Waveguide port definition for CB-CPW in CST: Propagating modes.
I now there are more modes in this structure but i guess the usual propagating one is the CB-CPW mode, from center strip to lateral and bottom ground planes. I always consider this mode as the propagating one... but CST calculates it as the third mode... this puzzles me a bit. 4th mode is a TM one, so higher order modes are of no interest for me.
I always thought first two modes were even and odd modes. Am I right? Or are they parallell plate/microstrip(1st) and CPW(2nd) modes?? Is CB-CPW a linear combination of them?? I've been reading several IEEE docs but I stil don′t understand it well... is there even a so called "CB-CPW mode"?? And the "microstrip mode"?? Is it embedded in any of the first two??
If i set the waveguide port to consider 3 modes, the CB-CPW is the third one, but if i set the waveguide as multipin and only declare one mode(1+ strip, 1- laterals and bottom), no warning message shows up saying i am not considering any other important mode. I don′t know if mode order correlates with it′s order of importance or they are just the solutions of the 2D eigenmode solver.
What is more, i design the feed with another app so that the impedance of the CB-CPW mode be 50 ohm... And the two port definitions (multipin and 3rd mode in normal waveguide) give a different impedance value, for the same port dimensions: 57ohm for the 3rd mode, 49,97 ohm for the multipin definition. So my questions are:
For the normal waveguide port:
1) Mode order implies mode "importance"?
2) Do these first 3 modes always propagate?? (They are QTEM, so no cutoff frec...)
For the multipin waveguide port:
3) Is what I call "CB-CPW mode" the third mode that the previous port calculates? Or is it a linear combination of these 3 modes, so that when i define it via multipin port i′m implementing that combination?
For both port definitions:
4) Why different port definitions (mode Nr. or multipin)for the same mode give different impedances if the port dimensions remain unchanged and the mode is also the same??
By the way... if a waveguide port is a "2D" eigenmode solver... is it possible for it to calculate eigenmodes that are not exclusively TEM?? I mean.. QTEM, TE, TM and hybrid modes have longitudinal components. Is there anything I am missing about this??
Many thanks!!!
Iban
hace 27 minutos
Hi,
I'm not an expert with CST but in my opinion...
1) No, mode order doesn't implies importance
2) If they are excited yes they always propagate
3) In my opinion the CBCPW normal mode of propagation is the even mode and not a combination of different modes.
4) For port impedance solver use to calculate integral of the E and H fields over the port area for power calculation. For current calculation integral line of the H field (perimeter of your port) and for voltage they use the impedance line as a guide for calculating it).
I hope this could help you a little bit
BR
Hi Marco
Thanks for your answer. I share much of your opinions... and i think there is still something I′m missing. Rather than being a pure CST software question, my doubts are related to EM. May be it′s a matter of nomenclature. For me, "even" and "odd" refer to classic CPW, with "even" mode meaning E lines from strip to lateral grounds. Apart from that,when I say "microstrip" mode I mean E lines from strip to bottom conductor, with "CB-CPW mode" being a combination of these "even" and "microstrip" modes. And If I understood well, you call "even mode" this very mode with E lines from strip to lateral and ground planes... yes, this is also the normal propagation mode for me, but I think it remains being a combination of two modes, in some sense... If you make the gap wider and wider, there`s a point in which the E lines from the center strip leave the lateral grounds and end in the bottom conductor... But, If it were a combination of the other calculated two modes, CST wouldn′t show it, obviously... Yep... probably it′s a mode on it′s own.
Anyway, and specially if these 3 modes are "linearly independent", my main concern is if I have to consider (and define) the other two modes when simulating a CB-CPW in a multipin waveguide port. Of course, it will depend on whether the other two modes are or not excited in my structure... but i think this again depends on where you place the vias from top to bottom planes, the interface with wich you connect to the CB-CPW (coax, 2wire, microstrip, slotline...if you connect your coax external conductor only in one lateral ground). Whenever you have a transition, there are local higher order modes excited, and it seems factible for me that this modes excite the other two QTEM modes, apart from the "even" one.
As to my knowledge, considering only this "even"(E from strip to lateral and bottom planes) mode gives good results and I'll continue simulating this way, but I think there′s something with the other 2 modes that...mmmm... I still don′t understand it completely. May be I'm wrong and they are excited only occasionally...
And with regards to the line impedance, if your mode is TEM or QTEM (in TE and TM modes you can define your impedance in different ways) no matter how you calculate it, I think it MUST give the same value for a multipin port mode or a mode from the normal waveguide port, as long as field line distribution is the same (that is, the mode is the same). It′s a matter of the material properties and the geometric configuration of the waveguide. If we talk about input impedance, then, of course, position along the waveguide makes this impedance change. But line impedance is a characteristic of your waveguide structure, and that′s why I don`t understand why same modes give different line impedance values. This question remains unresolved for me.
But thanks again!!! :)
Iban
Hi Iban,
the fact that there are some lines of the E field coupling to the back side ground plane is just a matter of how your CPW is designed (in my opinion that doesn't means that you have CPW even mode and ustrip mode propagating together as one mode). What I mean if as you mention you implement wider slots, some E field lines will couple to the back side ground plane but impedance will change sligthly as the same amount of E field will continue to be within your substrate. Both modes cannot propagate as a single mode as they have very different impedance and propagation constant (For microstrip mode E field is almost completly into the substrate while for the CPW even mode the field is aproximately half over the CPW and half into the substrate).
In my opinion the other two modes have to be defined and simulated because as you mention you have discontinuities on your layout that could lead to the excitation of other parasitics modes (such microstrip, CPW odd mode...). This is important because you could have energy leakage from the main propagation mode to other parasitic modes (this leakage could be seen as I.L. for your CPW even mode). Then in my opinion is important to perform a multimodal simulation in order to understand if there is leakage from the main propagation mode to other parasitics modes (you can extract this info from multimodal S parameters, how the interaction between modes take place). If you observe that there is not interaction between modes then is a good idea to go just for the CPW even mode in order to save simulation time.
For impedance line ( as I mention I'm not a specialist for CST) but what I could say is if you are performing a terminal based solution (HFSS nomenclature) port impedance is calculated by terminal voltage and current. In order to calculate the voltage and current at your terminal you can have N open integration contours and N closed integration contours respectively(one for each mode present at the port). The modes could include parasitics modes generated from your main mode and transmitted to the second port o reflected to the first one.
For waveguide ports you are just calculating the port impedance corresponding to the field distribution of the excited mode. That's the reason behind the difference between port impedance for terminals based solution and modal based solution.
Regards, Marco