JK BMS firmware 19.13b Battery Capacity Failure - Solved

Had this exact issue crop up on my van build last month — JK BMS showing wildly incorrect capacity readings after updating to 19.13b. Frustrating doesn't cover it when you're relying on accurate SoC for EV charging top-ups on the road.

After a fair bit of head-scratching, the fix that worked for me was doing a full capacity reset via the JK app before powering the system down, then letting the pack do a complete charge/discharge cycle to recalibrate. The BMS seems to need that reference point established under the new firmware or it just throws nonsense figures.

Worth noting — I'm running a Fogstar Drift 206Ah cells in a 4S configuration, so YMMV with different cell chemistry. My Victron Cerbo was reading SoC through the JK's RS485 integration, which made the wrong capacity values even more obvious since the two were constantly disagreeing.

A couple of questions for anyone else who's been through this:

• Did the fix hold long-term, or did you see drift creeping back in after a few weeks?
• Has anyone rolled back to 19.11 successfully without losing their cell balance data? I'm tempted to stay on older firmware until 19.13b gets a proper patch

Also curious whether this affects the JK PB series differently to the standard inverter BMS units — I've seen conflicting reports on the Facebook groups but those threads turn into a mess quickly.

Solid thread, good to know it's a known issue with a workaround at least. Would be nice if JK pushed a hotfix rather than leaving users to figure it out themselves.

@RussGreen had this on my tiny house setup a few months back, different firmware but same symptoms — capacity readings all over the shop.

Worth checking if your cell voltage calibration got reset by the update. Mine had defaulted back to factory values which completely threw the SoC calculation. Took me ages to realise that was the culprit.

Also — are you pairing it with a Victron MPPT? Found that the JK's reported capacity can conflict with what the Cerbo GX expects if the Peukert settings don't match post-update.

Quick fix that worked for me was doing a full charge/discharge cycle after manually re-entering the cell specs. Cleared the daft readings within a couple of cycles on my Fogstar cells.

What capacity are your cells? Might narrow it down.

@RussGreen this is a known regression in 19.13b — the cell count calibration routine got broken and the BMS essentially forgets to factor in your actual cell configuration when calculating pack capacity.

Fix is straightforward:

1. Power cycle the BMS completely (disconnect load and charge sources)
2. In the JK app, navigate to Battery Parameters → Rated Capacity and re-enter your value manually
3. Force a full charge to absorption then let it rest 30 mins
4. Trigger a capacity calibration cycle from the app

The BMS needs that full charge reference point to anchor the coulomb counting properly. Without it, the corrected figure just drifts again within a few days.

Worth noting — 19.14 appears to resolve this upstream, though I'd wait another week before updating given JK's track record with rapid releases. @PanelMatt's workaround likely bought time rather than permanently fixed the underlying calibration offset.

@DevonDweller hit the nail on the head about the calibration routine.

Worth adding — I noticed the same regression on my boat setup after a reluctant firmware update, and the fix that actually stuck for me was forcing a full capacity relearn cycle before touching any of the cell count parameters. JK's own documentation completely glosses over the order of operations here, which is maddening.

Practical steps that worked:

1. Discharge to BMS low-voltage cutoff
2. Full charge to absorption finish
3. Then re-enter your cell count and nominal capacity
4. Save and hard-restart the BMS

Doing step 3 first was causing the BMS to overwrite the freshly calibrated figures with garbage defaults. Took me an embarrassing number of attempts to work that out on the water with no mobile signal to search for answers.

@DevonDweller good to know it's the calibration routine — I had something similar crop up on my shepherd's hut build last winter, though on an older firmware.

One thing worth checking that nobody's mentioned yet — after you re-run the cell count calibration, make sure you do a full charge cycle to 100% and let it sit at absorption for a good while before trusting the SoC figures again. The BMS needs that reference point to anchor everything properly.

Also, if you're using a Victron MPPT alongside, cross-reference the SoC with the Victron app as a sanity check whilst the JK settles in. Saved me a lot of head-scratching.

@RussGreen rolled back to 19.11 on my Fogstar Drift setup and the capacity readings came back like nothing ever happened — 19.13b's a dumpster fire, avoid until JK actually fix it properly.