Victron SmartShunt showing different SOC to my BMS – which one do I trust?

Had this exact head-scratcher with my shepherd's hut build last winter. Two devices, two wildly different numbers — maddening.

Here's what I eventually landed on: trust the BMS for protection, trust the SmartShunt for trending.

The BMS knows the cell-level truth. It'll pull the plug if something goes dangerously low or high, full stop. But SOC percentage on most budget BMS units is frankly guesswork — they're often just reading voltage and calling it a day.

The SmartShunt does proper coulomb counting, so it's far more accurate over time, but it needs a solid reference point to anchor itself. If you haven't let it sync at 100% recently (full absorption, tail current dropping off cleanly), it'll drift. Mine was showing 67% when the BMS said 84% — turned out I hadn't had a proper full charge since October because the Scots weather had been absolutely dire.

Few things worth checking:

• Peukert exponent — is it set correctly for your chemistry? LiFePO4 should be close to 1.05
• Charged voltage threshold — must match what your charger actually hits
• Tail current setting — too high and it'll think it's full when it isn't

Once I dialled those in via the Victron Connect app, both devices settled within 3-4% of each other, which I'd call acceptable.

What BMS are you running? If it's one of the cheaper JK or Daly units, that'll change the conversation somewhat. And are you on Fogstar or a DIY pack? Curious whether others have found the same pattern.

@MPPT_Wizard had similar with my setup. The SmartShunt drifts if it never hits a proper full charge to reset — coulomb counting accumulates errors over time. BMS is reacting to cell voltages in real time, which is more honest at the extremes.

What sorted mine was making sure the SmartShunt charged efficiency factor and Peukert exponent were dialled in properly for my Fogstar cells. Default values are generic and rubbish for LiFePO4.

Also check your tail current setting — if it's not seeing a genuine absorption end, it'll never sync to 100%.

Once I got those right, both readings stayed within 3-4% of each other most of the time. Not perfect, but close enough.

@MarineGaz nailed the drift issue, but the bit nobody mentions: your SmartShunt's "charged voltage" threshold in the settings needs to match what your BMS actually allows the cells to reach — if your BMS tops out at 54.8V but the SmartShunt is waiting for 55.2V to call it 100%, it'll never sync properly and just wanders off into fantasy SOC territory like a lost sheep on Dartmoor. 🐑

Been through this exact farce with my Fogstar Drift 100Ah in the shepherd's hut — once I aligned the thresholds and set a realistic tail current (1-2% of capacity), the SmartShunt stopped lying to me.

@MPPT_Wizard @CamperCarl good points all round. One thing worth adding — your SmartShunt's tail current setting is just as critical as the charged voltage threshold. If it's set too high, the shunt thinks the battery's full before it actually is, and your SOC never properly synchronises. I had mine at the default 4% for ages wondering why it kept drifting optimistic. Dropped it to 1% on my LiFePO4 bank and the two devices agreed much more consistently. Worth cross-referencing whatever your BMS reports as the actual absorption tail with what you've got configured in the VictronConnect app.

@GrumpySkipper all valid, but there's another layer nobody's touched on yet — battery capacity accuracy.

Your SmartShunt is doing coulomb counting against whatever capacity figure you manually entered during setup. If you told it 100Ah and your cells are actually delivering 94Ah at your typical discharge rate (Peukert effects, temperature, age), the percentage figures will diverge regardless of how perfect your synchronisation logic is.

On my boat I cross-referenced against an actual measured capacity test before I trusted either device. Surprised myself — cells were notably below nameplate spec.

Worth also flagging: if you've got a LiFePO4 pack, the flat voltage curve makes voltage-based SOC estimation genuinely unreliable in the middle ranges, which is partly why the BMS and SmartShunt disagree — they're using fundamentally different methods, and neither is infallible mid-cycle.