Re: [gpsd-dev] RFC 2783

[Hal Murray]

> In Eric's writeup "Introduction to Time Service", he lists the 1PPS via USB
> as having a deviation of up to 1ms vs. 1PPS via Serial at only up to 5uSec.
> The USB GPS's I have will have THAT much latency on the PPS? That's a factor
> of 200, USB vs. Serial!  If so, I may shy away from using the PPS on the USB
> GPS and get another unit from adafruit... 

USB is polled.  For old/slow USB, the polling interval is 1 ms.  I think it's 
1/4 ms for newer, but all the GPS over USB that I've seen are old/slow.

If you are lucky, the "average" of several samples is 1/2 that.  If you are 
unlucky, the USB clock will be synchronized (or close to synchronized) to the 
PPS clock and all your samples will be at one end or the other of the 
sampling interval.

Geek alert.  The term is Hanging bridges.
  http://www.leapsecond.com/pages/m12/sawtooth.htm





-- 
These are my opinions.  I hate spam.