[time-nuts] PLL GPS locking shortcomings

[Poul-Henning Kamp]

In message <02BE24C395837C449971DFB9BEF1FA8AC74A42 at rakonmail.rakon.net>, "Murra
y Greenman" writes:

>The limitation on most GPSDOs is as Bruce says, lack of DAC resolution.

But for the timeconstants we are dealing with here, the DAC is 1000
times too fast, so oversampling works great.

I did that with Dave Mill's Loran-C receiver, I accumulated the
rounding error for the DAC output (as in floyd-steinberg dithering)
and update the dac on every GRI (15 times a second).  RC lowpass
filter filtered the high frequency (ie: 15 Hz) related components
out.

The effective increase in resolution was about 5-6 bits.

-- 
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.