RE: Help needed with USB GPS device

[Gavin Davenport]

Some interesting stuff here:
https://austinsnerdythings.com/2021/09/29/millisecond-accurate-chrony-ntp-with-a-usb-gps-for-12-usd/

I'll get the chrony stats file into splunk and start peering at offsets.

Thanks!

-----Original Message-----
From: gpsd-users-bounces+gavdav=gavdav.net@nongnu.org 
<gpsd-users-bounces+gavdav=gavdav.net@nongnu.org> On Behalf Of Gavin Davenport
Sent: 03 March 2024 14:15
To: gpsd-users@nongnu.org
Subject: RE: Help needed with USB GPS device

I recognise my questions are now really about chronyd config and I'm curious 
how I do this:
>> some indication of how well your GPS receiver is working as an NTP refclock.

What am I looking for ?

I have
refclock SHM 0 refid GPS precision 1e-2

and after restarts and a wait....
MS Name/IP address         Stratum Poll Reach LastRx Last sample
===============================================================================
#* GPS                           0   4   377    21  +1517us[+1853us] +/-   13ms

I don't really know how to interpret the "Last" and "sample" fields here. Is 
this a representation of how old the sample is, or how long it took to get it ?

Really apprecate the help


-----Original Message-----
From: gpsd-users-bounces+gavdav=gavdav.net@nongnu.org 
<gpsd-users-bounces+gavdav=gavdav.net@nongnu.org> On Behalf Of Bryan 
Christianson
Sent: 02 March 2024 22:38
To: gpsd-users@nongnu.org
Subject: Re: Help needed with USB GPS device

Hi Gavin

> On 3/03/2024, at 11:11, Gavin Davenport <gavdav@gavdav.net> wrote:
> 
> I'm now using this approach in chrony
> refclock SHM 0 refid GPS precision 1e-1 offset 0.9999 delay 0.2 
> refclock SHM 1 refid PPS precision 1e-7
> 
> And chrony can see the GPS source as follows:
> MS Name/IP address         Stratum Poll Reach LastRx Last sample
> ===============================================================================
> #x GPS                           0   4   377    18   -964ms[ -964ms] +/-  
> 200ms
> #? PPS                           0   4     0     -     +0ns[   +0ns] +/-    
> 0ns
> ^+ 2001:8b0:0:23::205            2   6    77    24    +23us[  +75us] +/-   
> 19ms
> ^- 2001:7b8:3:2b::4              2   6    77    25  +3106us[+3158us] +/-  
> 334ms
> ^- 2001:7b8:3:2b::5              2   6    77    24  +2865us[+2917us] +/-   
> 53ms
> ^- 2001:418:3ff::1:53            2   6   377    21  -1315us[-1263us] +/-   
> 96ms
> ^+ 2a05:b400:c::123:63           2   6   377    19   -296us[ -245us] +/- 
> 6816us
> ^* 131.111.8.63                  2   6   377    18   -437us[ -386us] +/- 
> 7137us
> ^+ 2a05:d01c:9d2:d200::be00:5    2   6   377    17   -370us[ -370us] +/- 
> 5314us
> 

You have specified offset 0.9999 delay 0.2 in your GPS refclock. This is why 
you are seeing the ~900ms offset in the chronyc sources output.
These attributes are used when you have a valid pps source, but are totally 
skewing the results in your case.

I would comment out the other ntp sources in the chronyd configuration and 
remove the offset/delay attributes from the GPS source. This will then give you 
some indication of how well your GPS receiver is working as an NTP refclock.

Please note - this is more a chrony configuration issue rather than gpsd. 
Please see https://chrony-project.org/documentation.html
gpsd is doing the best it can with your hardware.

Bryan Christianson
bryan@whatroute.net