[ntp:questions] help - clock jumps up to 5000 ms startup ntpd with usb gps linux

unruh unruh at invalid.ca
Mon Feb 20 03:03:01 UTC 2012

On 2012-02-19, David J Taylor <david-taylor at blueyonder.co.uk.invalid> wrote:
> "unruh" <unruh at invalid.ca> wrote in message 
> news:Zu90r.3511$zA2.2177 at newsfe02.iad...
>> On 2012-02-19, Rob <nomail at example.com> wrote:
>>> Ron Frazier (NTP) <timekeepingntplist at c3energy.com> wrote:
>>>> *GPS_NMEA(5)     .GPS1.           0 l    4    8   37    0.000  563.733
>>> Don't try to use NMEA for timesyncing!  It will only drive you crazy.
>> No. It is fine.
> NMEA alone may be little better than Internet servers, it may be worse.
>>> Some people are lucky and seem to be able to get a usable result, but
>>> they really are working outside the specs and can get bitten at any
>>> firmware change (as was recently seen with some Garmin firmware).
>> Yes, you may have to recalibrate the offset after a firmware change. So?
> The Garmin issue was more than a simple "recalibration required", but the 
> released a fix.  "So" might be: "so use PPS as well as NMEA, both to 
> improve performance dramatically and reduce or remove the need for 
> recalibration.".

Well, no. As I understand it, the garmin would issue the nmea sentence
so that it ended after the second that it referred to ended. This made
the simpleminded association of the garmin nmea time with its PPS
difficult. It however did not stop you from using a say 1.1 sec offset
to correct the nmea time using the nmea driver. 
Ie, as I understand it, it was solveable by a simple recalibration. (
unfortunately the pps driver could not do that)

> Cheers,
> David 

More information about the questions mailing list