[ntp:questions] Questions about joining pool.ntp.org

David J Taylor david-taylor at blueyonder.co.uk.invalid
Wed Aug 31 08:10:37 UTC 2011

> It is well below 0dB.
> ahh, here:
> http://www.northwoodlabs.com/AN101.pdf
> around -22dB
> you get a "win" by correlation with the CA PR pattern (~1000 bits).
> ( you need better than +6dB S/N afair)
> uwe

Thanks for that pointer, Uwe.  That -22dB is for a perfect receiver, as 
well.  The document lead me to the GPS specification documents, where I 
found this in appendix A, section 5:

"5.5 Time Transfer Accuracy Characteristics
Time transfer accuracy based upon the output of the position solution is a 
function of SPS timing errors with respect to GPS time, and GPS time scale 
errors with respect to Universal Coordinated Time (UTC) as it is 
maintained by the United States Naval Observatory. Current satellite SPS 
timing errors with respect to GPS time are on the order of 75 nanoseconds 
RMS (including propagation effects). The GPS Control Segment consistently 
manages GPS time coordination with UTC to better than 30 nanoseconds.

When the combined GPS time prediction error and GPS-UTC time 
synchronization errors are mapped into the position solution, the RMS time 
transfer error is on the order of 110 ns. The 95% UTC time transfer error 
should not exceed 250 nanoseconds, based upon measurements of the position 
solution time offset. The 340 nanosecond performance standard defined in 
Annex A provides the Control Segment with flexibility in the management of 
the GPS time scale."

How does this square with those who claim 4ns from their GPS devices?


More information about the questions mailing list