Building digital clock, need a clock pulse

[Eric]

I'm building a digital clock out of TTL 74xx chips (namely 7447, 7490,
7497, 7412), and I now need a clock pulse to run it. I was thinking of
using a quartz crystal resanator to run it, and I currently have a
32.768 Khz chip (which I would divide by 256 twice to get 1 hz) and a
1 Mhz chip I could also use. However I have tried finding a circuit
that will work with the crystals and can get none of the simpler ones
working, and don't want to take the time building a very complex one
unless I know it will work. If possible, could someone give me a link
to a circuit I can use to run one of these crystals. I am also open to
other suggestions as for different ways to get a clock pulse if
anybody thinks this is not the best method. Thanks is advance.

 

[John Fields]

I'm building a digital clock out of TTL 74xx chips (namely 7447, 7490,
7497, 7412), and I now need a clock pulse to run it. I was thinking of
using a quartz crystal resanator to run it, and I currently have a
32.768 Khz chip (which I would divide by 256 twice to get 1 hz)

---

32768 Hz
---------- = 128 Hz, so that won't work.
265

But,

32768 Hz
---------- = 1 Hz, so that will.
32768
---

 

[Tam/WB2TT]

Eric,

You must have a sine wave oscillator. The catalogs are full of oscillators
with TTL/CMOS outputs. Are you for real using 74XX logic? You should at
least look at 74HC/HCT. They (National) used to make single chip clock
circuits.

Tam

 

[Stefan Heinzmann]

---^---


---

32768 Hz
---------- = 128 Hz, so that won't work.
265

32768 Hz
-------- = 0.5 Hz, won't work either, but is closer ;-)
2 * 256

If you use real old TTL chips you need enough power to require a mains
connection, in which case you could use the 50/60 Hz line frequency as
the clock source.

 

[John Fields]

---

32768 Hz
---------- = 128 Hz, so that won't work.
265

---
Oops...

Misread the OP the first time, but 32768/256*256 = 0.5Hz, so it still
won't work
---

 

[markp]

I'm building a digital clock out of TTL 74xx chips (namely 7447, 7490,
7497, 7412), and I now need a clock pulse to run it. I was thinking of
using a quartz crystal resanator to run it, and I currently have a
32.768 Khz chip (which I would divide by 256 twice to get 1 hz) and a
1 Mhz chip I could also use. However I have tried finding a circuit
that will work with the crystals and can get none of the simpler ones
working, and don't want to take the time building a very complex one
unless I know it will work. If possible, could someone give me a link
to a circuit I can use to run one of these crystals. I am also open to
other suggestions as for different ways to get a clock pulse if
anybody thinks this is not the best method. Thanks is advance.

You could cascade two 74HC4060 chips to produce 1Hz, these are chips
specifically designed for use with 32k768Hz crystals and include 13 bit
counters internally. If you really want to re-invent the wheel and do it
discretely, use a CMOS unbuffered inverter such as a HEF4069UB or a
MM74HCU04:

http://he.fi/photo/ua5-dissection/sheets/mm74hcu04.pdf

Mark.

 

[Michael]

I'm building a digital clock out of TTL 74xx chips (namely 7447, 7490,
7497, 7412), and I now need a clock pulse to run it. I was thinking of
using a quartz crystal resanator to run it, and I currently have a
32.768 Khz chip (which I would divide by 256 twice to get 1 hz) and a
1 Mhz chip I could also use. However I have tried finding a circuit
that will work with the crystals and can get none of the simpler ones
working, and don't want to take the time building a very complex one
unless I know it will work. If possible, could someone give me a link
to a circuit I can use to run one of these crystals. I am also open to
other suggestions as for different ways to get a clock pulse if
anybody thinks this is not the best method. Thanks is advance.

Good for you! Use those 74xx chips (or CMOS functional equivalents).
They're a great way to learn. I built my first digital clock with 74xx,
back in 1972.

As for a clock source, use the voltage mains. Here in the U.S.,
half-wafe rectified secondary of the power supply transformer gives 60
Hz. Long-term stability of themains is excellent. Use some kind of
low-pass filter to block any hash, then divide-by-12 and divide-by-5
stages (a couple 7490's) to get your 1 Hz. That's what I did, and the
TTL clock with INCANDESCENT 7-segment display is still working all these
years later.

 

[David L. Jones]

Good for you! Use those 74xx chips (or CMOS functional equivalents).
They're a great way to learn. I built my first digital clock with 74xx,
back in 1972.

As for a clock source, use the voltage mains. Here in the U.S.,
half-wafe rectified secondary of the power supply transformer gives 60
Hz. Long-term stability of themains is excellent. Use some kind of
low-pass filter to block any hash, then divide-by-12 and divide-by-5
stages (a couple 7490's) to get your 1 Hz. That's what I did, and the
TTL clock with INCANDESCENT 7-segment display is still working all these
years later.

I'll second that.

Use the mains as a clock source, over the long term it is MUCH more
stable than a XTAL oscillator. A standard watch crystal will give you
something in the order of +/- 15 seconds per month or worse. My mains
clock (50Hz here in Australia) is accurate to tens of seconds a year
or better.
If you are powering your clock from a plugpack then use an AC plugpack
and tap off the AC voltage, clip it with some diodes and dropper
resistors and feed into your TTL divider chip. You might also need
some filtering as suggested.
My old TTL clock is still working 20 years on, and is still a talking
point, esp with it's 0.1 second display which ticks over ten times a
second. Non-Electronics types think that it must be super accurate!
:->
Mine also has an AM/PM display using another 7seg display ("A" or
"P"), and I have two set switches, one for fast and one for slow. One
really neat feature is that I use center-off switches and floating
4000 series CMOS inputs, so when I open the switch I get a third
"50Hz" "superfast" mode by picking up the 50Hz mains hum from the
switch wiring!

To lower power consumption I run my LED's at 0.5mA per segment and
they work great day or night. You have to find the right types though,
not all work off this low a current.

Forget all that rubbish about using FPGA's, CPLD's and
microcontrollers, they are sacrilege for this application!

Have fun.

Regards
Dave

 

[Michael]

... divide-by-12 and divide-by-5 stages (a couple 7490's) to get your 1 Hz.

Nope! Not 7490. That's a decade counter, of course. I don't recall
the number of the 74xx that I used but I do remember it's a ripple
counter.

 

[John Fields]

Nope! Not 7490. That's a decade counter, of course. I don't recall
the number of the 74xx that I used but I do remember it's a ripple
counter.

 

[Spehro Pefhany]

Nope! Not 7490. That's a decade counter, of course. I don't recall
the number of the 74xx that I used but I do remember it's a ripple
counter.

Whew, that was a LONG time ago.

7490: divide-by-5 and divide-by-2, can be used to make a divide-by-10
7492: divide-by-6 and divide-by-2, can be used to make a divide-by-12
7493: divide-by-8 and divide-by-2, can be used to make a divide-by-16

Best regards,
Spehro Pefhany

 

[Michael]

Whew, that was a LONG time ago.

7490: divide-by-5 and divide-by-2, can be used to make a divide-by-10
7492: divide-by-6 and divide-by-2, can be used to make a divide-by-12
7493: divide-by-8 and divide-by-2, can be used to make a divide-by-16

Best regards,
Spehro Pefhany

Yup. Too long for me. 7492 sounds right. But so does 7493. Darn
databook is a good 12 feet away, the casters on this chair don't work so
good, and my legs are worse!