[ntp:questions] fudge time1 for gps-18x-LVC?
snews at lordynet.org
Mon Feb 8 10:20:41 UTC 2010
David Lord wrote:
> David Lord wrote:
>> David Lord wrote:
>>> that I can't get parallel (or serial) pps to work with just
>>> ATOM driver flag3 0 | 1 (kernel discipline disable | enable)
>>> and no other refclock, ie I'd like to just use another ntp
>>> server as preferred with pps to condition the clock.
>> Don't know what I did wrong previously but just removed
>> 127.127.20.0 lines from ntp.conf, set me6000g as prefer,
>> restarted ntpd and had following:
>> oPPS(0) .PPSb. 377 -0.002 0.004 pps from GPS
>> *k6x400 377 0.246
>> +me6000g .PPSa. 377 0.106 pps from MSF
>> -p4x2400c 377 0.533
>> The above with parallel pps0 at ppbus and /dev/pps0
>> created using mknod same as tried earlier.
>> Also tried with p4x2400c as prefer and that's ok.
>> Replaced /dev/pps0 with /dev/pps0 -> /dev/tty00
>> disconnected parallel port, reconnected serial (DCD and RxD),
>> restarted ntpd and that's working ok.
>> Can't think what I've done different. Then again MSF
>> wouldn't sync end of last week after I'd made a change
>> to ntp.conf and then I noticed LED on receiver wasn't
> Short term memory loss I think.
> I'd no sooner posted that when oPPS(0) flipped to -PPS(0)
> and continued flipping with offset increasing from 4u
> to 145u. I then restarted with kernel pps selected
> (flag3 1). That was no better, and same when I went back
> to parallel port still with kernel mode pps. I even
> mentioned this behaviour in a much earlier message.
> Now back to parallel port with PPSAPI (no flag3) and so
> far this does seem to be converging with PPS(0)
> remaining selected as pps.peer for 25 minutes so far
> (no doubt until I hit the send button). Offset has
> converged from > 250us to 12us with jitter now at 4us.
Offset now 87us but up and down wildly over 520 minutes
alternating between syncronised to PPS(0) = 32 events
vs prefer peer = 32 events.
I might look at changes since version I'm running
(4.2.4p6-o) and try to compile most recent pre release,
but there seem too often to be changes required on
NetBSD for compile to correctly functional binaries.
More information about the questions