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

Ran into exactly this situation last winter with my garden office setup. Short answer: trust neither blindly, but the BMS wins for protection decisions.

Here's my reasoning. The SmartShunt is a coulomb counter — it's tracking charge in and out with reasonable accuracy, but it drifts over time if your peukert exponent isn't dialled in correctly, or if you're not hitting a proper absorption cycle regularly enough to reset the 100% reference point. Mine was reading 6-8% optimistic after a run of overcast days.

The BMS, on the other hand, is measuring cell voltages directly. It doesn't care about accumulated counting errors. What it does struggle with is the flat voltage curve you get with LiFePO4 — a small voltage difference in the middle of the curve represents a huge SOC swing, so its readings can look jumpy.

My current approach:

• SmartShunt handles day-to-day monitoring and feeds MQTT data into my Home Assistant dashboard
• BMS cell voltage readings are what actually trigger any load shedding or charge limiting
• I let a full absorption cycle run every couple of weeks specifically to resync the shunt

Worth checking a few things in Victron's VictronConnect app — make sure your battery capacity figure is accurate, and that the charged voltage threshold matches what your BMS considers full. A lot of discrepancy comes down to misconfigured parameters rather than either device actually being wrong.

What BMS are you running? JK, Daly, something else? Makes a difference in terms of how accurate the voltage-based SOC estimation actually is.

@CarlBaker makes a solid point about protection decisions sitting with the BMS.

Worth adding though — the SmartShunt's SOC accuracy lives or dies by how well you've calibrated it. Charged voltage, tail current, and Peukert exponent all need to match your actual battery chemistry precisely.

On my motorhome setup with Fogstar Drift lithium cells, I initially saw a 12% discrepancy. Turned out my tail current was set too high, so the shunt was calling 100% SOC prematurely.

Once dialled in properly, the SmartShunt is actually better for state-of-charge trending over time — the BMS typically uses voltage-based estimation which gets genuinely unreliable mid-discharge on LiFePO4's flat discharge curve.

So practically: BMS = hard limits and cell protection, SmartShunt = accurate capacity tracking for day-to-day planning. Run both, cross-reference regularly, and investigate any divergence beyond ~5%.

My narrowboat has both disagreeing with each other and with reality on a regular basis — at this point I just check the kettle still boils.

Good thread. One thing worth adding to @CarlBaker's point — the accuracy of both devices degrades over time without proper calibration.

My SmartShunt drifts noticeably if the battery never hits a genuine 100% absorption finish. The shunt can't self-correct without a proper full charge cycle to reset its reference point. I've seen mine read 94% when the BMS was screaming high-voltage cutoff.

Practical fix on my LiFePO4 setup:

• Let the bank hit full absorption occasionally (not just float)
• Set your SmartShunt's charged voltage threshold accurately — most people leave it at default
• Sync them once a month manually if needed

The BMS wins for protection, absolutely. But treat the SmartShunt as your fuel gauge, not your protection layer — and calibrate it properly or it'll lie to you consistently rather than occasionally.

@T6Solar raises something I've not seen mentioned yet — the synchronisation point matters enormously for the SmartShunt specifically.

Mine was consistently reading 8-12% higher than my BMS until I realised the "charged voltage" threshold in VictronConnect was set slightly below where my battery actually hits absorption. It was never triggering a proper sync, so coulomb-counting errors were compounding over weeks.

Worth checking three settings:

• Charged voltage — must actually be reached during a full charge
• Tail current — set too high and it'll sync prematurely
• Charged detection time — default 3 minutes is often too short for lithium

Once I corrected those, my SmartShunt and Fogstar Drive BMS were within 2% of each other consistently. The disagreement wasn't a hardware problem at all — just misconfiguration. Check those before assuming one device is faulty.