I already did some complete discharges in my KD buckle v4 Q5.
But I only did it after first carefully measuring the input current and the battery voltage drop under load. Is seems to work without any problems with my light and my 10440s. But I recommend not to do this if you do not know exactly what you are doing. With a LED with slightly lower forward voltage or a battery with lower impedance the flashlight could be damaged.
And even when the flashlight is not killed by the batteries it is very likely that the flashlight will kill the batteries by overdischarging them.
But it is amazingly bright running on 10440s, more than twice as bright than whith my Eneloops. And even more astonishing, I get a runtime of almost 30 minutes with one of my two 10440s fully charged to 4.2V.
Runtime plot with the better one of my two 10440's. I used my USB scope to measure the photocurrent of a photodiode placed some centimeters in front of the light. I clamped the head of the light in a vise so that I could change batteries without moving the light. After 27 minutes of slightly dimming very bright light and three more minutes of constant brightness the light suddenly went very dim. When you look at my current draw curve on NiMHs you see that the battery voltage must have suddenly dropped to below 0.5V.
Runtime on Eneloop NiMh's. Approximately the same runtime but less than half the brightness.
I repeated the measurements 3 times and always got nearly the same result.
Runtime plot using alkaline primaries. It's just a complete waste of money to put alkalines in this light.
Current draw and battery voltage using a 10440 battery. Current measured as voltage drop over a 10 milliohm resistor. This time I ended the measurement before the battery was over-discharged. But you can already see the voltage dropping and the current rising after 30 minutes (crazy software: 29minutes 60seconds). That is the beginning of the approx. 3 minute constant brightness period.
Eneloop current draw and battery voltage. After 30 minutes I took the lamps head form my test setup, the battery voltage rose immediately.
Look at the values after 10 minutes of runtime:
LiIon 10440: 3.4V * 0.62A = 2.11 W input power
NiMH: 1.13V * 1.9A = 2.15 W input power
(input current is nearly pure DC with only a barely visible ripple from the boost converter running at approx. 1 MHz, voltage shows no visible ripple)
How can it be that the input power is nearly identical but the light is more than twice as bright running on 10440s?
There are two possible explanations:
- driver efficiency is very low using NiMH's.
- something is seriously wrong with my measurements
as if something was shorting, and flickered!
