More testing complete with the solar panel and batteries in circuit.

This is a little graph (actually quite big) of the 4v solar panel in circuit and charging the batteries. I hooked up my multimeter to a laptop to record mAs and rigged up and Uno + Ethernet shield to record the voltage. I only ran this for just over an hour because it was a day with patchy cloud and I was worried about rain (not waterproof yet). Still, the results are encouraging.

The voltage output from the solar panel is pretty steady and around the level I need, certainly enough to charge the batteries. The current level fluctuates a lot, this is down to cloud cover and angle toward the sun.

If I can source a diode with a lower voltage drop diode, like a good Shotkky diode or perhaps use a reversed P Channel Mosfet bodge, then the voltage from the panel to the batteries will be more than sufficient.

In terms of current the average making it into the batteries over the time period was 4ma, sounds a bit crap right? Well in just over an hour on a cloudy day the battery voltage increase by 0.05v and on a sunny day it would be much more efficient. You can see from the graph where the sun appear from behind the clouds which yielded a whopping 21ma at peak. If we could chuck that into the batteries we could run the lights for 40 hours. If you recall the voltage drop on the batteries running the LEDs for 2 hours was 0.005v.

Should I worry about over charging the batteries? Not really. NiMh batteries tend to lose some of their charge each day and a trickle charge of 0.1c (100th of the battery's capacity) is fine according the the battery manufacturer.

I forgot to take a picture outside... sorry.

With a bit of careful part choosing for the diode or equivalent and some circuity to switch on and off the LEDs (Probably a transistor from solar panel) this setup should meet the requirements.

Woahh! Steady on there, this panel is producing a massive 0.06 watts at peak!!!!!!