mixer

I want to upconvert a 100 MHz signal to 2.25 GHz. Has anyone got idea on how to design a mixer for this. Can a single mixer do or I need more than one mixer stage. I am trying to design a single balanced diode mixer. Which hybrid should I use, rat-race or branch-line. How should I solve the bandwidth problem of the hybrd.

One thing that controls your design is if the 100 MHz or the 2.4 GHz signals cover a wide range of frequencies. You will have enough problems with mixing images as it is. Then with the feedthrough of a single balanced mixer things will be worse. Have you thought of a transformer-diode ring type mixer?

To get rid of the feed through and image problems I would suggest a double conversion method. If you design it properly the local oscillator for both mixers can be the same one (and frequency).

The bandwidth of the Rat-race or branchline coupler using quarterwave transmission line is narrow due to the property of transmission line.
If you want wide bandwidth, then use transformer balun. You can get more wide bandwidth.

BR.

The bandwidth of the signal is about 50 KHz.

This narrow bandwidth is an advantage to you. Now the only unwanted outputs will be the feed through of the two signals going into the mixer and the unwanted mixing products.

You will have to do the calculations of various filter types (resonators) and how much they attenuate these unwanted outputs and compare that to the system specification. If you cannot meed the specification in one conversion, you will have to use two.

Your application is very close to an Amateur Radio project: 13cm to 2m transverter.
Do a web search for this definition, and you will find a lot of examples. I remember, long time ago I made something like this, to use on amateur satellites.

In your case the only problem that you will have after up-conversion, is how to reject the level of the LO (in your case 2.15GHz or 2.35GHz) from the desired signal (2.25GHz). You have to spend time in some quality filters.

What about hybrid coupler. I am trying to avoid tranformer balun because it is usually difficult to design or obtain a tranformer with the required coupling coefficient and inductance.

For filter, I am using microstrip coupled line filter. Is it O.K.

Coupled line filters are adequate. If you do the double conversion scheme the filters will not be required to have as many resonators. Also you can use single balanced mixers (nulling out the LO) and have adequate performance.

What about hybrid coupler. I am trying to avoid transformer balun because it is difficult design the balun having required inductance and coupling co-efficient.