[ntp:questions] Re: Bought a GPS receiver--What next?

David Schwartz davids at webmaster.com
Tue Nov 2 10:31:13 UTC 2004

"Wolfgang S. Rupprecht" 
<wolfgang+gnus20041101T201445 at dailyplanet.dontspam.wsrcc.com> wrote in 
message news:x7k6t4zxhe.fsf at bonnet.wsrcc.com...

> I always found that it was sad that the GPS interfaces didn't provide
> a simple 64-bit (or better yet 128-bit) cpu accessible register that
> simply provided the GPS's current internal time.  The PPS output is a
> pretty cumbersome output to try to use.

    It's not combersome to use with dedicated hardware. However, why not put 
that hardware in the GPS chipset where it, for all intents and purposes, 

> Does anyone have the ear of any of the GPS chip guys?  Now that
> computer buses are turning into high-speed serial buses, would it be
> possible to design a low pin-count GPS chip that dropped into the
> PCI-e or hypertransport bus and just reported the current time in,
> say, nanoseconds (or better yet 2^-32 seconds) since the GPS epoch?

    USB2.0 would probably be cheaper and easier to implement. It's plenty 
fast enough and latency is low enough. Accuracy would typically be measured 
in small numbers of microseconds and could easily be bounded at a few 
millseconds worst case. (And you would know how long the USB transaction 
took, so you would have a bound on the inaccuracy.)


More information about the questions mailing list