Originally Posted by
n4zou
...Overcharging the 4 Ni-MH batteries is possible. I've found that in actual use this is not a big problem. When using just the LED headlight and LED taillight with just 120mA of recharging current and traveling anywhere other than an interstate highway (illegal to be on one on a bike in the USA) it's almost impossible to overcharge the batteries...
I modified my circuit as shown below. I used 4 2000 mAh AA cells from Radio Shack. They were much better behaved under charging than I'd expected. As the battery reached full charge the voltage went to 1.36V per cell and the charging current dropped to 40 mA. (My circuit dissipates the excess power in a zener shunt.)
Link to larger image
I don't have personal engineering experience with NiMH cells, but I read that the newer batteries are designed to better tolerate a certain amount of overcharging at trickle currents. So the stern warnings we see about the dangers of trickle charging with NiMH might be somewhat overstated with respect to current NiMH cell designs. At least that is what I'm thinking right now.
If overcharging isn't a serious concern, n4zou's circuit might have an additional advantage. Its charging current is full wave rectified DC, which might amount to a pulse charging scheme. That is, it puts a pulse of charge into the cell and then briefly removes the charging current. The interval between charge pulses might allow better recovery from polarization effects. So maybe I'll try removing the 4700 uF filter capacitor in my circuit.