[time-nuts] Thunderbolt meets ntpd

Tim Cwik timenuts at stnhbr.com
Mon Jun 23 08:36:23 EDT 2008


Chris Kuethe wrote:
> On Sun, Jun 22, 2008 at 11:30 PM, Hal Murray <hmurray at megapathdsl.net> wrote:
>   
>> Does anybody have any Linux code to print the stuff it's sending?  If not,
>> I'll probably have something soon.
>>     
>
> 1) dig around on trimble.com - there's a link to iQSource.zip which
> may be useful
> 2) rip some code from gpsd - we can decode a fair number of the
> packets, and there's a lot of machinery you can reuse.
>
> I think I ordered early enough that sometime in the coming weeks a
> thunderbolt will be headed my way, after that i can see about making
> gpsd do the right things. already i know i need to do auto parity
> detection - we assume TSIP is 8O1, which may not be true in the case
> of the thunderbolt. Given some of this past week's list traffic, I
> think Trimble goofed if they decided to go to 8N1.
>
>   
Hi Chris,

Trimble is using 8N1. Patching the code as you suggested works. To keep 
gpsd as config free as possible, I suppose you will need the auto parity 
detection.
I used the patch Wayne Knowles provided for decoding the super packet 
and that allows gpsd to detect when it has a fix.
The PPS output from the Thunderbolt I have is a short negative going 
pulse. The documentation says the pulse is configurable but I could not 
find a way to do this using Tboltmon. Instead, I purchased a FATPPS kit 
from TAPR which is able to stretch the PPS pulse and can also invert the 
polarity. Using the FATPPS, your parity correction, and Wayne's patch, I 
now have ntpd using the PPS from the Thunderbolt.
ntpd rejects the GPS clock itself. I am still trying to understand this. 
I hope when your unit arrives, the fix will be obvious to you. Here are 
my results:
ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 15949  9314   yes   yes  none   outlyer   reachable  1
  2 15950  9414   yes   yes  none  candidat   reachable  1
  3 15951  9314   yes   yes  none   outlyer   reachable  1
  4 15952  9414   yes   yes  none  candidat   reachable  1
  5 15953  8015   yes   yes  none    reject  clock expt  1
  6 15954  9614   yes   yes  none  sys.peer   reachable  1
  7 15955  9014   yes   yes  none    reject   reachable  1

ind assID status  conf reach auth condition  last_event cnt
===========================================================
  1 15949  9314   yes   yes  none   outlyer   reachable  1
  2 15950  9414   yes   yes  none  candidat   reachable  1
  3 15951  9314   yes   yes  none   outlyer   reachable  1
  4 15952  9414   yes   yes  none  candidat   reachable  1
  5 15953  8015   yes   yes  none    reject  clock expt  1
  6 15954  9614   yes   yes  none  sys.peer   reachable  1
  7 15955  9014   yes   yes  none    reject   reachable  1

Thanks for all of your help through several versions of gpsd. I am quite 
excited to finally have this sort of time reference available at home!

Best regards,
Tim



More information about the time-nuts mailing list