The idea consist to create a switcher from a power divider using varactor diodes to adjust impedance adaptation. For a 6 ways power divider, four outputs works well but two output don't work despite the same condition used for worked outputs.
The structure is symmetric and has a cylinder shape.
How to excite all ports? I think that problem is related to dispersion of EM waves in the substrate which is asymetric?
I would like to excite all ports. How to correcte the distrubition of fields?
Notes that iwould like to excite TM surface wave with a substrat er=2.2 for 5.8 Ghz.
4 output work and 2 not work can explain that one of higher modes is exicited?
How i can excite the fundemental mode to distribute field in symmetric ways?
Have you looked into circulator (>3 port) , they are commercially available for your range too.
If you would like to look into the analysis , have you come across Manley-Rowe relations.
This is meant to calculate energy vs frequency in waves . Check Microwave devices and circuits by S L Liao (may be an old book but is too good)
The idea to feed in the center of a circular by fed coaxial. So, the circulor can't be made by this configuration
I guess then you have to look for:
1.circular waveguide design for specific modes (TE/TM) with port spacing = 360 /N .
2.Since you say dielectric may be surface waveguide !!!
3.I know Wilkinson is resistance based , they are commercially available though .
4. Directional couplers , sorry I have worked with 3-port ...
Dear Mohamed Aymen,
can you share an image of the structure?
Best regards
Is this with miscrostrip lines or waveguides?
I don't see the input port, I suppose that it is in the center of the structure for the symmetry?
it's fed by coaxial cable at the center.
I use microstrip line.
My idea is to eliminate TE01 modes (no waveguide structure and no SIW also)
Because it is symmetric I don't think you should have different results for the different states at the different output ports.
In order to validate your results I suggest, if you have not done it, to look at what is the insertion loss that you get (S21) and also to check if the simulation converged.
Maybe you can also check that all your lumped components are well defined.
For selected Sij (4 ports), I have a peak at desirable frequency (around -8dB) and the others disabled ports are around -20 db (the worst case -15 dB). but, two output port haven't the same behaviour as the others (a differents insertion Loss response) .
and varactor diode parameters are correct becaue first i used a datasheet and varactor diodes are serial RLC (i changed the capacitor value to have Transmission or not for one port.
The results should be the same for all ports.
Maybe you can try to rotate the lumped components by 60 degrees around the center of the structure, maybe this can help you to find if there are differences in the modelling in some places.
In the image of the structure I see a box around your structure with two ports touching the boundaries: maybe this can be an explanation?
If you use radiation box, the boundaries should be at least at a quarter wavelength distance from the structure.
I use CST MWS where boundaries conditions are defined automatically (All Et=0)
But the boundaries are touching two of your ports: probably these are the ports with unwanted responses?
If so, then you just need to move the structure away from the boundaries
By the way, this idea of using the TM mode surface wave for designing a reconfigurable power divider looks like a nice idea
Try to use open (add space ) boundary condition for all directions.
The problem is solved (Open boundaries). Now, the device switchs all outputs.
Thank you very much.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering