[time-nuts] Turning off PPS when not enough satellites

[Poul-Henning Kamp]

In message <000001c6b659$d4d289b0$03b2fea9 at athlon>, "Ulrich Bangert" writes:
>Bj=F6rn,

>Motorola specified the VP Oncore with:
>
>130 ns observed (1s) with SA on
>In position hold mode, < 50 ns observed (1s) with SA on =
>
>Ok, there IS a certain improvement with position hold. 

The important thing about position hold mode is not that
you get better 1PPS but that the receiver knows how good
the 1PPS is.

In navigation mode, a satellite mistake will affect your
1PPS.  It will still be within 130nsec of the 4D equation
solution, but the 4D equation solution is wrong because
one of the inputs is wrong.

I pos-hold mode, the receive can and will spot the
wrong input, throw it out and do the math again and
because it's doing Nsat equations with one unknown instead
of Nsat equations with four unknowns, it will get a
much better fix on that one variable.

But the important thing is that it know the solution
makes sense.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.