stable clock timer..for a different LED clock

[dacflyer]

i am looking to build a rather crude strange clock
kinda sorta like the the big ben clock but less fancy...i been dreaming this way before the big ben was put into N&V...i like it tho..<p>anyway heres my idea,,,sounds good anyway...feel free to tell me of my flaws<p>i want to use 6 (4017 decade divider chips)
2 for a 60 sec count
2 for a 60 min count
and 2 for the 12 hr count<p>60 sec count will advance the min by 1
and 60 min will advance the hrs. by 1
make sence?<p>what i need is a stable counter/timer to drive the
seconds to make all this work..was considering using a 555 timer..but i do not know if its stable enough?? is there something better to use?
perhaps a circuit to suck on the 60HZ from house to keep time stable?<p>BTW..will be having the leds for this clock on one at a time,, not a arc like the big ben...leds will be in a matrex form sharing the 2) 4017's<p>please feel free to tell me your comments

[Edd]

About as simple as it gets….take your pick of the 28 LED model or the 72 LED model.
Of course there are many options on the time base/clock/ref..
http://ourworld.compuserve.com/homepage ... /clock.gif <p>http://ourworld.compuserve.com/homepage ... e.htm#menu
For the whole data spiel as well as the 72 LED unit… Is listed under LED circuits.
Of course if you had unlimited time you could go for broke and do:
http://www.luberth.com/analog.htm <p>
73's de Edd
:p<p>[ October 21, 2003: Message edited by: Edd Whatley ]</p>

[desterline]

A 555 might be useful for testing, especialy since you can run the freq arbitrarily high to see 'hours' in just a few minutes. But a 555 doesn't have the precision for a real clock.<p>Power line is a great place to get a freq standard for a clock. It's *very* accurate. Most plug in clocks use it (digital or mechanical). <p>The other option is a crystal. There's lots of options here that are 'prety good', but most profesionaly engineered crystal based clocks use a "32.768 kHz watch crystal". The're very small, should cost less than a buck, and are about the most precise 'standard' crystal you can buy.<p>32.768 sounds like an arbitrary freq, but it's actualy 2^14. So a 14 stage divide by 2 counter gets you a nice 1 hertz signal. Look around and you'll even find chips that are designed for just this purpose, they have the counter and oscillator parts built in. You just add the crystal and a 1 hertz square wave comes out the other side.<p>I'm drawing ablank on part numbers right now, If you can't find any, let me know and I'll look through my notes.<p>-Denny

[blueeyedpop]

if memory serves me correctly, the utilities adjust the 60 cycles to keep a proper number of counts throughout the day.<p>Mike

[dyarker]

For real reliability use the 60Hz power line for the frequency source, and divide a crystal oscillator to 60Hz. Then when the power goes off, batteries provide backup power, and a missing pulse detector switches in the crystal pulses until 60Hz power returns. Putting battery power to the LEDs through a push switch extends battery life when the lights are out.

[dacflyer]

i am not worried about power failures..but if someone can provide me a schematic for the line powered 60Hz timebase that would be great..
like i said i want to keep this simple..<p>and i will be using 2 circles of 60 leds = 120
and 12 segments for the hrs...
and when i am done, i will be happy to share this with everyone...

[desterline]

I don't have a schematic handy, but it's not complicated so I'll try and describe it.<p>Basicaly you need to connect to the secondary (low voltage) side of your power transformer with a single diode. That will provide you with a 1/120 sec pulse every 1/60 of a sec - half wave rectification.<p>Then the pulse is fed through a resiston divider to ground to give you a voltage that'll work with your logic chips (TTL-74xx series 5 volts, CMOS 40xx series no more than Vdd) and push it into a buffer. A schmidt trigger is better, but I've seen a regular buffer work ok.<p>Hope that helps. <p>
BTW, here's some trivia for you. What is the technical definition of a "Jiffy"?<p>It came about to describe the time of one power cycle at 60 Hz. So it's 1/60 of a sec.<p>-Denny

[i_myself_matthew]

I once tried to derive 60hz by desterline's method, but for some reason it never would work. What I ended up doing was use a 555 chip to basically amplify the ripple peaks on the unregulated side of a 7805. It took some planning, and I needed a scope to get the 555 to trigger right, but it worked (until the high voltage part of the project found a way to the low voltage supply). The final result was a nice, clean pulse at logic level. It is more complex, so if you are interested, e-mail me, and I will draw up a schematic.<p>Matthew

[dyarker]

Instead of resistor divider, use resistor to a 5.1V or 4.7V zener to ground after the diode connected to a transformer lead (low voltage side of course). That gives squarer pulses. A low value capacitor in parallel with the zener filters power line noise and thermal noise generated by the zener.<p>If you are using a DC wall wart for power, then you have to use i_myself_matthew's 555 circut. Because the only 60Hz signal is ripple.<p>I can do a schematic. But what is the secondary voltage of your transformer?

[Bernius1]

Sorry, I can't help the thread. But Mike, That postscript is hysterical. I laughed for 10 minutes.
The only clock I've wanted to copy is one (avail.from Germany) which is like a ceiling fan mounted sideways on the wall. LED's on the 'blades' blink in phase, so it appears that a regular clock face is floating in air, with hour/min./sec. hands, the 12 numbers, and "Space Invaders" crawling around the perimeter. Wild.

[desterline]

All those alternatives are plausable. I was just tossing out the simplest one I could think of.<p>The zener diode design is better as it provides a safety against high voltages into your logic, and of course reverse voltages. But square-er pulses realy isn't a problem if the signal is run through a schmitt trigger, or a comparator.<p>With careful design I'm sure you could also pick the ripple off a supply with a comparator too (instread of a 555) <p>My concern with either of those methods though is they're more dificult to make tolerant of input voltage changes. <p>Good luck
-Denny

[dyarker]

The zener circuit is absolutely tolerant of input variation. Zero to 5.1V to zero every time.

[k7elp60]

I see several approaches for the clock. If you use a 32.768Kc crystal and a 4060 IC and divide with the maximum 2^14(16,384) the ouput would be
2Hz. Use 1/2 of a 4013 as a D flip-flop to divide by 2 you would have 1 Hz just right. You could use a 4017 to divide by 2 also.
I have used the 555 as an stable precision oscillator, but at low frequencies it is pretty hard to get a capacitor that doen't have leakage current to affect the timing. What I have done is set the 555 up to a higher frequency and then divide the output down to the frequency I want. As long as the external components are stable, the 555 is very stable and won't change frequency
because of temperature and power supply fluctuations. I find by using a good quality film capacitor the 555 is very stable. You may consider setting the 555 to 4.096Kc and dividing it with a 4060 by 2^12(4096). If you have a frequeny counter that will be easiy. If you want to use the 60 cycle line freqency and divide by 60 another approach would be to use a op amp set up as a comparator and the output would be a fast rise time square wave to trigger your 4017 counters.
Ned

[dacflyer]

hey i think i got what i want... ADVICE>>><p>how bout a 4024 i salvaged from a traffic signal FLASHER...blinks wig wag style,,this flasher controlled 2 120 volt circuits,,triacs fired by opto isolators..,at about a 50-50 duty cycle,,,,
this works off 120v directly resistered to logic level...and the out put comes off of pin 4. Q6
since it has 7 out puts i should be able to adjust it???
someone let me know

[dacflyer]

still waiting for a comment :p