Howabout a DIY atomic clock?

[Duo]

Well, this idea of mine isn't actually for a true atomic clocke per-se.

However, the idea is to use a microwave signal generator and to decade it down to milliseconds, seconds, minutes, hours, days, weeks, months, and years.

I think that by decading a high accuracy source of around 4Ghz, I should get a very accurate second timer output.

The frequency standard I have in mind is an HP8616A signal generator. It is a cavity tuned industrial signal generator and I believe it will stay on a given frequency for a long time.

Does anybody have any ideas or comments about this idea?

Any circuitry ideas?

[janneman]

"I believe it will stay on a given frequency for a long time."
Well, if that's all you know, how do you know it will be more accurate than, say, a $9,95 Japanese watch?

Look upo the accuracy of the signal generator, it will probably be given in ppm (parts per million) where 1000 ppm = 0.1%.

Then check up long term stability.

For ultimate requirements you can't do better then a radio-controlled/synchronised clock; these are in effect atomic-controlled.

However, if you want to have fun, go ahead and build your design.

Jan Didden

[Duo]

Well, actually, my idea is after that of the clocks used by the NIST in colorado for the WWV UTC broadcast system.

I simply want to see if I can make a fairly accurate clock from this nice old sig gen I have sitting around, simply for something to do with it.

I haven't yet designed anything. I am just thinking up ideas right now and I believe it would be a very good project to attempt.

[Circlotron]

Here in Australia at least, many of the TV stations have their horizontal sync frequency (15,625Hz here) derived from a rubidium frequency standard, accurate to 2 parts in 10^11. Dangle a wire across the back of the TV, have a little tuned cct to get this frequency, multiply it by x64 with a 4046 phase locked loop for 1.00000000000 Mhz, (more or less) then just do what you like from there. Worked for me. Just use a nice long settling time on the pll loop filter.

Edit: Info was from a mid 70's Electronics Australia magazine, but probably still true enough.

[Rob M]

http://www.ntp.org/

[Duo]

Circlotron: Very interesting idea, I like that idea. Still, I want to see what I can do with this signal gen.

Rob M: I already have that.

[Rob M]

How about this one?

http://www.leapsecond.com/pages/atomic-nixie/

[OliverD]

OEM GPS receiver boards provide a 1ms and 1s "tick" signal used in applications to update the screen synchronically to the UTC second.

Very accurate.

[jackinnj]

there is an EBay T&M for Calibration and Standards and rubidium oscillators, complete HP clocks go for $40 to $200.

for me, I just keep my HP3586C "Measurement Receiver" running all the time -- it has an ovenized oscillator and I can "beat it" to any receivable standard. I use the 10MHz clock to drive my HP5334B counter and HP3336B generator.

[Tensop]

How about a microcontroller that can interface to tcp/ip protocol, wifi standard? you can receive it via a computer network via Network Time Protocol? Surely Maxim IC make something capable of this.