Timing drift

Any information on timing drift. I'm recording temperature every second on ten minute intervals and noticed that there is a random timing drift from 2-7 mS per measurement which adds up to about 700 mS per 10 minute interval. Any software fixes that would address this?

Thanks

Justin Tabatchnick

Details are usually welcome

sounds like you're using delay() which won't account for processing time.

if you use millis(), set the next timestamp to the current one plus the desire interval

Does the board you're using use a resonator or a quartz crystal for timing? Resonator based boards like the UNO drift like crazy.