Dividing a 32768 Hz crystal frequency

You can make an oscillator out of pretty much any inverter. I've not seen it done, but I expect you can use the reset input and Q output of the 4013, which forms an inverter.

Just be aware that CMOS has little output drive and is very sensitive to voltage spikes. If you have any external input or output signals to these parts, they need to be well protected.

mandag den 26. juni 2023 kl. 16.20.48 UTC+2 skrev John Woodgate:

get a 32768 Hz oscillator instead of a crystal, saves alot of headaches

Yes. Oscillators made from a crystal and a chip and some caps and maybe a resistor, tend to not oscillate. And certainly won't be PPMs on-frequency, as a purchased oscillator usually is.

Maybe play with the math and use some other oscillator frequency and some easy divisor?

If you want to divide by twenty you need five stages of division. The CD4096 contains only two bistables, so it won't do it

A 4017 is a divide by ten, so if you use one of the J-K bistables in in the 4096 as a divide by two you've got you divide by 20 and can use the other half to drive your oscillation.

A 4018 is more flexible.

A small programable logic device could do it in a single packages, but you have to work out how to program it, which isn't all that difficult.

I was fond of the Philips cool-runner parts about twenty years ago, and they seemed to get even easier to use when Xilinx took them over, but I never bothered to work out how to program them from my home computer (which looked as it it would be easy enough).

There's probably a chip which does exactly what you want but I've given up reading the Farnell catalogue thoroughly enough to know what it might be today.

Lasse is probably right about the 32768 oscillator. It won't be as cheap as the crystal on it own but Farnell has lots

is one of many, and fairly cheap at about $1 each in small quantities. For test gear. why mess around making something you can buy?

Yeah 32768/20 seems like kind of weird reference frequency, anyway. 1 kHz is easy from a 4.096 MHz crystal and a 4060.

Yes, use a 74HCxxxx if it's available unless there's some pressing need to use the old CD series, hard to think of a good reason to though if there's a silicon-gate equivalent.

I'm not sure what you mean by 'a 32768 Hz oscillator'? Part number? Web site?

I think I'd be very lucky to find a better combination of crystal frequency and division ratio. But if you could suggest a simple way of getting closer to 1591.55 Hz I would be very interested.

Thanks, Bill, I will investigate,

Those clock/watch crystals are good, but only at wrist temperature; 100ppm frequency variance is close enough for a wall clock.

The 'component bridge' application, though, needs frequency purity, which a CMOS clock does not have; you'll get 32 kHz square waves, not a pure sinewave.

There's ways (''magic sines') to do a sinewave synthesis, but in the end, you'll undoubtedly want to have a trimmed LC filter in addition to a crystal oscillator, and maybe a bit of gain control.

Phase-locking an XR2206 to a crystal clock kinda works, too.

7490? Div by 2, and 10 Bit, rusty. But I've used 7490 and 4013 to get odd divisors in the past.

Cheers

You can buy a programmed XO and divide by something easy.

Or, Digikey has a zillion standard XO frequencies starting at 1 Hz. You could take your frequency and keep multiplying by 2 until you hit something interesting.

CD4059/HC4059 is a divide-by-N, which could be used after one of those many XO choices. Add a flop if you need a square wave.

something like these just supply power and you are done,

all SMD but it is only 4 pins

I'd just use a PIC or similar running a simple loop toggling an output pin. The eight pin parts (always?) have Xtal oscillators. Some have Numerically Controlled Oscillators which might be handy instead of the simple loop.

eg PIC16(L)F18313/18323

As others have said, watch crystals are good at watch temperatures, not so good otherwise, so a better spec higher frequency may be the way to go. Or if you want even more betterer, a GPSDO will give a stable 10MHz if the antenna can see satellites.

Hey, John,

Nice to see you back on SED, man! An HC40103 will do it, if you don’t mind a 5% duty cycle.

Cheers

Phil Hobbs

Plus a 74HC1G04 or something for the oscillator.

Cheers

Phil Hobbs

Thanks, Phil. I left SED when it turned into a political forum. I will certainly look at 40103, but the duty cycle is a problem. I want to filter the output to make a sine wave with not too much distortion. I don't seem to be able to reply to this group by email. Is that the way it is set up?