Behavior at very low State of Charge

Been thinking about this after noticing some odd behaviour with my shepherd's hut setup lately, so curious what others have experienced.

Running a fairly modest system — Victron SmartSolar, 200Ah Fogstar Drift LiFePO4, and a JK BMS. I've intentionally left my minimum SOC quite low (around 5%) because the hut is only used occasionally and I'd rather squeeze every usable amp-hour out of the bank when I'm there.

What I'm noticing is that behaviour gets a bit unpredictable as the battery approaches that floor. The JK BMS occasionally throws a low-voltage protection event before the Victron even registers we're near the limit. They don't always seem to be "talking" to each other cleanly at the tail end of discharge.

A few questions for the group:

• Is there a recommended minimum SOC floor for LiFePO4 that balances longevity vs usability? I've seen everything from 10% to 20% suggested.
• For those using a JK BMS alongside Victron kit — are you using the BMS assistant in VenusOS, or relying purely on BMS hardware protection?
• Does the flat discharge curve of LiFePO4 make SOC readings genuinely unreliable at low charge, or is that mostly a calibration issue?

Suspect some of this is just the nature of LiFePO4's voltage curve making SOC estimation a bit of a dark art near empty. But keen to hear whether anyone's found a reliable way to manage the low-SOC handoff between BMS and inverter/charger without the system getting twitchy.

@Jock30 what BMS are you running? That's usually where the odd behaviour comes from at low SoC rather than the Victron kit itself.

My van build taught me the hard way — had a cheap BMS that would randomly disconnect under load when it hit around 10-15% and the Victron MPPT would then throw a fault because it lost its battery reference voltage. Took me ages to figure out it wasn't a wiring issue.

What counts as "odd behaviour" in your case? Voltage dropping suddenly, BMS tripping, or is the SmartSolar doing something unexpected?

Also worth checking — are you using the Victron battery presets or a custom charge profile? Some of the default LiFePO4 settings are a bit conservative and can make the SoC readings look weird near the bottom end.

@GoldenMechanic makes a fair point about the BMS, but also worth checking your battery voltage under load at that low SoC. LiFePO4 has a pretty flat discharge curve right up until it falls off a cliff — so what looks like 20% SoC can actually be nearly empty once you factor in voltage sag under any meaningful draw.

With 200Ah you haven't got masses of headroom to begin with. I'd suggest logging your battery voltage overnight (Victron Connect makes this dead easy) and see if you're getting sudden dips when something kicks in — a fridge compressor starting up can be enough to confuse a BMS into thinking it's hitting undervoltage protection.

What's your low voltage disconnect set to on the BMS? That'd help narrow it down considerably.

@Jock30 had something similar with my cabin setup last winter. Below about 10% my Fogstar 100Ah would get really twitchy — voltage would sag under even modest loads and the BMS kept doing these brief disconnects.

Turned out my low voltage cutoff on the SmartSolar was set slightly too aggressive for the actual chemistry. Worth checking your DVCC settings in VictronConnect — if the charger and BMS aren't talking properly you get that weird hunting behaviour.

Also, what's your load like at the point it acts up? Resistive vs inverter loads behave very differently at low SoC.

Hey @Jock30, interesting one! Just to add to what @GlenWard mentioned — I've noticed with my Fogstar Drift 200Ah that behaviour below 10% can vary quite a bit depending on temperature too. Cold mornings here in Scotland made mine act really strangely last winter, almost like the BMS was seeing a lower voltage than it actually was. Worth checking if the odd behaviour coincides with colder periods? Also, are you seeing this under load or when the system's relatively idle? Makes a difference in terms of whether you're looking at a genuine low SoC cutoff situation or something more like voltage sag causing a false alarm. I'd personally try to keep it above 20% as a regular floor anyway — saves a lot of headaches and is kinder to the cells long-term! 🙂