Unequal-split Wilkinsons

Revised November 25, 2009

Click here to go to our general discussion of Wilkinson power splitters

Click here to go to our unequal-split power divider calculator, it handles Wilkinsons, rat-races and branchline couplers!

New for August 2009! Here's a page on N-way, unequal split Wilkinsons

New for May 2009! We're reorganizing the content on this topic (slowly), so our apologies if you can't quickly find what you need.

Our material on this topic is (or will be) spread onto the following pages:

Split Tee Power Divider - the original unequal splitter from 1965

Choose ZA solution - provides the most freedom to the designer

Maximally-flat transformers used in unequal split Wilkinsons

Kouzoujian's N-way splitter - uses unequal splitters in a planar layout

Traveling wave combiner - coming soon

We have a brand new page on unequal-split Wilkinsons, which derives the math behind the impedance calculations. We'll show you an extra degree of freedom that the "classic" unequal-split Wilkinson equations don't provide.

We are talking about two-way Wilkinson splitters here, but of course you can make unequal split N-way dividers. The way to make power split unequally requires two things: the quarterwave sections must be of different impedance, to encourage more of the signal to travel in/out the lower-impedance arm, and a second set of quarter-wave sections are needed, to transform the arm impedances back to 50 ohms. The structure looks similar to a two-stage Wilkinson without the second isolation resistor (here "R_W" is the Wilkinson isolation resistor).

The above image and many more are available in a Word document you can get in our download area (for free).

What constitutes a valid solution to the unequal splitter network? There are really just three criteria that must be met:

1. All ports are perfectly matched at center frequency
2. Infinite isolation is achieve at center frequency
3. In the even mode (exciting Port 1) the splitter is lossless.