I culled through a bunch of AA and AAA cells to find any that were dying or nearly dead. If I found any that were below 1.2 volts they were considered dying, and below 1 volt, they were considered dead.
I then put some of the dead ones into Joule Thiefs to suck out the last few milliJoules of energy left. Since these cells were so low voltage, I used Joule Thiefs having a germanium transistor, so that the Joule Thief’s LED would still be lit when the cell’s voltage is well below half volt, which is typically where silicon transistors quit.
Germanium transistors typically have lower current gain than silicon; usually less than 100, which is half to 1/4 of silicon, which is typically 200 to 400. The standard silicon JT circuit uses a 1000 ohm resistor for silicon transistors with gains of around 300. For germanium, the resistor should therefore be lower, 470, 330 or possibly less.
The Joule Thief I used had a 330 ohm resistor. At less than 1 volt the LED was bright. Over a period of a few days the led grew dimmer and dimmer and finally went out. So I took the cell out and put in another “dead” cell. After an hour or so, I checked the voltage of the first cell and it had climbed from less than 0.25 volts up to more than a half volt. The cell had recovered or rebounded back to where it could run the Joule Thief. I removed the second cell and put the first cell back in. The JT lit up for about a half minute, and then dimmed down to nothing.
I have done this before with nearly depleted batteries, and usually I can repeat this recovery and depletion a few times before the cell finally dies. I think what’s happening is the chemicals in the cell get used up until the small amount of chemicals left can only power a light load. Even a low power JT becomes too heavy of a load for the cell. The cell’s internal resistance goes up and the JT dims out. Then after recovery, the internal resistance drops enough to light the JT for a short period.
I thought that if my theory is correct then if I reduce the JT’s running current, the cell will not dim and recover, but instead keep the JT lit longer with less or no dimming and recovery. This is easy to do; just increase the resistance of the base resistor. The JT will then run at a lower current. The existing resistor was 330 ohms; I added a second 1k, 1/8 watt resistor in series for a total of 1.33k. The germanium JT ran fine with this much resistance, and the LED was bright with the cell voltage of less than 1/3 volt. I could have put a very small trimpot in place of the 1k resistor – a 2.5k or 5k might be a good choice. Or a small switch across the 1k to jumper it and go back to 330 ohms.
I think this topic needs further experimentation, but that can wait for another blog.