[ntp:questions] Performance?

dhavey dhavey at gmail.com
Tue Dec 30 21:40:05 UTC 2008


On Dec 30, 1:36 pm, dhavey <dha... at gmail.com> wrote:
> On Dec 30, 9:39 am, David Woolley
>
>
>
> <da... at ex.djwhome.demon.co.uk.invalid> wrote:
> > Unruh wrote:
> > > Garmin defines the leading edge as the transition from 0V to 5V on the PPS
> > > line. Now serial has two levels -12V and +12V. with capacitive coupling,
>
> > There's no capacitive coupling; it is assumed that the transmission line
> >   and load behave as capacitive, but that is in parallel.
>
> > The transition region for RS232C is -3V to +3V into 4 to 5k.
>
> > > the garmin signal would be something like -2.5V to 2.5 V which really is
>
> > However, the expected hysterisis is in the 10s to 100s of mV range, and
> > the actual threshold for control signals is supposed to be above zero,
> > such that zero volts will give an unambiguous off state.
>
> > In practice, therefore, with real world RS 232 line receivers, there is
> > no problem in feeding from TTL levels, as long as the cable is short
> > enough to avoid ringing.
>
> > > way out of spec for the serial port. (Or with the typicallysmall duty
> > > cycle, more like -1V to 4V)
>
> time.qnan.org seems to indicate that
> flag2 0=pps edge on assert
> or High to Low
> is standard behavior for GPS devices.



I don't see much difference in offsets between the machines either
way:
+192.168.0.29    .GPS.            1 u    3   16  377    0.090
3.084   0.254
*GPS_NMEA(0)     .GPS.            0 l    -   16  377    0.000
-0.276   1.722

+192.168.0.28    .GPS.            1 u   16   16  376    0.001
-2.921   0.124
*GPS_NMEA(0)     .GPS.            0 l   15   16  377    0.000
0.286   1.399

Something else must be wrong.




More information about the questions mailing list