Three-Day Flat Battery on a Lithium-Equipped Adventure Bike
A 2020 850 cc adventure bike with 19,500 km arrives on a trailer: the engine will not crank after three days parked in a garage. The owner has fitted heated grips, a USB socket and a GPS tracker, the last wired straight to the battery positive with an inline 2 A fuse. The battery is a 12.8 V, 8 Ah LiFePO4 with a built-in BMS; rested voltage is 13.1 V and it starts the bike readily after a 20-minute charge. Charging voltage at the battery terminals is 13.9 V at 3,000 rpm with the lights on, against a factory spec of 14.0-15.0 V. With a datalogging meter in series, key-off draw settles at 38 mA after the 12-minute CAN sleep, but spikes to 850 mA for 40 seconds every 6 minutes. A load test of the battery at 0.5 C shows it holds 5.1 Ah before the BMS disconnects. The bike is used daily in summer and stored outdoors at -5 °C in winter.
1. Quantify the parasitic draw and show whether it explains a flat battery after three days. State what a 38 mA reading alone would have predicted and why the datalogging meter changed the conclusion.
Read the worked answer
The datalogging meter is the critical instrument here, because the steady-state figure is not the whole load.
Average the spikes: 0.85 A for 40 s in every 360 s cycle contributes 0.85 × 40/360 = 0.094 A, or 94 mA. Add the 38 mA baseline and the true average key-off draw is roughly 132 mA, about 3.5 times what a single steady reading suggests.
Now the battery. Rated capacity is 8 Ah, but the load test shows only 5.1 Ah before the BMS disconnects, so 5.1 Ah is the usable figure to work with, not 8 Ah. Time to disconnect = 5.1 / 0.132 = 38.6 hours, roughly 1.6 days. Three days is comfortably explained, and in the cold the margin is even worse.
A steady 38 mA reading alone would have predicted 5.1 / 0.038 = 134 hours, or 5.6 days, which looks almost acceptable and explains why the previous shop found nothing wrong. It is still out of spec, most manufacturers specifying under about 3 mA after sleep, but on its own it does not explain the complaint. The datalogging meter converted a pass into a fail.
The lesson: on any modern machine with always-on modules, a single steady-state current reading is not a valid test.
2. Describe the isolation procedure, the safe technique for a CAN-equipped bike, and how you would identify the responsible circuit.
Read the worked answer
Technique. Never break and remake the battery negative with a meter in series on a CAN bike unless you accept waking every module and re-timing the sleep cycle. Use a clamp meter where resolution allows, or a series meter with a bypass switch so the circuit is never interrupted. Note the 12-minute sleep; every door, key or scan-tool event resets it.
Procedure.
- Fully charge the battery, then park the bike as the owner does with nothing touched, and let it sleep for 15 minutes.
- Datalog for at least 30 minutes so periodic wake-ups are captured, not just the baseline.
- Pull fuses and relays one at a time, re-datalogging after each, and record which removal eliminates the spikes.
- Check the tracker separately: it is wired direct to the positive with its own fuse, so it is outside the fuse box and will not appear in a fuse-pull sweep. Disconnect it at the battery and re-measure.
- Check for a stuck relay on the heated grip circuit, which would show as a constant load rather than a periodic one, and check that the USB socket is switched.
Interpretation. A periodic 850 mA spike every 6 minutes is the signature of a module waking on a timer, typically a tracker or alarm with a keep-alive or a cellular heartbeat, rather than a shorted diode or a lamp left on. A constant current with a fuse pulled identifies a discrete circuit; a spike that survives every fuse pull but disappears when the tracker lead is lifted identifies the tracker.
Document the baseline, each removal, and the resulting current.
3. Assess the charging system and the lithium-specific risks, then give your recommendation before the bike is released.
Read the worked answer
Charging. 13.9 V is below the 14.0-15.0 V specification, but before condemning the regulator, verify at the battery posts rather than the dash and run a voltage-drop test on the positive and ground paths. Then measure AC ripple at the battery with the engine at 3,000 rpm: anything above roughly 0.5 V AC points to a failing rectifier rather than a low set point. A stable 13.9 V with clean ripple is a regulator marginally out of spec, not a charging failure.
Lithium specifics.
- 13.9 V will not fully charge a LiFePO4 pack and will never bring the cells into balance. The pack will sit at perhaps 80-90 % state of charge, which is acceptable for cranking but leaves no reserve for a parasitic draw plus a cold start.
- Charging below 0 °C damages LiFePO4 cells, which is why good packs have a low-temperature charge cut-off in the BMS. A bike stored outdoors at -5 °C may simply refuse to accept charge until it warms.
- The BMS low-voltage disconnect is silent: the bike will not even light the dash, with no slow-crank warning. The owner reads this as a failed battery.
- 8 Ah nominal is fine for cranking, because lithium delivers high pulse current, but it is a small reservoir against a 132 mA average drain.
Recommendation. Fix the tracker or move it to a switched supply, correct the draw to under about 3 mA, then re-test the charging system at the battery. Release the bike only with a logged sleep current under spec and a lithium-compatible maintainer supplied to the owner for winter storage, with instructions not to charge below freezing.