schufti wrote:@TO:
if your ESP was only ~4 min wrong for the year, you were lucky. My esp-01 was about 4min wrong per day!
I also tried to correct it in software but had to realize that the diffrence was much depending on ambient temperature. And this simple method gets very complex if your battery powered IoT gadget uses deepsleep.
So yes, one could save on the rtc but only under very specific circumstances...
Since I do have three DS3231s, I have decided I may as well use them and be reasonably sure that the time will be stable regardless of environmental factors. Clearly it is slightly harder to include one in a project than not but they are so cheap that having to add a small amount of complexity isn't a big deal.
It would be interesting to know to what extent performance varies by device and depending on environmental factors.