Hello again,
Our competitors over at Nordic Semiconductor have recently released an article on coin cells as part of their newsletter, which can be found here (page 9).
They have also identified the internal resistance of the battery as the primary state variable of a lithium coin cell, but they have made some different assumptions from us, which causes their conclusions to be a bit different from ours.
First, the manufacturer-provided internal resistance plots are measured for a 2-second pulse, which is very different from the <1 ms active periods of Bluetooth low energy and other low-power protocols. A steady-state "law" such as Ohm's law might not be a good tool to analyse pulsed operation like this. From personal experience, I can say that batteries are quite complex electro-chemical systems and can sometimes exhibit strange behaviours. I remember especially some primary Lithium cells I worked with years ago that showed overshoot/ringing with time periods measured in seconds - strange stuff!
Nordic's analysis is based on data only from a single battery manufacturer, while our results showed big differences between different CR2032 manufacturers (with effective capacities varying from over 100% of the rated capacity to only 50% of related capacity, both at 15 mA and 30 mA peak currents).
Unfortunately, it seems that so far there has not been a lot of focus on pulsed operation of Lithium coin-cells. This is understandable, as the pulse periods can be very different for different applications, and battery manufacturers are not application experts. I believe our lab results give different enough results from analysis using steady-state methods that it looks like a more detailed model of coin-cells may be needed for pulse operation. Hopefully, battery manufacturers and semiconductor manufacturers can work together to gain further understanding of how coin cells respond to pulse currents. Maybe this could even make a good research project for an MSc or PhD student?
Best regards,
Karl H.