Is there an option, or a secondary component, that will tell qucs that there is a "perfect" transition from waveguide to coax for 50/75 ohm applications? It would also be useful to have a transition that was non-ideal and be able to specify the break points.
On a side note, I have access to both ADS and Ansoft's HFSS. The one test that I've run comparing the rectangular waveguide on qucs to an HFSS model was very good. I'll continue using all three and comment if i ever find discrepancies. Thanks for this wonderful program.
You can just tie them together right?
Or do you mean that you want a 50 to 75 ohm transformer? there is also a perfect transformer available
What I've found is that when you tie a 50 ohm source directly to waveguide, you don't get the true response. The impedance of the waveguide also seems to change with frequency, so simply change the source to say, 377 ohms, while it gives a better response, still seems to be lacking. In full disclosure, I work for a company that builds coax to waveguide transitions, so that when you put a transition on both ends of a piece of waveguide the full bandwidth of the waveguide can be utilized with a 1.02:1 or better VSWR. I have yet to see this kind of transformer, unless the perfect transformer your talking about is the answer?
I don't think I was as clear as I should have been. Using the Ansoft HFSS modeler, I've simulated WH1150 waveguide (11.50" x 5.75"). Then using QUCS, I gave it the same problem. The difference is that HFSS gives an option to see a non-normalized solution, which in this case makes more sense, since it shows very little loss and a good return loss (<-40 dB) across the operating range of this size. When I run this problem in QUCS, I get mostly a 0dB return loss, with a few spikes varying from -14dB to -40dB and terrible insertion loss. Incidentally, when I tell HFSS to normalize to 50ohms it gives me the same response. So basically, without some sort of transition that takes into account waveguide characteristics, any simulation I run is going to be riddled with poor VSWR and loss performance that isn't really there. I'll be happy to provide screen grabs of what I'm talking about if requested. It seems to me that the transition would have to be nothing more than another piece of waveguide with one end normalized to some value (coax or stripline connection), and the other not normalized (waveguide connection).
Sign up for the SourceForge newsletter:
You seem to have CSS turned off.
Please don't fill out this field.