Been there with my Fogstar Drift cells and a Victron SmartShunt — the BMS was cheerfully reporting 98% whilst the shunt was essentially screaming "mate, you've got about enough juice left to boil a kettle... once."
The root issue is that most BMS percentage readouts are voltage-based estimates, whereas a proper Coulomb-counting shunt (like the SmartShunt or a BMV-712) is actually tracking electrons in and out. When the two disagree wildly, trust the shunt — the BMS is basically just eyeballing it.
Classic scenario where this bites you:
- BMS says 100% ✅
- Shunt says 31% 😬
- Multiplus wakes up on shore power and absolutely launches 1,400W at your bank trying to top up what it correctly knows is missing
- You panic, post on forums at midnight (we've all been there)
The fix that sorted mine was properly synchronising the shunt — make sure it's set to register a full charge only when the battery hits absorption voltage and tail current drops appropriately. Otherwise it just... drifts. Hence the name, presumably.
Also worth checking your charge efficiency factor in VRM if you're Victron-based — lithium should be around 99%, not the 95% default which is more AGM territory.
Anyone else running a mix of BMS Bluetooth readouts alongside VRM and finding them permanently at odds? Curious whether this is a Victron/Fogstar quirk or if folks with Pylontech stacks or Liontron batteries see the same thing — particularly those of you with garden offices drawing overnight loads where the drift