Frequency shifter?

[Plasmateur]

Hello,

I am designing heterodyne a circuit which will down convert a 40GHz signal to a lower frequency that can be read on an oscilloscope.



One way I envision doing this is the following -

Utilizing the following,

40 GHz from signal Generator
10 MHz clock from signal generator
2 mixers, one up convert and one down convert.


With the up converting mixer
40 GHz - RF
10MHz - LO
40.01 GHz - IF

With the down converting mixer
40 GHz - RF
40.01 GHz - LO
10 MHz - IF


It seems like this could be easier if I could use another device to shift the 40 GHz frequency rather than using a mixer. Does such a thing exist, and are they expensive?

Thanks!

[dmills]

Might want to remember that you get TWO sidebands and that a square wave is a comb of odd harmonics.

That upconverting mixer will produce 40GHz +- (10M,30M,50M,70M) plus some (significant) 40GHz leakage, not a clean 40.01GHz. Good luck filtering 40.01GHz from all the other products.

Also consider what the phase noise on a 40GHz osc usually looks like at 10MHz offset. 

Have you actually done much work in the high microwave bands?
That stuff is not simple, every line is lossy, every passive an inductor (Except inductors which are capacitors) every transition reflective and every trace a waveguide (possibly also a filter, they often have weird propagation modes up there), hard soldered brass waveguide is your friend up there.

Besides, time domain at 40GHz, really?
Anything up there looks like a sine wave simply because for it to be anything else it needs to have harmonics, and most microwave doings are just not that wide. Now if you were building a receiver, that would be different, but usually instruments for use up there work in the frequency domain or mix down to near DC for modulation domain analysis.

[Plasmateur]

Might want to remember that you get TWO sidebands and that a square wave is a comb of odd harmonics.

That upconverting mixer will produce 40GHz +- (10M,30M,50M,70M) plus some (significant) 40GHz leakage, not a clean 40.01GHz. Good luck filtering 40.01GHz from all the other products.

Also consider what the phase noise on a 40GHz osc usually looks like at 10MHz offset. 

Have you actually done much work in the high microwave bands?
That stuff is not simple, every line is lossy, every passive an inductor (Except inductors which are capacitors) every transition reflective and every trace a waveguide (possibly also a filter, they often have weird propagation modes up there), hard soldered brass waveguide is your friend up there.

Besides, time domain at 40GHz, really?
Anything up there looks like a sine wave simply because for it to be anything else it needs to have harmonics, and most microwave doings are just not that wide. Now if you were building a receiver, that would be different, but usually instruments for use up there work in the frequency domain or mix down to near DC for modulation domain analysis.

Thank you for the response.

I have not work in microwave bands.


Can I multiply the 10MHz clock frequency to help reduce phase noise and make filtering easier with a band pass filter? Is there some kind of high frequency oscillator I can drive with my clock?

There will be a few transitions. SMA to waveguide - collimate- waveguide - back to SMA. With a few mixers in there as well.

I may just have to work in the frequency domain for now, but would like to say something about the time domain.

[ale500]

That stuff is not simple, every line is lossy,

every passive an inductor 

(Except inductors which are capacitors) 

every transition reflective and every trace a waveguide

There is a book on RF that has the words "black magic" in its title... I wonder why .

[dmills]

It's not that bad, but 40GHz is not where I would do a first microwave project, start down where FR4 is good enough and the geometry is big enough to be visible to the naked eye (And modifiable with a scalpel and copper tape).

Also, at a Ghz or so, test equipment is somewhat affordable, a 40GHz VNA (And CAL KIT), generally not so much.

Finally, by 40G you really need **Expensive** software support to do the 2.5D or 3D field maths to simulate in a somewhat usable way.

Regards, Dan.

[Alex Eisenhut]

Why not buy a SD series sampling head?

http://w140.com/tekwiki/wiki/SD-30

https://www.ebay.ca/itm/Tektronix-TEK-SD-30-Sampling-Head/122970467366?hash=item1ca19c6c26:g:pUIAAOSwvSBaXk-1

And if you think that's expensive, you need to pick another hobby!