[time-nuts] Re: [tacgps] IC for Clock Input?

Poul-Henning Kamp phk at phk.freebsd.dk
Thu Sep 9 08:33:18 EDT 2004

In message <Pine.BSF.4.44.0409062049130.49792-100000 at king.cts.com>, John Kolb w
>On Mon, 6 Sep 2004, Brooke Clarke wrote:
>> I'm trying again to make a clock that will use the output from a
>> precision frequency standard (1, 5 or 10 MHz) as it's heartbeat.
>> At first I was going to use the 74HCT4046, like was done by Brooks
>> Shera, but I've discovered that the through hole version of this part is
>> out of production.  Brooks' has them in stock to support his board, but
>> I would like to find a new part.
>> In my case all the input IC needs to do is convert the input RF sine
>> wave into a clean TTL drive for the microprocessor.
>> Can I just use a Schmitt gate like the 74HCT14N?  To center the A.C.
>> coupled input a 68 Ohm resistor to ground and a 200 Ohm resistor to +5
>> Volts will bias the gate to the center of it's hysteresis and any
>> excursion greater than about 1 Volts (about 20 mw) will cause it to
>> switch.
>That should work as long as you have a low impedance source
>driving it. I would probably use a fast op-amp, however, to
>insure there's enough gain to square up the sine wave well.
>One of the op-amps that runs on a +5V supply only, and let
>it bang into the stops.

I have found that some amount of hysteresis is necessary to avoid
jitter around the zero crossings.  I tend to prefer LT1016 for
as clock buffer.

Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.

More information about the time-nuts mailing list