[ntp:questions] Re: Audio refclock+linux+soundcard = yuck

Tim Shoppa shoppa at trailing-edge.com
Thu Oct 2 17:10:05 UTC 2003


Adam Myrow <amyrow at midsouth.rr.com> wrote in message news:<bRmeb.1139$oC5.453 at clmboh1-nws5.columbus.rr.com>...
> Have you tried the CHU audio driver?  My experience was that if CHU was 
> putting in a decent signal, I often got better performance from it than 
> the WWV driver.  Actually, for me, the CHU driver worked while the WWV 
> driver did not.  The best I ever got from WWV was the date and part of 
> the time code.  Since the CHU driver samples less often, I suspect it's 
> less sensitive to buggy drivers or hardware.

It's probably a different problem than yours, but my K7S5A's motherboard
on-board sound is way out of specification in terms of the sample
rate... several hundred ppm.  The audio refclock docs say that being off
by more than 125 ppm will make it impossible for refclock_wwv to lock up.
And that's what I see... it will see occasional second blips but never
does so reliably and never sets the minute pulse seen flag.
The CHU decoder is much more simplistic (and I suspect tolerant).

When WWV is coming in exceptionally clearly, all the digits turn up good
in a minute or two after getting the minute pulse locked up.  When it's
not so clear (but still mostly audible), I've seen (in my limited
experience) it take hours to get all the digits.  How long did you wait,
and how clearly was WWV coming in?

Some SW receivers aren't so good about letting the 100Hz subcarrier through.
My RX-320 does very well in this respect.  Some other receivers I tried
obviously regarded 100Hz as being outside the passband (though there
was still enough to eventually lock on).

Tim.



More information about the questions mailing list