[time-nuts] IC for 1 PPS Output

[Tom Clark, K3IO (ex W3IWI)]

John Ackermann took my name in vain:

> Hi Brooke --
>
> The circuit I use is stolen from Tom Clark's TAC -- paralleling 
> several sections of a 74AC04 hex inverter through 47 ohm resistors.  
> For the current TADD projects, I'm using three sections that way, so 
> each chip gives me two outputs, and I use a 47 ohm SIP resistor pack 
> to keep the board space down.
>
> Tom came up with the 47 ohm value (though IIRC he used four sections 
> in parallel) based on compromising for maximum voltage into a 50 ohm 
> termination, and best rising edge.
>
> The 74AC chips don't have internal source resistors, but I think some 
> of the op amp distribution amp chips might.  For square waves at these 
> frequencies, I'd use 74AC parts over an analog circuit.
>
> John 
One of my desires was to have a source impedance BELOW 50 ohms so that 
the driver would push at least 3½ volts into a 50 ohm termination. So in 
later versions of the circuit, I have used as small as 22 ohms for the 
series resistors. The main reason for using series resistors to to allow 
the output current to be shared between all the gates; although I have 
seen people put CMOS gates directly in parallel, it never seemed 
aesthetically right.

The output impedance of an AC series gate (i.e. 74AC04) is in the 5-10 
ohm range. With several stages in parallel, the source impedance of the 
gate array is then ~10 ohms. So when feeding a 50 ohm load (with logic 
running at 5v), you will get ~5v*50/(50+10) = 4.16v which met my >3½ 
volts criterion. It also turns out that with 22 or 47 ohm series 
resistors, a shorted load will not cause the chips to become instant 
"crispy critturs"  ;<}

The 74AC gate array source doesn't need to be matched to the 50 ohms, 
since the terminated cable will have no reflections from the load.

Regards, Tom