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

David J Taylor david-taylor at blueyonder.co.uk.invalid
Thu Feb 23 18:00:18 UTC 2012

>> - ground and +5V
>> - ground and RX
>> - ground and TX
>> - ground and PPS/DCD
> pairs.
> That is what I used and then got poor result.  (Yes I have a scope)  I
> figured the reason has that all the conductors called "ground" above
> where tied to gather at the DB9 shells pin-5 at each end of the cable
> So the return current divides.   You then don't have balanced twisted
> pairs.    You are right about it being a capacitor but really it's
> several caps that all interact.
> I think the better way to use cat-5 is like this
> RX, TX, DCD, ground
> With each of the above in a "solid" color wire. I think you get less
> capacitance that way
> Also while many RS232 ports will accept TTL signals, if they do they
> are operating way off spec and you are just lucky it works.  You
> really should be sure the PPS, TX and RX signals are at rs232 voltage
> levels.  out of spec voltages and long cables are a bad combination.
> I've converted to rg58 drives with 50R driver for the PPS and then
> convert is very close to the computer to RS232.  No ringing.
> Chris Albertson
> Redondo Beach, California


Thanks for that information - most interesting.

Could you please explain just how you are doing the wiring?  You are using 
the cable just as a four-way cable, and not using the second conductor of 
each pair, is that it?  And you aren't sending +5V up the wire?  Using 
separate earths which this implies may have caused the problems you 
describe, I only have one end of the system grounded here.  Unless you 
have a long run and fast edges, I don't believe that the capacitance 

Completely agree about using RS-232 levels wherever possible, and that's 
what I use on my Sure GPS Evaluation board, and the modification I 
described.  The slower RS-232 edges and specified receiver hysteresis are 
more tolerant of cable runs.

My gut feeling is that for the average stratum-1 user, correctly 
terminated RG58 and matched drivers and receivers for the PPS is beyond 
the precision level required.


More information about the questions mailing list