Posted on

Why an E-bike BMS cannot accurately measure battery capacity

Better BMS with warranty handling support

Electric Vehicle (EV) batteries can determine their capacity accurate enough to rely on for warranty purposes. EV’s have a built-in battery analyzer that also reads and interprets BMS data, which is vital for warranty evaluation. This article explains which limitations an E-bike BMS has, how to deal with this and how to -theoretically- resolve this.

Accuracy of components and calibration

A good battery analyzer uses highly accurate voltage- and current sensing resistors with a very low Tc (temperature coefficient). These resistors are expensive, relative to the cost of an e-bike battery BMS. Good battery analyzers were calibrated individually, which has a significant impact on system accuracy. Performing this would easily double the cost of the e-bike BMS. However, that is just the beginning…

Coulomb counting

Both battery analyzers and BMS’s use the same method to measure the discharge capacity: Multiplying voltage, current and time yields the capacity (in Wh). In a workshop, where the battery analyzer is used, the ambient temperature is relatively constant, on an e-bike it obviously changes a lot. This influences battery temperature and therewith the battery capacity. As significant as the resulting inaccuracy is, the worst still has to come…
A battery analyzer always charges to 100% and discharges to 0%. With this procedure, the battery capacity can be accurately determined. The battery also charges to 100% during e-bike operation, but the driver rarely discharges to 0%. This is because they do not wish to run without power or because they realize that regular discharging to 0% is bad for battery lifetime. The absolute value can only be determined when discharging to 0%.

Relying on BMS data

When a BMS reports an actual capacity in Wh rather than as a % of the last known current capacity, the accuracy depends greatly on when the battery was last discharged. When a consumer offers an e-bike for trade-in to a dealer, the dealer has no idea of the date of last full discharge. As dealers learn by experience, they do not longer accept trade-ins without having a battery analyzer measure the real current capacity, in addition to the internal resistance, which is a measure of the health of the battery.

Improving BMS data accuracy

To bring the accuracy of e-bike BMS data to an acceptable (+/- 5-10%) level, these measures are the solution:

  • Use precision voltage and current sensing resistors with a low temperature coefficient
  • Characterize the battery as part of the QC procedure: measure the voltages at different loads and temperatures over the full SoC (state-of-charge) range. Then, the BMS can more accurately determine the SoC based upon measuring the voltage, current and temperature, provided a recent full discharge was performed and the date hereof and the number of cycles thereafter are recorded. This is the procedure used with high end EV (electric vehicle). For Akkus of 10,000 Euro and higher, this time consuming (several days) characterization is worth the cost, but for e-bikes it is not cost-efficient at all.

Best practice for e-bike BMS

When BMS designers would register the last known capacity, measured by a battery analyzer as well as the date and cycles thereafter, this would allow the dealer (and the consumer) a better insight in actual battery capacity. The dealer can then do a quick test to measure the internal resistance. The combination of capacity and health (signified by the internal resistance) closely simulate the performance under street conditions and therewith the (remaining) value.