[time-nuts] Buffering the 1pps output from a Z3801A

[Bruce Griffiths]

Didier Juges wrote:
> The Thunderbolt's output impedance on the PPS output is much less than 50
> ohms, more like 5 ohms.
>
> If you terminate the other end of the cable with a good 50 ohms termination,
> you will be OK, even though it would be better to put a ~47 ohms series
> resistor right at the TB output before driving the cable. This will reduce
> the available amplitude, but will reduce ringing even more.
>
> Check this page:
>
> http://www.ko4bb.com/Test_Equipment/CoaxCableMatching.html
>
> I used the PPS output of the Thunderbolt as a fast rise time pulse generator
> for cable matching/reflections experiments.
>
> It will show you what to expect when you use that output to drive a long
> cable. 
>
> Didier KO4BB
>
>   
A 5V PECL to CMOS level translator capable of driving a 50 ohm load is 
shown:

http://www.eds-fl.com/~bruce/5V_PECLtoCMOS_50_OhmDriver.html 
<http://www.eds-fl.com/%7Ebruce/5V_PECLtoCMOS_50_OhmDriver.html>

Input transistors can be 2N3907 although they will be a little slow 
since the transistor capacitances will be larger than usual because of 
the low collector base voltages, however the input stage propagation 
delay will only be a few nanosec. For a faster input stage use high ft 
low capacitance transistors (eg BFT92/93).

The values of the output series resistors can be changed to 240 ohms to 
provide a better match to a 50 ohm load if the load and consequently 
lower reflections termination has significant shunt capacitance.

Bruce