[ntp:questions] Oddities in termination of cable from gps18.

unruh unruh at invalid.ca
Thu Feb 23 16:06:24 UTC 2012

On 2012-02-23, David J Taylor <david-taylor at blueyonder.co.uk.invalid> wrote:
> "Uwe Klein" <uwe at klein-habertwedt.de> wrote in message 
> news:1nbh19-jvp.ln1 at klein-habertwedt.de...
> []
>> Is the 110Ohm twisted pair the only connection to your pps sink ?
>> If you connect one conductor to GND on both sides you just introduce
>> some funny groundloops.
>> i.e. the assymetric source driven transmission line works as a baloon
>> and produces a symmetric signal at the other end. If you short circuit
>> that to gnd on one conductor...
>> uwe
> What you say is true, but only when the cable is a significant fraction of 
> a wavelength.  For the relatively short lengths and ~200 ns (or slower) 
> edges being discussed here it's unlikely to matter.  For the ~1 us 

The edge is much faster than 200ns. It is at the 20-50ns level which is
less than a wavelength away.  From the
Sure, it is at the 5ns level.

> accuracy of the PPS edge likely it would be better to simply slow the edge 
> down to avoid any ringing effects (and would the ringing matter in any 
> case when the receiver has an RS-232 specification, likely including 
> hysteresis? ).

The receiver is the ACK pin on a parallel port, not on a serial port.
And 1 V is just on the hairy edge seemingly without hysteresis. 

> Different if you're chasing a 50 ns difference, of course.  You try to do 
> the job properly then.
> Cheers,
> David 

More information about the questions mailing list