If I remember correctly, I have already seen such transformations between Wilkinson-like and Quadrature-like using 90° shifters.

However, I find it simpler to go straight to a quadrature coupler for two reasons. For wideband designs, the 90° phase shifter is not trivial to design. For higher power designs, the grounded resistor of the quadrature couplers is simpler to handle than the floating resistor of the Wilkinson.

Have a nice day,

Many thanks for your answers.
@Dale xiao: could you please give us the link towards the Chinese website ?

It depends:

* I always put annotations in my schematics, as text, and even images

* Lot of words documents in my project folder.

* Excel as soon as I mix numbers/calculations/explanations.

* Notepad++ (text files) for small content.

Keep it simple and convenient for success.

Hadrien

Good points but several things are mixed. I try to make an executive summary since I know a bit about this topic:

* The microstrip does NOT support a pure TEM mode at any frequency, due to the difference in propagation velocities between the two materials. cf. Pozar.
* It is not a matter of filling factor or material resistivity.
* The key parameter is the difference of speed between the two materials.
* This quasi-TEM mode is close enough to a TEM mode so it can be approximated by a TEM mode, and all relevant concepts of TEM modes (voltages and currents) can be applied without trouble. Say otherwise, there are not so much difference between the different ways of defining the voltage. Hence the name quasi-TEM.
* Particularly, for low enough frequencies, this quasi-TEM mode can be calculated in the same way as TEM modes.
* But this quasi-TEM mode gradually changes when frequency is increased.
* This dispersion needs to calculate the still quasi-TEM mode using full wave mode solvers.

But don't worry too much. There are lots more of real problems to care about.

Dear colleague,

First of all, quantify your approximation : how much non-real is your caracteristic impedance ? If, for instance, my characteristic impedance is 50 + 0.1j Ohm, I would simply drop the imag part.

Stay in touch,