How to divide a 1.7GHz oscillator signal by 2 w/o using chip

[multanova51]

RF freq. divider

Hi Guys,

I am looking for a solution to divide a 1.7 GHz osc. signal by 2 without using a digital divide-by-two chip.
Has anybody an idea to realise it with a more simple approach with normal lumped components ?
I know the other way around is more easy (multiplier).

regards and Seasons Greetings to all...,

Mike

 

[flatulent]

only two ways

The logic method is the cheapest and easiest. The other way is complex and expensive. Run your signal into a mixer. Filter the output for the half frequency and run it also into the mixer. The half frequency will subtract from the full frequency producing the half frequency. This method has the lowest phase noise and goes back at least 60 years.

 

[klug]

Hi friends!

Thanks, flatulent. Very interesting method - it is new for me. But how they got half frequency signal just after switching on of this device?

Another possible method is to use second generator on 0.85 GHz synced by signal from first 1.7 GHz. It was used on low frequencies long ago, but maybe it possible on high. But it will be too complicated in comparing with IC divider.

Best Wishes! klug.

 

[g579]

There is another cheap method.
Make a varactor (varicap) circuit, tuned out for 1.7 GHz for one port, and for another port tuned out for the 1/2 freq. of 0.85 GHz.
If 1.7 gHz drive is strong enough (>0 dbm approx.) the capacitance change will produce a negative impedance and you will get the 1/2 freq. at the 0.85 Ghz tuned port.
Disadvantage of the method is that it is not easy to make it stable.

 

[flatulent]

regenerative divider

The method I described above is called "regnerative division" and since it is a feedback loop, it starts with the noise just like many oscillators ramp up from turn on with the noise. By the way, you can divide by N+1 by filtering the output to the frequency you want and then multiply by N and filter before going back to the mixer. In all cases, the loop gain around the mixer-filter-harmonic generator-filter and back to the mixer should be greater than 1. This may require an amplifier. The loop gain should be high for low level signals which means that it is best to use the LO port for the original signal and the RF port for the loop.

 

[multanova51]

RF division of 2

Guys,
I never thaught there were that many solutions. Flatulent , I was very much intrigued by your suggestion of using a mixer. At first I was a bit sceptic about it but you are absolutely right. I found following articles as an example on the net :
**broken link removed**
**broken link removed**
I will do some more reading on the subject and try something out.
Thanks,

Mike

 

[mwpro]

Hi,

The method suggested by flatulent is indeed an old method that exists for a long time. Indeed, you can have 1/3, 1/4 division. The only problem (critical) is bandwidth. For 1/2, probably you can get 10% bandwidth. For others, the bandwidth is minimal.

Still this class of method found its application in ultra high microwvae frequency where static division by flip-flop is not feasible.

For 1.7GHz, static division can do the job nicely. Thus, the best way is to use static division.

mwpro

 

[flatulent]

areas of application

This regenerative method is still practical. about 5 years ago 2-10 GHz binary dividers were $200. Still today it is the best method for frequencies over 10 GHz and for low phase noise. There are several ways to extend the bandwidth. One is to use full wave doublers which have high rejection of the fundamental input. Another way is to use tracking filters. In some applications the narrow bandwidth is no problem. Many commercial and military bands are about 10% wide.

 

[VSWR]

Re: RF freq. divider

Using a static prescaler IC cannot be 'a more simple approach' nowadays: take a close look on e.g. the D602 by SiGe Semiconductor (www.sige.com). It runs on 3 V, is useable up to 6 GHz and requires only a few external components. Samples might be requested free of charge.

Hi Guys,

I am looking for a solution to divide a 1.7 GHz osc. signal by 2 without using a digital divide-by-two chip.
Has anybody an idea to realise it with a more simple approach with normal lumped components ?
I know the other way around is more easy (multiplier).

regards and Seasons Greetings to all...,

Mike

 

[strabush]

unusual frequency dividers

In this article you can find some unusual frequency dividers and how to push logic to divide at higher frequencies than they were designed to.
Article from Wenzel Associates site.

 

[multanova51]

divide by 2

Gents,

Thanks for your kind feedback on possible solutions.
At first sight it might look silly to ask for such a solution because of the dig dividers around even with good phase noise.
My particular application needs to be extremely low cost (consumer product) and furthermore a digital divider is likely to give more EMC issues as an analogue system. In particular high harmonics of the incoming freq. might radiate as previous experience learned me. Also an EMC fix needs to be low cost if required.
However, I am very thankfull for the different valuable tips and I will try to do a few tests with actual circuitry.

regards
Mike

 

[zorro]

U can search for "subharmonic" and "injection locked" (oscillators).
Z