[SOLVED] [HFSS] How do you terminate a microstrip line with a given load impedance?

Hello,

I am currently building a demo of a microstrip line with a tuning stub. The transmission line is terminated in a load impedance of 100-50j ohms. The microstrip line has a Z0 of 50 ohms, and the stub has a Z0 of 100 ohms. I know my designs for the microstrip lines work, but I am having trouble terminating the transmission line with the fixed load. I have designed the stub length and distance from the load with a smith chart, but the magnitude of my S11 parameter is about 0.54. I think my main problem is how I am simulating the load. Oh, and I'm also feeding the structure with a waveport at one end. I have used two geometries for the load. One was to assign one outer face of the model, both the air and substrate, as an impedance boundary. The other design was to have a square going down into the substrate, and assign it as an impedance boundary. For this last design I made the air box and substrate longer than the transmission line - the impedance boundary was not touching the outer boundary of the model. In both cases my network was not matched. In short, hod do you simulate a fixed general load impedance in HFSS?

Michael Kelley

One of my office mates found the solution. The impedance boundary is defined by square, which means that you have to make a square - not a rectangle. The two words are sometimes used interchangeably. What this means is that in HFSS whatever shape that is drawn is not set to the impedance that you define, but rather the shape's impedance is scaled by a some square unit. When I made my width equal to my height the stub example that I created worked like the theory predicted. Before this I had a rectangle because my microstrip width was larger than the height of the substrate. In short, if you want to simulate a load with a specific impedance make a perfect square, and don't put it near any of your boundaries - except connecting to your ground plane.

Michael Kelley

I don't know for sure, but it might be possible to just put a port where you want the termination, and define that port to have the impedance you require. It defaults to 50 Ohm resistive, but there's no reason it can't be any complex value you want.

Dave

Dave,

My office mates and I tried that with a wave port and a lumped port, but the result deviated from theory too much. My guess is that it has something to do with having a wave port at the boundary. I'll try to remember to post a picture of the final VSWR that I got tomorrow.

Mike

Here's a picture of the geometry that I created and the subsequent VSWR. The distance "d" from the load is from the load to the middle of the stub.

m.kelley said:

Here's a picture of the geometry that I created and the subsequent VSWR. The distance "d" from the load is from the load to the middle of the stub.

Click to expand...

do you know the calculaton for that stub geometry? and location?

thanks!

If somebody uses QUCS simulator, it may be useful for them: as in QUCS there was a problem to define anything different than 50Ohms (complex part was not set correct), i created s-parameters file and now add file-defined 1-port device.
I think it will work for all simulation package, that support S-parameter file-defined devices. Just create your own 1-port s-parameters file, fill it with required impedances and attach it to the line.