RTC loses 1-2 seconds per day on MKR1000

I have a project that uses the MKR1000's RTC. When it starts, it obtains the time from NIST, then assumes that the RTC will maintain an accurate clock from then on (for months).

It seems that the RTC on the MKR1000 loses 1 to 2 seconds per day. Is this normal and is there anything that can be done to correct it (other than updating the time from NIST periodically)?

Thanks.

Some sketches can disrupt the RTC.

If yours does in any way (cannot see it from here :wink: ) then the update from NIST would be the way forward.

Almost certain this type of subject has come up for MKR series before so maybe a search of the forum for "mkr rtc" would find you some more clues.