My ttl clock is gotten to the point where I need to get an accurate one second pulse to clock the 74s390 counter. I have been using a 74124 VCO and it works fine for testing but I think it's not too accurate. I ran into this idea of using a discarded clock.

http://www.josepino.com/?one_second_timebase

I got the parts and built a circuit (I used a different transistor than that author suggested) that can blink an LED. I can't figure out how to get this pulse to clock the 74s390. Hooking up from the e or c of the transistor to the 74s390 clock input doesn't do anything. One thing I noticed is that my circuit flashes the LED for a much shorter time than the 74124 did. Is that my problem? How can I connect the pulse to the counter input and how can I adjust it to work? Thanks.

 

The cheapo clock guts with the watch crystal will be a lot better than an RC oscillator, but still could loose or gain a few seconds per day. Using the 50/60Hz power line driving a "zero-crossing detector" followed by counter is actually better than an untrimmed watch crystal.

Connect the clock input to your one-second counter to the collector of the NPN transistor. You might have to ditch the flashing LED, or parallel it with a 10KΩ resistor. The collector will be low (~0V) during the wired-or pulse from the motor outputs, and will be high (~5V) at other times. If you were to look at it with a scope, it would look like an inverted clock signal, but since your one second counter likely responds to positive-going edges, it will count just fine...

 

I hooked the emitter into a 7414 schmitt trigger input and then the output to the 74s390. Works great! Clocking away like a champ. I would never had known to even try that without this board. Working and I'm onto the next step, to get a 10 led bar graph to count up the seconds until the minute ticks over. Thanks All.

 

MikeML, I didn't want to use a 60hz reference from the mains because I have a 9V 500ma DC wall wart that I can feed through a 5v regulator to run the whole show. And this is a zero budget project that the rule is to use things I already have. I admit I cheated after I burned out the first clock mech I had and spent another dollar for the one I'm using. It's all about learning for me, and I find setting limits can help. Thanks for the suggestions, I am learning.

 

Add a diode in series with the 5V regulator upstream of the input filter capacitor. This will give you a 120Hz waveform. Feed that into your Schmitt trigger gate, and into a counter chain to get a divide by 120. Now you have a one second pulse that will be accurate to better than one second per month, year, 10 years,