David Woolley wrote:
[]
> The problem with USB latency is that it is unpredictable.  The
> uncertainty in the direct serial port latency is 1/16 th of the
> signalling unit length (about 6 microseconds at 9600) plus the
> uncertainty in the interrupt and process scheduling latencies.  (If
> the UART clock is free running on the GPS device, you might need to
> add another 1/16th, so about 12 microseconds.)

Thanks, David.  As NTP deals well with the unpredictable latency on the 
internet, it's not impossible that a USB/GPS device could produce 
significantly better results than a consumer internet connection alone.

I'd like to see actual NTP results.

Cheers,
David 

_______________________________________________
questions mailing list
[email protected]
https://lists.ntp.org/mailman/listinfo/questions

Reply via email to