So we've managed to put a couple of first complete prototypes together. The first attempt failed, so we just put the MCP73871 on for testing the charging only in version two. Unfortunately this competition is in the middle of our very gray winter in Sweden when we very rarely get direct sunshine, and then not for very long. But we actually got one chance to put it in the sun, with code to report  batteryvoltage via AIO. It had positive results, the voltage increased slightly, but we didnt have the Huzzah32 in deep sleep during that only run. Weve since have put deep sleep on, but no sun. We are suspicious though, because you would think that a 2 W solar panel would give something held very near an artificial light source, even though most of the light sources nowadays might not have the right wavelengths for producing current. That is, if i remember correctly, something ive seen before.

Also, we have a drop from 6 V no load to 3,8  V after the diode, when plugged in. we would expect it to average 4,5 V which the voltage divider to VPCC is set to, just like Adafruits own design. We havent been able to measure the same experiment on their solar charger that uses the same chip.

So, inconclusive as for now, but the design should work, albeit that there are possibly some redundancies with the fuel gauge and the INA. Maybe there is a component that does both of their jobs. Or some other combination of fuel gauge and charger. I have found one that is usable for solar though. Tips are welcome!

Happy new year, folks!