That is true, especially your last point. For periodic wake-up it is still useful even when that period may not be accurately known.There are use-cases even if it is wildly inaccurate.
The primary one is polling very slow events for power saving purposes - go to powerdown state "for a while", wake up, poll something, go back to powerdown if it hasn't happened yet.
Also even a wildly inaccurate RTC can still be more useful than no RTC at all
I suppose one thing which should be added to the list is how LPOSC does vary over time as that's important for battery powered situations. if a 100 ms wake-up period is desired then being 99-101 ms won't make much of a difference in most cases. But, if it's variability is much greater than that, it could.
Uncalibrated, according to the datasheet, it seems LPOSC could be as much as +/-64% out from what it should be. Even calibrated it could vary by as much as +/-36%, though one would hope it would not be quite that bad in most cases.
Statistics: Posted by hippy — Thu Nov 14, 2024 7:09 pm
