Electronics > Power/Renewable Energy/EV's

Lithium Ion 18650 Cell Idle Voltage


I am looking for idle voltage over charge charts for COTS 18650 cells.
Datasheets usually feature voltage over charge charts for certain charge or discharge for certain currents. I am sure that there will be differences between cell types, at the moment I am looking for general hints.

Background: I have a home brew driver for a
daughter wall clock. The circuit idles around 2µA and only draws about 30mA for a couple of seconds every minute to step the minute hand. Design goal was for a couple of El Cheapo Fleabay "3700mAh" cells to last for one DST/ST season (6 months).
The clock has been in operation for a while and I keep track of accuracy and cell voltages. Over the last three months I have seen cell voltage decrease from 4.06V to 3.96V. (Software compensated error for the quartz RTC is less than 1s/month.)
The clock was manfactered by Telenorma and must have graced some university lecture hall for decades, before it was discarded into the trash container from where I saved it. The whole thing is pretty ugly and dirty, yet loaded with nostalgia. I don't dare to clean it any further, because the paint is peeling off.

This is called open-circuit voltage. Average of charge and discharge curves (curve drawn in the middle) is a pretty good estimate because the voltage drop over the internal ESR is roughly symmetrical for both charge and discharhe. Use the curves recorded with the minimum possible currents.

Another way would be to extrapolate how a "0C" discharge curve would behave, by looking at discharge currents at two different currents.

Alternatively, a straight line from 4.20V (100% SoC) to 3.40V (0% SoC) is a decent estimate for many purposes.

https://lygte-info.dk/ has recorded discharge currents for many COTS cells, e.g. https://lygte-info.dk/review/batteries2012/Common18650comparator.php , with surprisingly low currents. Such low current discharge current isn't too far off from the open circuit voltage curve to begin with.

Siwastaja, thanks for the search term and the hint on how to guesstimate the open-circuit voltage from the existing curves. At the same time I keep recording the data from those cells.

Alex Wolf:
The only reliable way to obtain a discharge curve for specific conditions is to measure it yourself. While specifications and datasheets provide very approximate and average data, especially if they have nothing to do with your specific product from a dubious supplier. However, keep in mind that cells of questionable origin may be susceptible to poor repeatability due to inconsistent quality. Of course, with such a low discharge current this would take an unacceptably long time, but with a discharge current of say 10 mA you can get very close to the desired values in a couple of weeks (and using a multimeter or oscilloscope in datalogger mode, this can be done automatically). The zero charge state can be defined as the point of 10% from the start of the steep voltage drop on the discharge curve at a constant discharge current. The rest of the curve can be seen as a fairly linear relationship between charge level and voltage level (for a given cell chemistry). A simple constant current sink circuit can be build on a TL431 (see its datasheet from TI).

Well, thanks for your input. The cells are cheap, for sure. Since the current discharge cycle has been running for a while, I will continue to monitor it. Attachment Cells_231216 shows how their voltage has changed during the last three months. Measurements are taken using normal probe tips with a cheap multimeter, while the whole device hangs on the wall above my head - so don't interpret too much into outliers.
Today I took the time to take the clock off the wall and have a look at the cells' voltages. Since the cells react to temperature and orientation, I had to wait several minutes before I could get meaningful measurements.
Cell_Top_10mA_2312161546 shows two events. Between -1m00s . . -0m50s we see the drop in voltage when the minute pulse occurs. For a short time, the device's current consumption rises from 3µA to 25mA. After dropping by 3mV, the cell voltage quickly recovers.
From about -0m48s 10mA load current is drawn from the cell. This slope is magnified in the next picture Cell_Top_10mA_2312161548. After only 2 minutes of 10mA discharge, the cell voltage takes several minutes to recover and stabilize.

Because I don't want to disturb the running clock too much, the above measurements include the connector resistances and the internal resistance, 140mR in total. With an additional pair of hands I could show that the cell resistance is about 70mR at 25mA. (Sanity check: The datasheet of the Panasonic NCR18650GA gives us about 40mR at 2..8A at 25°C).

Where would my clock's cell be on a 2A discharge curve right now? 3.92V-(2A*0.07R) = 3.78V
I'll keep in mind to mentally subtract 140mV from my no-load measurements. So much about my guesstimate, I'll keep recording.


[0] Message Index

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod