CMOS Synthesizer

Hi Ivan,
That is really interesting. In other words I could used that as a PC-based ultra-accurate local oscillator for a receiver.
I feel a VISA coming on ...  
Very best regards from Harry

Hi Harry,
I have read the promise, too. I really want to fill in a customs declaration and pay VAT and duty, you know .
The web says "output reference frequency can have almost any value between 400Hz and 810MHz". Tuning is accomplished via USB. There is also another model with two outputs in the shop.

VBR from Ivan

Hi Ivan,
I would not worry about BREXIT. According to the site:

"BREXIT: In simple terms, nothing will change until at least the end of 2020. There will be no customs, VAT, import/export or any other changes until January 2021."

I am also buying from the UK and have yet to have any problem. The last item I bought was just three weeks ago.

It looks like a nifty little unit. I didn't look too closely but it would be interesting to see what reference frequencies it will deliver.

Best regards Harry

Hi all,
I found this beast recently. It may be a smart solution when one needs better accuracy and stability than that of an XCO. Set to a suitable frequency from its range and with a divider added, it would do the job.
The drawback is it needs view of the sky to catch GPS signals.
I am conidering to order one if it is still posible after brexit.

VBR from Ivan

Hi Craig and Ivan,
This sounds like an interesting project, but as Ivan stated, the capture/lock time at such low frequencies could be long, even if you have a filter-bypass circuit to speed up the lock time.

But looking at your requirements: "115.2 Hz. etc. Accuracy +/-.01 Hz." = 1 part in 11520, which is not very much in terms of accuracy. A simple crystal oscillator can give you 1 part in 100,000 which is about 10 times better than you want.

Perhaps a more simple method would be of interest: a crystal oscillator and a programmable divider. For examply, using a 10MHz crystal divided by 86806 = 115.1994Hz, or 10MHz divided by 86805 = 115.2007Hz. the advantage of dividing is that you do not have any lock time, but the disadvantage is that the frequency steps become greater as the higher frequency you want. But at 115Hz then a simple divider will give you approximately 10x the accuracy you are asking. Components: one crystal oscillator module and one programmable divider (4040 plus diode/switches). But this is just a suggestion  

The 4046 and 4059 would work to make a nice PLL. Better accuracy than the above , long lock-time , no calculations required ,  more complex and expensive , Easier to change frequency .

Very best regards from Harry SM0VPO

Hi Craig,
it is well possible to use a 0,1 Hz timebase derived from a crystal oscilletor or even a rubidium atomic clock and multiply it by 150 - 1200 using a PLL consisting of 4046 and 4059. The drawback is a long settling time of the PLL after startup or change of the frequency.
A direct dividing of an output of a precision oscillator doe not do?

BR from Ivan

Hi,
What I'm trying to do is make a very limited range, low frequency synthesizer: 15.00 Hz. to 120.0 Hz. but need it to generate to four places, 27.82 Hz. for example or 115.2 Hz. etc. Accuracy +/-.01 Hz.
From what I can read, it would seem like a 4059 between the in / out of a 4046 might work?
Have used a sig. gen. in place of crystal osc. just to try it and followed your CMOS RF Synthesizer circuit.
Really am not getting anywhere?
I obviously do not have any experience with this type of circuit, perhaps you might point me in the right direction?
Thanks, Craig