Hi,
is there any solution to have permanently a correct clock without gap in the data acquisition? A solution like xntp is possible in CRBasic? Several tests we have made have failed: tests of collecting data and then making a clock change lower than the minimum time between two acquisitions have all lead to data gaps.
Thanks,
LGS
If you jump the clock of the logger forward beyond the time when it is due to store data then you risk not storing that data. You can avoid the worse of these issues by only allowing changes of the time well away from the data is about to be stored, but the logger may still skip over a measurement scan although in most circumstances this is not significant (as averages etc are calculated based upon the actual number of samples).
There are several ways of synchronising the logger time.
It does support NTP (for an IP connected logger it is via the NetworkTimeProtocol instruction). However, you still need to avoid calling this instruction near data storage times and also using too tight a deviation limit, as NTP can have some jitter especially on laggy links.
You can also sync back to a PC but this needs to be done carefully especially as PC clocks are known to drift even within the normal times they are sync'ed to a server. Virtual servers appear to be even worse in this respect.
The ultimate solution for the datalogger is to use a GPS time source. All of our modern loggers (except the CR200) can use a GPS system (with PPS output) as the main clock driver which totally eliminates the clock jumping. This does though require some extra power and also sight of the sky for the GPS receiver.