Long Life Lithium, charge and discharge thresholds

[bronsonj]

So many of us know what kills lithium. Overcharge, overdischarge, overheat, physical damage and using the pack while it's really cold.

But I was reading online and found some interesting stats on the life of a lithium cell. The person using the cell had several small lithium ion cells and charged/discharged them to very specific voltages to see the effect on charging compared to the life of the cell.

When charged to 4.2v, the cell life cycle matched what was advertised and it had 103% of rated capacity.
When charged to 4.15v it had 1.4x the cycle life and 96% of rated capacity.
When charged to 4.1v it had twice the cycle life and 92% of rated capacity.
When charged to 4.05v it had 2.8x the cycle life and 86% of rated capacity.
When charged to 4.00v it had 4x the cycle life but only 75% of rated capacity.

All tests were done discharging to 2.5v (too low for our use) but they were all discharged to that spec. The tests were done at room temp and the cells never had an elevated temperature that was humanly noticeable.

To that end, what do you charge your cells to in your pack?

At the moment I am charging my pack to 4.06vpc for 56.8V at the pack.

Additionally, the discharge cutoff can have a large impact on cell life and we shouldn't discharge anywhere near the 2.5v done in the test I mentioned. But I have not seen a simple cycle life and capacity comparison based on discharge voltage. What is the max discharge you take your cells to? I stop at 3.25vpc (under load) for 45.5V pack voltage.

[bronsonj]

Oh, another thing, the moving lithium 'knee'.

I also saw on a similar series of tests with smaller lithium cells the knee voltage when the battery was discharged at 0.5C and a 1C. For these smaller 18650 cells discharged at 0.5C the knee voltage was remarkably different than the voltage for the same cells discharged at 1C.

For their tests they used different brands of cells to compare against each other as well and the voltage at the beginning and end of the knee also varied greatly with what brand/capacity they were using for that test.

Vaguely averaged their tests and saw:
Discharged at 0.5C the cell had 20% remaining capacity at 3.6v and was basically at the bottom end of the knee.
Discharged at 1C the cell had 20% remaining capacity at 3.35v and was also basically at the end of the knee.

I can't draw a useful conclusion for the end-of-the-knee voltage though since the various brands were so much different.

[Sergio]

That appears to be the experiment from Chalmers University in Sweden from the battery university article:

https://batteryuniversity.com/learn/...f%20Technology