Charge / Discharge Test

Battery Charge/Discharge Testing

The ePure battery is rated at 1875 watt-hours and nominally 50.4 volts.

The discharge test was conducted at a constant 10.0A rate. This rate was easy for me to achieve and is reasonably representative of the average current draw while free-riding. Interestingly, this equates to an average power delivery of well under 1 horsepower.

The discharge took 187.75 minutes and the battery yielded 1610 watt-hours. This is obviously less than the battery's rated capacity. It's likely that withdrawing energy at a slower rate, or intermittently, would yield more energy. Conversely, withdrawing energy at a faster rate would yield less energy.

When recharging the battery, it accepted 1823 watt-hours from the DC side of the charger. Some of this energy is wasted as heat by the internal BMS performing cell balancing.

The energy required from the AC line was 1953 watt-hours. This equates to a charger efficiency of better than 93%. However, the power factor is far from unity. I measured the 5.7's charger at 0.66 (leading) with a Fluke 39. This is representative of a typical non-power-factor-corrected off-line switched-mode power supply (e.g. desktop computer).

Interestingly, it took about the same time to charge the battery (180 minutes) as it did to discharge it.

The charger initially operates in a constant-current mode (at 15.5A) until it reaches a battery voltage of about 58 volts. This period is when the lion's share of the energy gets transferred (taking about 100 minutes). The charger then operates in constant-voltage mode until the battery reaches about 58.7 volts.

After charging is complete, the charger's fan turns off and it enters maintenance mode, delivering about 0.83A (50W) to the battery. I don't know how long this might continue. Possibly indefinitely? It's possible the BMS would consider cell temperature and open the charging relay. The charger itself continues to draw about 13 watts after it is disconnected (unplugged) from the battery.

The spreadsheet on left is the discharge test. The spreadsheet on the right is the charge test.

ePure 10A Discharge Test - Copy.ods
ePure Charge Test - Copy.ods

Error Sources

Data was recorded manually by observing digital meters and typing the readings into a spreadsheet every 5 minutes. The numbers are constantly changing and not all the readings are recorded at the same instant.

The digital meters (part numbers shown below) are very inexpensive items mass-produced in China. There is no way to calibrate them, and I have not checked them for accuracy. However, the values seem reasonable when compared to other equipment I own. I expect a few percent of error here.

Test Equipment

The photo below shows two active loads as they were applied when testing the EM 5.7 battery.

The active load was manufactured by AC-DC Electronics (P/N EL750B). It's rated to a maximum of 750 watts. Its maximum voltage rating of 55 VDC was slightly low at the start of testing. However, part of the unit's calibration procedure talks about testing its over-voltage protection circuitry which should kick in between 57 and 70 volts. Mine was set for 62 volts, so I felt safe using the unit briefly, slightly above its rated voltage.

The digital meters were manufactured in China by Peacefair. The following three meters were used.

AC power measurements: PZEM-061 (0 - 100 A)

Note: Having 100A capability is much more than required and possibly introduces a greater error than Peacefair's 20A unit (PZEM-051). However, their 100A unit uses an external current transformer whereas their 20A unit uses an internal shunt resistor. I originally designed this meter as a general-purpose lab instrument and did not trust running high AC currents directly through the 20A version.

DC load measurements: PZEM-051 (0 - 100 A) which uses an external shunt.

DC charging measurements: PZEM-031 (0 – 20A) which uses an internal shunt.

The charger's DC side uses a Neutrik NL4MP connector. The battery-side mate is a panel-mount NL4MX connector. Connections are as follows:

  • Terminal “1+” is Positive

  • Terminal “2+” is Negative

  • Terminals “1- and 2-” are jumpered together inside the plug.

Test apparatus: AC-DC Electronics EL750B electronic loads, PZEM-051 power meter, stopwatch, spreadsheet


As mentioned in the Battery section, I believe the cells used in ePure Race are Panasonic NCR1850A. Their datasheet shows discharge characteristics down to 2.5 volts, so I think discharging them to 3.0 volts is reasonable and safe. Also, the bike's onboard battery capacity monitor shows 0% charge remaining at 3.0 volts per cell.

I have estimated there are 168 cells in the battery. The NCR18650A cell is rated at 3100 mAh. At the nominal design voltage of 3.6 volts per cell, the total capacity would be 1874.8 Wh. EM rates the battery at 1875 Wh.

My testing yielded 1610 Wh (about 86% of its rating) for a battery manufactured in 2019 and tested in 2022.

For comparison, Cindy's EM 5.7 battery is rated at 1178 Wh. My discharge testing yielded 953 Wh (also tested at a 10A draw, down to 3.0 volts per cell). This is about 81% of its rated 1178 Wh. That battery was manufactured in 2014 and tested in 2021.

Originally, I had been concerned that Cindy's battery had experienced significant degradation, but now feel it's reasonable when compared with the newer ePure battery. (It's also worth noting that the 5.7 battery is lithium-polymer whereas the ePure battery is lithium-ion.)

I have kept a log of the AC-side power consumption during each charge for both EM batteries. A correlation between the AC side and the DC side of the charger has been established so I'll no longer bother to record the DC side. The 5.7 charger is about 87% efficient whereas the ePure charger is about 93% efficient.

Cell Life

Panasonic's datasheet for the NCR18650A shows the cell's capacity drops off steadily. They consider the end of life as 500 cycles. From this chart we can calculate:

  • 0 cycles = 100% of full capacity

  • 100 cycles = about 92% of full capacity

  • 200 cycles = about 85% of full capacity

  • 300 cycles = about 82% of full capacity

  • 400 cycles = about 78% of full capacity

  • 500 cycles = about 76% of full capacity

Credit: Panasonic