[ntp:questions] ESR looking for good GPS clocks

unruh unruh at invalid.ca
Tue Mar 6 16:42:16 UTC 2012

On 2012-03-06, Ron Frazier (NTP) <timekeepingntplist at c3energy.com> wrote:
> On 3/6/2012 4:45 AM, Rob wrote:
>> unruh<unruh at invalid.ca>  wrote:
>>> He is suggesting that the on board gps chip have ntp on board, so that
>>> your computer timestamps the request, the gps chip then timestamps when
>>> that trequest was received and returns it. No gps chip that I know of
>>> has an onboard ntp server. So if you want to get a chip fab to make you
>>> a special purpose chip which contains a gps receiver and ntp onboard,
>>> you could probably do it-- but it might cost a bit more than $100 ieach
>>> in
>>> lots of 100.
>> Of course not.  NTP is software, not hardware.
>> The only thing you need is special firmware for any existing GPS chip
>> that includes the NTP functionality you want, and maybe drops some other
>> functionality to make room in the flash.
>> Some manufacturers provide a software development kit that allows the
>> user to write his own code, link that with the object modules written
>> by the manufacturer, and create a new flash file for the GPS device.
>> At least SiRF provides such a development kit to its commercial customers.
> I haven't been following this thread extremely closely, but I did read 
> ESR's blog post and exchanged a few emails with him.  He says SIRF GPS's 
> exhibit a "wobble" of variance of outputting the NMEA time of 170 ms or 
> so.  I've personally observed with my USB BU-353 this effect, where the 
> Meinberg Server Monitor shows my computer locked into the GPS time 
> within a few ms and a number of internet servers that I have programmed 
> will show offsets of say 50 ms.  However, when I originally set the 
> fudge factor, I had almost all the internet servers showing single digit 

One does NOT use nmea to do accurate time. One uses the pps. 

The suggestion above is that one reprogram the onboard computer on the
gps chipset to act as an ntp server. Assuming that the computer can
reliably turn around an ntp request in a minimal time, and not interfer
with the handling of the gps sattelite signals, and that the usb
translation of the signals does not suffer too long a delay, the
postulate is that one can get us accuracy from the chipset. 
This is NOT using the nmea sentences.

> offsets when the PC is locked to the GPS time.  Sometimes, it shows a 
> positive offset for the internet servers and sometimes it shows 
> negative.  Anyway, ESR says SIRF GPS's aren't suitable for timekeeping.  
> I have not done anything to corroborate his statement, other than 
> observing this strange behavior.  I'm going to start a new thread 
> talking about what I've seen in more detail, since it's really a 
> different topic.  I just mentioned it since you brought up SIRF.
> Sincerely,
> Ron

More information about the questions mailing list