I’d need to leave a datalogger running for a while to measure internal and external temps to see if forced ventilation will provide any benefit (as we know, without a heat pump, it is not possible to cool below ambient).
Regarding the solar panel, I built the box specifically to hold up the panel (hence the roof angle) - that’s part of its purpose.
Water cooling the box would be an interesting project in its own right, but realistically, it wasn’t the heat that killed the Pi, it was the copper heatsink causing a short.
An alternative way to look at it is:
- the heatsink adhesive should not have failed.
- the fan shouldn’t have failed - if it hadn’t, the adhesive would have lasted (probably).
(For the want of a nail …)
There are some interesting comments on the Pi’s thermal managment where it describes what happens between 80~85°C (thereafter it threatens you with a red thermometer!)
Peltiers are great, but come with their own problems. They are power hungry, sure, but if it’s humid, you’ll get condensation on the board.