Anyone else had issues with JK BMS dropping cells below 2.8V before cutting off?

Been running a 280Ah LiFePO4 pack (8 cells in series, 24V) in my shed setup for about six months now. Using a JK BMS — the 200A active balancer version — and overall I've been pretty happy with it. Balancing is decent and the app is straightforward enough once you get your head round it.

The issue I keep hitting is that during heavy discharge (running a 2kW inverter with a kettle or similar), one rogue cell is dipping to about 2.75V before the BMS actually trips the under-voltage protection. I've got the cell UVP set to 2.8V and the protection release at 3.0V, but there's clearly some lag or hysteresis happening that's letting it creep below the threshold before the contactor opens.

I've read a few threads suggesting the "protection delay" setting is the culprit — mine's currently at 500ms — but dropping it lower makes me nervous about nuisance trips under transient loads. Anyone found a sensible middle ground? Also wondering whether this one cell is just a weaker one that needs replacing, or if it's worth trying to top-balance the whole pack again first before pointing the finger.

@NeilSmith interesting one — I've got a JK BMS on my boat (100A passive version) and I had something similar early on. Turned out my cell undervoltage protection was set correctly but the cell undervoltage protection delay was too long, giving weak cells time to sag further before the cutoff actually triggered.

Worth checking in the JK app:

• What's your protection delay set to?
• Are the cells actually weak/imbalanced, or is it purely a BMS response lag issue?

Mine was the delay — dropped it from 5s to 1s and the problem stopped. Though on a boat with big inverter loads the voltage sag can be dramatic, so I also had to reconsider whether my protection voltage threshold was actually realistic under load vs at rest.

What loads are you running when it happens?

@NeilSmith the default undervoltage protection on JK units is often set quite conservatively from the factory — worth checking your cell undervoltage cutoff isn't sitting lower than it should be. For LiFePO4 I'd set protection at 2.9V and recovery at 3.0V, not the 2.7V/2.8V defaults some firmware ships with.

Also check your undervoltage delay setting. If it's set to 0 seconds, transient voltage sag under load can trigger the cutoff prematurely — but equally, if it's too long, cells genuinely do get pulled low before the BMS reacts.

On my garden office pack (16S, Victron SmartShunt monitoring) I run 2.95V protection and haven't had a single cell touch 2.8V in 18 months. The JK app makes changing these straightforward enough.

Hey @NeilSmith, worth checking whether you've got the cell undervoltage protection and the pack undervoltage protection set consistently — sometimes they can conflict and cause unexpected behaviour. Also, are you running any significant loads at the moment of cutoff? A high current draw can cause a temporary voltage sag that recovers almost immediately, which might mean the cell wasn't genuinely at 2.8V — just dipping under load.

If you haven't already, try logging the individual cell voltages over time using the JK app — the Bluetooth monitoring is decent enough to catch patterns. Mine took a bit of tweaking before the settings felt properly dialled in. What firmware version are you running? There were some known quirks in earlier versions that were sorted in later updates.

@NeilSmith one thing worth checking that nobody's mentioned yet — have you set a recovery voltage that's too close to your protection threshold? If the recovery voltage is only a few millivolts below the cutoff, the BMS can start cycling rapidly under load, which can let cells dip further than expected before it latches properly.

Also worth looking at your cell undervoltage delay setting. If it's set too short, transient sag under heavy load can trigger a disconnect before the BMS has properly assessed the true resting voltage. Conversely if it's too long, cells sit low for longer than ideal.

What loads are you running when it trips? High current draw will exaggerate sag on weaker cells considerably. Might be worth logging with a Bluetooth app to catch exactly what's happening at the moment of cutoff.