[time-nuts] Any Spectracom hackers out there?

[Poul-Henning Kamp]

In message <417C0616.9080100 at febo.com>, John Ackermann N8UR writes:
>Poul-Henning Kamp wrote:
>> In message <417BFF07.5050203 at febo.com>, John Ackermann N8UR writes:
>> 
>> 
>>>Oh, by the way -- you can see the 8170 jump when phase lock is lost at 
>>>http://www.febo.com/stats/ntp -- scrolling down through the graphs will 
>>>show you the current day plus two previous days of offset information 
>>>via ntp.  I plot performance of my two Z3801As as well as the 8170.
>> 
>> 
>> Stupid question time:  Isn't the problem here that the internal
>> 10Mhz should be calibrated better ?
>
>Good point, it might be possible to trim it closer, but it's still a far 
>less stable source than the external reference.  Spectracom apparently 
>knew this could be a problem, since they had an option to allow an 
>external reference to kick in when lock was lost; my unit doesn't have 
>that option, so I'm experimenting...

I don't think so.  I think the external reference were for long
periods of dropout (lightning in the antenna), not for short blips.

Also, it is very likely that your short blips are caused by the
oscillator being right out at the edge of what the PLL can pull.

-- 
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.