Frequency Reference for NTP Use?

[Alex Satrapa]

I'm trying to hook up my Garmin GPS-II+ GPS to use as a reference clock 
for xntpd.  Unfortunately, there's just way too much jitter (the NMEA 
driver has to parse the time out of a lot of cruft including such 
useless information as the current physical location).  I'm hoping that 
it might be possible to get a reference of some sort that can provide a 
1PPS signal to discipline the GPS clock (ie: use the GPS to provide the 
time stamp, use the PPS to provide the exact time).

Has anyone else been looking for frequency references/standards?  All 
the things I've found insist on being GPS clocks too, and they generally 
cost in excess of $US1000.  I just don't have that much money lying 
around to spend on a clock.

Ideally, I'm looking for a 1PPS source with an RS-232 or Parallel port 
interface, that has accuracy as low as a few seconds a month (which 
means I'd be happy with a Quartz oscillator).  I would love to have a 
Caesium clock, but I just can't scrape up the $1M odd that I'd need ;)

My biggest reasons for avoiding GPS-based external clocks is that:
1) I already have a GPS
2) I don't want to rely on the GPS network for my time (I may not be 
able to receive the signal)

I'm wondering whether I might get away with building my own oscillator 
from a 10MHz crystal and just triggering the "CTS" line (or whatever) 
every 10M cycles - but I'm thinking it would be cheaper to buy someone 
else's proven technology.

Alex
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 225 bytes
Desc: not available
Url : http://lists.samba.org/archive/linux/attachments/20020712/53f43cd9/attachment.bin