I began to experiment with this kind of charging over a month ago. Here is what I have seen.
The battery is a 1.2v Nicd battery with 500mAH printed on its label.
It was in a sorry state where it could only be recharged to 1.28v then would stop working at 1.23v.
So, I thought it was a good candidate for this kind of test.
Bedini says Nicd's recharge well with spikes.
I charged it from the secondary of a jtc which gives 70V spikes at a rate of 45khz.
After recharging, I either let it rest a bit or put it directly into a germanium based joule thief with 1 led in the C-E and count the time the light is on.
At first the light continued to turn off when the battery was at 1.23v, and it did this in less and less time.
I needed to take drastic measures to break into the lower end crystals.
I actually used the jtc with the germanium transistor to pull down the battery.
It did not work. The light went off at 1.23v
What really worked was to put the battery into a jtc running a LoA bulb from its secondary.
This circuit continues to drain a battery well below the turn on level of the transistor.
This worked and on the 5th cycle I had a working range of 1.19v -- 1.31v which immediately relaxed to 1.27v
However, by the next day the battery under load only flashed briefly then went out.
That was rock bottom. (and a little discouraging)
The range remained very narrow and I continued to drain the battery with the jtc with the LoA bulb on the secondary.
Finally on the 9th cycle, the range began to widen and the length of time under load increased.
After this 9th cycle, the Nicd filled to 1.36v after 10 hours of recharge.
The time it lit a jtc was just under an hour.
Each day after that it improved more.
The time to support the load increased and also the brightness of the light increased.
At the end of the
16th recharge the light went on very very bright and stayed bright for longer than 2 1/2 hours.
then I packed it away and must have shorted it.
-----------------
But, I did not need to start ALL over again.
It took 5 cycles to bring it back to the 2 1/2 hour load time.
I popped a normal healthy and charged Nicd into the same circuit and it light much brighter for over 5 hours, so there is a way to go.
Today, after 10 recharges since the short, the led is at full brightness for more than 2 1/2 hours.
---------------
I have devised a way to measure the brightness of the led.
I hold my hand or a paper a distance from the top of the bulb and measure the distance to where the light is too diffuse.
Soon, this will not be a good measurement, because yesterday the light hit the ceiling of the room for over 2 hours.
=============
Next time I will lower the charger circuit to have spikes just twice the height of the full battery level. I betcha I wasted about 67v in each spike.
thank you,
jeanna