Trying to draw the schematic for a power divider circuit

Question

I want to draw both the schematic and layout for the following power divider circuit using microstrip lines in ADS. First of all I want to see if my schematic agrees well with theoretical results at all.

Here I have drawn the schematic. I'm not 100% confident whether I've ended up with the right schematic configuration to begin with. Particularly I'm not sure whether or should I need to use step junctions, junction Tees as required?

To check my schematic I did an S simulation and drew magnitude of S11 for comparison. The left shows my simulated result while the right depicts the result obtained from theoretical calculations given in this paper. As you see it doesn't well agree with the theoretical results :(

Note that as far as S11 concerns the series connected R and C does not matter at all (a result verified by even- and odd-mode analysis given in the paper). So for simplicity I didn't include them in the schematic. So now I don't know what I should do to get as close as possible to the actual S11 response. What practical considerations have I missed?

Answer 1

As you suspected and as mentioned in the comments and user287001's answer, you should use the proper components to model any junctions in your microstrip lines, including any steps in line width, etc. In the case of this particular circuit, there's no good built-in microstrip component to model the junction where the split occurs, so that's something that would typically be EM simulated.

In any case, I think your dimensions are off... Using the dimensions from the paper and adding in the parallel \$Z_{ISO}\$ results in something reasonably close to the paper. See below schematic and plots. I simplified the short stub sections to a single MLIN and an ideal short to ground, which is not exactly what they do in the paper, but seems to be somewhat close to the results presented in Fig 7. of the paper. Note my use of an MCROS and an MTEE to model some of the microstrip junctions where the stubs connect. These are very important when working with the analytical transmission line models included in any commercial microwave CAD tool.

Answer 2

Seemingly you haven't got a flood of pro quality answers from advanced microwave designers. But here's a guess. It's essentially the same as I wrote as a comment.

Drawn joints between transmission line parts:

I guess the joint assumes some default wave behaviour. It can well be ok if the strip continues by having only a change of width. But three line joints should have some defined geometry. I have no access to high end design software, so this is only a guess. I also guess you have at least thought the same when you wrote about maybe missing T-parts.

I suggest that you try to reveal possible fundamental error by trying something simpler which you can check by yourself or which is generally well known and presented identically in textbooks.