Sometimes, the parts list says it all. 777 transistors, 1223 resistors, 136 LEDs, 455 crimp connectors, 41 protoboards and 500 grams of solder. That’s what went into this transistor logic clock build.
While additional diodes and capacitors were tolerated in this project, a consequent implementation of a discrete transistor logic clock, of course, does not contain a quarz oscillator. Instead, it extracts its clock signal from the mains frequency in its power supply. Because mains frequency is slow, it can be stepped down to a clock-applicable 1 Hz by a simple counter unit which already spreads its discrete transistors across 4 protoboards.
In total, 28 Flip Flops were assembled on individual boards. Most of them went into the counters for hours, minutes and seconds. The counters are orchestrated by reset boards that know how far each counter shall count and reset them when necessary while incrementing the next counter — a transistor-based overflow interrupt. The output of the counters feeds into a multiplexer (3 boards), clocked by a 300 Hz timing circuit based on a 555 timer, also implemented from discrete components. Eventually, a 7 segment decoder (2 boards) sends the numbers to the displays.
Fine tuning the clock speed may require a quick call to the electricity provider isn’t even necessary due to the long term stability of the mains frequency maintained in large parts of the world (Thanks Tom!). The end result is a beautifully crafted, modular transistor logic clock!
This is good news for those of you who have been coveting the MOnSter6502 project. Practice your SMD transistor PCB layout with this project before you try to recreate that monster of a board layout project.
Thanks to [BB] for the tip!
“Fine tuning the clock speed may require a quick call to the electricity provider, but the end result is nevertheless a beautifully crafted, modular transistor logic clock!”
We’ve had this discussion before: The mains frequency is highly accurate when averaged over a couple of days, explicitly because a lot of clocks use it for timekeeping.
I thought USA were abandoning long term frequency compensation
Maybe one of Trump ideas : because clocks are used by terrorists.
I’m not a fan of the man, but let’s keep unnecessary political rhetoric out of HaD, please.
I agree.
Same here. All political(ed) out.
I guess some people don’t like sarcastic humor, but I do.
Eureka, you’re right! Thanks!
that’s a big fan for a 700mA 5V supply.. very nice project, must have kept someone busy for quite some time.
Well, it wasn’t being used for anything else presumably. You should see the fans I have in my junk collection!
Hahaha! Ridiculously cool build, but I have a suggestion. Now that you’re certain the circuit works…
Do it dead-bug-style!
I honestly would probably buy it from you if that were to happen.
Great job. My favorite image is the back of the multiplexer, the blue jumpers are very neatly curved.
I’m pretty sure he did know that it work just fine. According to his site this is a copy of the big-pcb transistor clock…
” Another transistor clock – This is a copy of the Kabtronics transistor clock. http://transistorclock.com/ “
Why would you want it in dead bug style?
I have a thing for it. Most are terrible; but done right, magic!
Very, very nice. Nice construction, nice presentation. And das blinkenlights! I’m a big fan of status LEDs on each flip-flop. But “of course” no crystal oscillator? Come on now, seven flip-flops were spent on dividing 50 Hz down to 1 Hz. It would have only added eight more to divide down from 32,768 Hz, plus the oscillator itself. But then there would have been a number of flip-flops running too fast to see, so maybe this was for the best. The fully-discrete 555-equivalent was the frosting on the cake, but maybe overkill – a free-running multivibrator could have been done in four transistors, including buffers.
> But then there would have been a number of flip-flops running too fast to see […]
Now, if you can see 300 Hz (heck, if you can see 37.5 Hz!) I’d like to have a talk with you. What’s your optical bandwidth?
:-P
It’s not as difficult as you may think: take a LED that blinks at that frequency and move it really fast with your hand, preferably in the dark. You’ll see spots instead of a continuous line. You can even roughly calculate the frequency based on the apparent size of the dots and the speed of your hand.
Long term stability is the word here. A crystal would drift, not the mains ;)
> free-running multivibrator could have been done in four transistors
You would only need 2 transistors or 1.
http://www.sentex.ca/~mec1995/tutorial/xtor/xtor7/xtor7.html
could have done it with 555 transistors
</sarcasm>
After a second look –
Wow, he even made a discrete 555
Mains may be longterm stable but in the short term it’s frequency can vary quite a bit. [ For a clock anyway. ] He might want to bring down the frequency of his discreet 555 and maybe add a temperature sensor to trim the frequency.
Even then, the mains frequency tolerance and stability is going to be much better than what you get out of uncalibrated discrete 555 timer circuit. The transistors are not going to be matched as closely as you’ll get inside a chip, so all bets are off on the voltage/temperature sensitivities.
Before he can do any temperature compensation, he has to characterize how the timing varies with temperature.