Yes, you are correct - there is no true drop to low, I just stopped the previous test when the light got dim. Here's a runtime graph I made with the V2 light on unprotected 14500, and it was heading to zero. I stopped it because it's an unprotected cell! :
[...]
I made this very unscientifically, with a solar cell and my multimeter hooked up to the PC logging voltage off the solar cell with the light about 1.5 inches away on high.
The cell has 2.68 volts now; do I dare recharge it?
.
Thank you for the measurements. I hope your cell will be well
It was a relatively high current discharge so if you haven't discharged the battery for too long and too low (~2V under load) the voltage should raise above 3V soon. If it is above 3V, you can charge the battery in the standard way (and you should do it soon so it doesn't sit discharged).
If the cell doesn't want to go back above 2.90V-3.00V, it was overdischarged. It is generally recommended to start charging of a cell that is below ~2.80V resting voltage with the low current of ~0.1C (~50-70mA with a 14500 cell). The cell shouldn't heat up and its voltage should raise above 2.80V in less than 30min. - 1h. If it doesn't want to go above 2.8V in that time or heats up, I would consider such a cell dead. When it raises above 2.8V it has recovered and can be charged with the standard 1C charge. I don't have experience with reviving abused cells so you should confirm what I wrote in this paragraph.
The problem with using solar cells to measure the illuminance (lux) is in their characteristics which make it really complicated, if not impossible, to relate the voltage or the current to the illuminance. You can get a relatively cheap photodiode (e.g.
BPW21R) that is dedicated for the task. Note that the voltage usually has a logarithmic relation to the illuminance which makes it possible to measure wide range of illuminances but makes it hard to do it precisely. The current measurement is sometimes better suited for more precise measurements.
EDIT: Even with all the problems with the precise interpretation of solar cell measurements, the graph is still really valuable. It shows that the output is not perfectly flat (would be easier to see with a little bigger image
) It's hard to tell precisely what the change in brightness was but with human eye insensitivity to such changes it probably isn't noticeable anyway.