This unit is as close to an atomic reference as one can get for NTP. It takes the rubidium-based clock source in the GPS satellite network and locks with time tagging to UTC time of the GPS network. Ultimately this is governed by the Caesium clock at NIST (for me being based in NA) computer time is only good to about 0.02-0.10 seconds as computer and equipment clocks are not accurate at all for keeping time, hence the need for the NTP server. The magic of this unit is it uses the GPS to drive the NTP server we are going to use it on summer field day for some digital modes that need a NTP server when we don’t have internet access. Can server NTP from a GPS antenna. the radios and computer will need to sync time about every 20 seconds to a time accuracy of 1 sec Cesium gives a time accuracy of about 2-3 parts in 10^-14 Rubidium give a tame accuracy of about 1 part in 10^-12 This allows me to give the computers, servers, and equipment a good clock for time stamping. I am not running this as the 10Mhz reference. The Stanford Research SRS725 (Atomic Rubidium) unit is still providing the 10Mhz for the lab as this is the most accurate time base that I have available here. The LEO unit is syncing time date clocks. The RF time bases are on the atomic reference. Zen

@@ZenwizardStudios You do know this is a rabbit hole par excellence? 😋 (Says one who owns 2 GPSDOs, an Efratom HLN Rb and for good measure a Meinberg DCF77 driven NTP box which also produces a 10 MHz ref. So should GPS go down.. )

In theory it should never be off as this is part of the self-updating chain back to NIST. This goes back through the GPS Rubidium to the Cesium at NIST so it should be always accurate and self-updating.