, and whether these budget units just aren't calibrated properly for lithium chemistry.

Has anyone had similar grief with cheap shunts on lithium? I'm debating whether to just bite the bullet and grab a Victron BMV-712 but that feels like a lot of money to spend when I'm not 100% sure that's even the problem. Is there a middle-ground option, or am I missing something obvious in my setup?

LauraCole | 47 posts

@RussStevens Oh yes, very familiar with this! The main culprit with those cheap shunts is usually the shunt resistance tolerance — they're often rated at 500A but the actual resistance value is slightly off, which throws your Ah counting out from the start. Also worth checking whether it's accounting for the LiFePO4 charge efficiency factor (around 99% typically) — many budget units just assume lead-acid defaults.

One thing that genuinely helped me was doing a proper full charge to 100%, letting it rest for 30 mins, then resetting the SOC there. LiFePO4's flat voltage curve means voltage-based correction barely works mid-cycle, so you're almost entirely reliant on accurate coulomb counting between resets.

What's your discharge current typically like? High sustained loads can expose calibration errors faster.

@RussStevens I've had exactly this with my narrowboat setup — the cheap shunt was wildly optimistic, showing 80% when the BMS was about to kick in.

What's your shunt actually calibrated to? A lot of the no-name ones ship with incorrect shunt resistance values in the firmware, so the current measurement is off from the start. Also wondering whether it's accounting for the flat discharge curve on LiFePO4 properly — most budget units seem designed around lead-acid voltage profiles.

Did you set the battery capacity correctly in the display settings? Some of those little LCD units default to 100Ah regardless of what you enter.

Ended up swapping mine for a Victron BMV-712 and the difference was night and day — proper Peukert compensation and Bluetooth monitoring too. Is that kind of spend realistic for your setup, or are you committed to making the budget option work?

DazHughes | 203 posts

@RussStevens Worth checking whether your shunt is actually rated for continuous current at that 500A figure, or whether that's a peak rating — a lot of the cheap ones are, and they'll run warm under sustained load which throws the resistance off and skews your readings.

Also, LiFePO4's flat discharge curve genuinely makes coulomb counting almost mandatory for decent SoC accuracy — voltage-based readings are nearly useless across most of the range. If the unit is primarily voltage-referenced rather than properly counting amp-hours in and out, you'll get rubbish results regardless of shunt quality.

Personally I ditched a similar setup and moved to a Victron SmartShunt — not exactly budget, but the difference in accuracy was night and day. Might be worth considering if the frustration keeps up.

PaddyDixon | 12 posts

Running a similar setup on my narrowboat and had the same headache. One thing nobody's mentioned yet — have you verified the shunt's actual resistance value with a decent multimeter? Some of the no-name units are stamped 500A but the resistor is spec'd completely wrong, so the display is calculating SoC from duff data before you've even started.

Also curious whether your BMS is reporting anything different to the shunt display? On mine there was a noticeable discrepancy which helped me figure out where the fault actually sat.

Ended up going with a Victron SmartShunt in the end — pricey compared to the Amazon jobs but the Bluetooth integration and proper Peukert compensation for LiFePO4 made it worth it on a liveaboard where accurate SoC actually matters for planning.