Hi:
I want to obtain a 1Hz Clock From A 10Mhz Oscillator Module. I would like to use
CD4000 series CMOS since I have a good supply of these. I can't really determine
which chip(s) to build the frequency divider chain with. Could someone please advise?

Thanks,
M

 

I want to obtain a 1Hz Clock From A 10Mhz Oscillator Module. I would like to use
CD4000 series CMOS since I have a good supply of these.

Even with a supply voltage of 15V, CD4xxx aren't guaranteed to operate at 10MHz.

From Fairchild:

 

You would have to start your divider chain with some part that are not from the CD4000 series. I would recommend the 74LVC1G to do at least a dive by 8. That will get you to 1.25 MHz. which should be within the capacity CD4000 series CMOS parts with a 5V supply. You still have a problem with the divisors from there because 1.25 MHz is not a multiple of so you are going to have to find the prime factors of 1.25E+06 and proceed from there to get exactly 1 Hz.

\( 1,250,000\;=\;2^4\cdot5^7 \)

You're going to need some divide by 5 stages to go with the divide by 2 stages. Good luck with that one.

 

Even with a supply voltage of 15V, CD4xxx aren't guaranteed to operate at 10MHz.

From Fairchild:
View attachment 313538

Thanks Dennis. I guess I should get a 1Mhz oscillator module. M.

 

I guess I should get a 1Mhz oscillator module.

If you're willing to use a lower frequency module, why not get 32.768kHz?

 

If you're willing to use a lower frequency module, why not get 32.768kHz?

Thanks again. Yes that sounds good. M.

 

Start with an 74HCT390, dual high speed CMOS with TTL-threshold for the first stage. It will divide down to 100Khz, which should be sufficient for standard CMOS at 5 V.
It is very important that you procure the H C T variant. All the 10Mhz sources I am aware of, will output a 3V waveform, which is not sufficient to reliably clock a CMOS-threshold counter.

 

Start with an 74HCT390, dual high speed CMOS with TTL-threshold for the first stage. It will divide down to 100Khz, which should be sufficient for standard CMOS at 5 V.
It is very important that you procure the H C T variant. All the 10Mhz sources I am aware of, will output a 3V waveform, which is not sufficient to reliably clock a CMOS-threshold counter.

Thanks!

 

The problem with a 32 kHz crystal is that there is no way to get a 1 Hz signal without using two chips. Do you have a CD4521? This, plus a higher freq crystal, gets you 1 Hz in 1 chip.

If not, then still consider a higher freq crystal than 32 kHz. A 32.768 kHz crystal is not a true crystal. It actually is a tuning fork, and is less accurate and less stable than a real crystal, especially over wider temperature ranges. A CD4060 (oscillator plus divider) plus any of the other multi-bit dividers (CD4020, 4024, 4040, etc.) lets you use a 1.048 MHz or 2.097 MHz true crystal for better accuracy and stability.

ak

 

If you're willing to use a lower frequency module, why not get 32.768kHz?

Is there a spice model for a 32.768 Khz crystal by any chance? I like to simulate
a circuit prior to breadboarding it.

M

 

Is there a spice model for a 32.768 Khz crystal by any chance?

Yes, there is one

 

Yes, there is one
View attachment 313552

Thanks. I can't find this library.

 

Thanks. I can't find this library.

Are you familiar with the collection from @Bordodynov ?
You can find it here:
For LTspice users. Libraries of models, examples, etc (ltwiki.org)
It is the top link and about a 20Meg file

 

Are you familiar with the collection from @Bordodynov ?
You can find it here:
For LTspice users. Libraries of models, examples, etc (ltwiki.org)
It is the top link and about a 20Meg file

Thanks I found it.

I set up the attached sim but I get an error message "Could not open library file 'xtal_par'.sub.
Not sure how to fix this.

 

Is there a spice model for a 32.768 Khz crystal by any chance? I like to simulate
a circuit prior to breadboarding it.

Can't help you with that. I rarely use circuit simulators. I prefer to do them in my head and breadboard instead of spending time trying to make simulations work.

 

The problem with a 32 kHz crystal is that there is no way to get a 1 Hz signal without using two chips. Do you have a CD4521? This, plus a higher freq crystal, gets you 1 Hz in 1 chip.

If not, then still consider a higher freq crystal than 32 kHz. A 32.768 kHz crystal is not a true crystal. It actually is a tuning fork, and is less accurate and less stable than a real crystal, especially over wider temperature ranges. A CD4060 (oscillator plus divider) plus any of the other multi-bit dividers (CD4020, 4024, 4040, etc.) lets you use a 1.048 MHz or 2.097 MHz true crystal for better accuracy and stability.

ak

There is actually. I got a CD4536B Programmable Timer chip to give a 1Hz output recently using pins 3 and 5 in the standard CD4060 type oscillator circuit, although it needed slight changes to the resistor values (a 1M feedback and a 330k in series with the Xtal), and it ran quite happily, but needing about 4.5V minimum to start reliably!

 

Can't help you with that. I rarely use circuit simulators. I prefer to do them in my head and breadboard instead of spending time trying to make simulations work.

Thanks Dennis I am going to go with 32.768 Khz.

 

There is actually. I got a CD4536B Programmable Timer chip to give a 1Hz output recently using pins 3 and 5 in the standard CD4060 type oscillator circuit, although it needed slight changes to the resistor values (a 1M feedback and a 330k in series with the Xtal), and it ran quite happily, but needing about 4.5V minimum to start reliably!

I don't have a CD4536B or a CD4521 but I can order them once I have a large enough order.
I was actually trying to get a datasheet for the CD4536B but Digikey doesn't have it on their site.

 

I was actually trying to get a datasheet for the CD4536B but Digikey doesn't have it on their site.

Here's the Motorola datasheet.

 

Here's the Motorola datasheet.

Thanks.

Here's the Motorola datasheet.

Thanks!