site:www.edaboard.com branch line coupler
I have a question about the 90 deg branch line coupler again.
The attached figures show two tests of the branch. The (a) is the traditional one. The simulations show pretty good isolation, return loss and power splits.
The (b) is the other one I tried. The two outputs line are basically rotating 90 degree for the layout purpose. But the simulations show this structure has a bad isolation and non even power splits.
I am wondering whether there are anyone design the (b) structure before. Are there any tricks to get a good performance for that coupler?
Thanks.
That is odd I would think that would be ok, reflections will be slightly different but not a lot, what happens if you move the other two inputs around on the (b) picture so all ports are going in and out of the same direction just a different direction than in (a)?
Let me know if you look at that.
Yes you can do it with optimization on the line widths and corners.
Added after 48 seconds:
Note that your 50 ohm line widths are differ for port 3 and 4
I looks like you are rotating the lines to close to the stucture. Why not rotate away from the structure?
You are reducing the line length and impedance the way it shown in figure B.
Yeah that is what I was getting at
Sure b4bb4ge, sure.
Just kidding
After a few rounds of simulations, I found that the rotated arms at port 3&4 cannot be too closed to the coupler. No matter how I optimized the coupler structure itself in (b), it doesn't work.
So the only way for me is to more the rotated arms a little bit away from the coupler, even a very small fractions of wavelength improved the performance a lot. Quite surprising.
Are you still face the problem? I did not understand.
Sorry for that.