[ntp:questions] PPS and NMEA

unruh unruh at invalid.ca
Tue Sep 11 00:01:26 UTC 2012

On 2012-09-10, Richard B. Gilbert <rgilbert88 at comcast.net> wrote:
> On 9/10/2012 5:02 AM, Rob van der Putten wrote:
>> Hi there
>> ksprabha wrote:
>>> Kindly let me know what is the difference between NTP with PPS and NTP
>>> with out PPS.
>> NMEA can be a bit off with some GPS receivers. PPS tends to be very
>> accurate.
> The the PPS (assuming we are talking about the Pulse from a GPS Timing 
> Receiver) is specified to be within plus or minus 50 nanoseconds.

That depends on the gps receiver. The Garmin 18(x) talks about 1usec.
The pulse also depends on the route to the computer ( light travels
through the coax at about 20cm/ns so a 10 m run is about 50ns late). It
also depends on the relation between the input impedence of your
computer, vs the cable impedence, and the output impedence of your gps
receiver. (if it is high, then the cable is like a capacitor, and it
takes time to charge up, the longer the more time it takes).
Also if you put pulse conversion (ttl to RS232) it can also take many
many ns. But the worst thing is the interrupt handling on your computer.
It is liable to take 1-10usec variable. 
Thus do not expect better than 1us from the pps. 

Thus, ntp with pps will discipline your computer's clock to a few usec,
with nmea alone, 10s of msec. (ie 10000 times worse).


More information about the questions mailing list