Anyone else running a Victron Cerbo GX with mixed battery brands — how are you handling state of charge accuracy?

Just swapped out my old AGM bank on the narrowboat for a mix of Fogstar Drift 100Ah LiFePO4 cells (two of them) alongside a remaining 110Ah AGM I haven't replaced yet. The Cerbo is reading SoC from a SmartShunt, which is fine in theory, but because the charge profiles are so different between the two chemistries, the shunt's coulomb counting seems to drift badly after a few days. Showing 74% when the AGM is clearly gassing and the Fogstar cells are sitting pretty — proper misleading.

I've tried tweaking the charged voltage and tail current thresholds in the SmartShunt settings (currently at 13.5V and 4% respectively), but the re-sync events are unreliable because the AGM reaches absorption before the Lithium bank does. The Cerbo sees a blended voltage that doesn't cleanly represent either battery's true state.

Has anyone tackled this properly — separate shunts per bank feeding into the Cerbo via VE.Direct, or is there a smarter way to configure the DVCC settings to account for mixed chemistry? Genuinely not sure whether the "right" fix is a second SmartShunt, a full battery replacement (inevitable but not this month), or whether there's some Cerbo firmware wizardry I'm missing. Would love to hear how others have approached it before I start buying more kit.

@RustySpanner the mixed chemistry setup is going to give your Cerbo a headache — AGM and LiFePO4 have completely different voltage curves so any SoC reading from the Cerbo's voltage-based estimation will be essentially meaningless.

Honestly the only clean solution is a Victron SmartShunt on the whole bank. Coulomb counting ignores the voltage curve problem and just tracks what goes in and out. Still not perfect with mixed chemistries but dramatically better than voltage-based.

Longer term though, ditch the AGM. Even running in parallel you're forcing the LiFePO4s to charge at AGM-friendly voltages which is leaving capacity on the table. I went all-Fogstar Drift in my shepherd's hut and the Cerbo SoC accuracy improved massively once I had a single chemistry and a properly configured SmartShunt — set the charged voltage and tail current correctly and it's very reliable.

@RustySpanner been down this exact road on my boat — mixed banks are a nightmare for accurate SoC.

Honestly the only thing that sorted it for me was fitting a BMV-712 on the main negative and letting the Cerbo pull SoC from that via VE.Direct rather than calculating it internally. Shunt-based coulomb counting doesn't care what chemistry you've got, it just counts electrons in and out.

That said, your charge profile is still going to be a compromise — AGM and LiFePO4 want different absorption voltages. Worth considering whether that last AGM is worth keeping in the mix or just retiring it early. Sometimes the faff isn't worth it.

What's running your alternator side? If it's unregulated you've got another headache waiting for you once the LiFePO4 starts hitting full charge fast.