[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: FreeBSD Garmin 18x LVC receiver - precision -20

From: Gary E. Miller
Subject: Re: FreeBSD Garmin 18x LVC receiver - precision -20
Date: Wed, 21 Oct 2020 11:29:10 -0700

Yo Philippe!

On Wed, 21 Oct 2020 12:33:42 +0200
"Philippe Maechler" <> wrote:

> Hello Gary and List
> >> While googling I saw other outputs with a precision of -3 
> >> and the like.   
> > -3 is the assumed precision for time from the serial stream.   
> >> So I tought it has something to do with “how accurate 
> >> the position is”   
> > Your thought is correct.  PPS is kinda sorta accurate to -20.   
> >> Since I had the impression that the receiver is not “precise” I
> >> debug ntpd further😊   
> > What do you think you need to debug?   
> Oh boy, I should *not* attend online meetings and answer emails at
> the same time^^ I’meant that “Since I had the impression that the
> receiver is not precise, I “did not* look into the ntpd configuration

The GPS 18 LVC has PPS and can get the time on a server under 1 microsecond
jitter.  Maybe even to 100  nanoseconds with work.  The limitation is the
host, not the PPS.

> > I guess you are using NTP Classic, not NTPsec.  With NTPsec a peer 
> > and a server are the same thing.   
> Yes I’m running the «default” ntpd from base freebsd. Now, after
> reading lots of articles and how-to’s I’m thinking about switching to
> chrony or NTPsec…

We are biased to NTPsec here.

> >> server mode 1 prefer 
> >> fudge time1 0.000 refid PPS   
> > That is configuring ntp direct to the serial port.  gpsd is not
> > being used by ntpd.  But if gpsd is connected to the receiver, then
> > ntpd can not also directly read the receiver.  This is not per the
> > gpsd howto's as it does not use gpsd!   
> The system is now working. This comment made me thinking about using
> the receiver without gpsd and after a bit of googling I found *the*
> list of all the 127.127.t addresses available.

Or just read the doc.

> *SHM(1)          .PPS.        0 l   13   16  377  0.000 -0.252   0.085

A tad high, maybe it will settle with time.

>  And when I query from another host (client of this server)
>  pmaechler@NTP1-1:~ % ntpq -p
> *s1-ntp1.glattne .PPS.        1 u   29  256  377    0.388 +0.571   0.129

About as good as you can get with NTP over ethernet.

> Now comes the part where I fine tune the system 😊 If any questions
> arise, I’ll come back to a list, but I guess then I’m better suited
> on the ntpd list and not gpsd 😊

Or, better yet, the NTPsc list.

To do much better you need to start doing things like maintaining
room temp very closely, forcing the CPU load to always be the
same, etc.

Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

Attachment: pgpA7JpKEOeYb.pgp
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]