David J Taylor wrote:

- thanks for the notes on the GPS 18 LVC. Mine wasn't set like yours, and by chance I got the version with the longer 5m lead.

Yes mine has a 5m lead too which is handy.

- kernel recompile took 5 hours on this 133MHz/48MB system

Ouch!

- without the kernel recompile (as I understand), the PPS line looks just like another Internet source, so the jitter might be milliseconds rather than microseconds. You /have/ to do the recompile for best accuracy.

Compile finished, but it doesn't seem to have improved unfortunately :( This GPS is *really* innacurate...

xGPS_NMEA(0) GPS_NMEA(0) 15 l 25 64 377 0.000 -215.79 10.321

- you /did/ wire the PPS to the DCD pin, and not somewhere else? Even a multimeter should show the signal (well, an analog one might). IIRC, the GPS delivers the PPS signal even if it's not position locked.

Yes.

- the pulse is nominally 200ms long, so if you have the "wrong" number of inverters between the GPS and the RS-232 connector, you may observe a 200ms offset, and you may need to tell NTP to lock onto the other edge.

Not sure what that means.. :) I connected it up as per the instructions on the google article I linked to.

Tony
_______________________________________________
timekeepers mailing list
[email protected]
https://fortytwo.ch/mailman/cgi-bin/listinfo/timekeepers

Reply via email to