Sunday afternoon I made a few more updates to the solar library. I soldered a thermistor to the charging circuit and taped its sensor to the battery. According to the documentation, this should stop the battery charging if it gets too cold or too hot.
I also hooked up my new 3.5 Watt solar panel. The initial panel for this project was a mere 1.4 Watts and my limited testing with it suggested it wasn't even worth doing an overnight test with it, so this is the first time this project has actually been solar powered for more than a couple hours. The new panel is looking promising so far. It's been mostly overcast and rainy the past few days. Sunday afternoon I started with an almost full battery. Despite the clouds, the solar panel was able to provide enough juice to stay off battery power until about 6 pm. The battery then discharged overnight as expected until about 7 am when the panel started providing enough juice to hit equilibrium. It was another cloudy day, but by 10 or 11, the panel was providing enough current to charge the battery. By 2 pm it was full. Shortly afterwards we had a storm roll through and we switched back to battery power. After the storm, the sun came out and quickly topped the battery back up. Here's that first day in graphical form:
Day 2 (today) was also mostly cloudy. The battery had just started to charge at 9 am when we had another storm come in. But after that storm, the battery charged back up and was full by about 1 pm, which isn't bad for a mostly cloudy day.
I'm pretty happy with the current situation. I think it's time to just leave it in place and monitor it for a while. The question remains how it will perform in winter with longer nights, I still see the potential of running out of charge if I get a few real cloudy winter days in a row. But only time will tell.