need to match this dipole array URGENTLY

[kae_jolie]

Guys,

I really would appreciate your help in matching this 5-element semi-circular dipole array which is excited by 5 lumped ports in HFSS. What offset excitation port phase under Edit Sources do you use when matching it? Is it the same phase I want to steer it at? I read in Stutzman book that arrays are usually matched at broadside or phase = 0° regardless of where you want to scan them later.

Also, if I choose the offset excitation port phase angle to be 45°, what scan angle does this correspond to? What equation are you using?

thanks all.

 

[Darktrax]

The un-zipped file was not accepted by the version of HFSS I was able to try it on. I had to forcibly abort HFSS to escape the lock-up.

 

[kae_jolie]

I use HFSS 13. What version do you have?

 

[DeboraHarry]

The un-zipped file was not accepted by the version of HFSS I was able to try it on. I had to forcibly abort HFSS to escape the lock-up.

I have not tried loading it, but it would appear to be version 13.0, as you can see if you open the .hfss file in a text editor.

 

[kae_jolie]

Can someone help with this please?

Thanks.

 

[Darktrax]

The HFSS I had temporary access to was version 11.

While HFSS is a well-known industry tool, viewers here who might have suggestions will have a variety of other ways not necessarily locked to Ansoft. It might help if you could post an image of the general layout, maybe captured from part of your screen display. You would have to do the simulations yourself, but I am sure there are many in this forum who would be able to suggest strategy, and ways to optimize.

We need to cut through the number of combinations we consider. Therefore..
1. Are the dipoles independent, each having its own excitation source?
2. Are the dipoles identical, or is this a multi-band device?
3. When you say "match this dipole array" do you mean finding a network with a 50 ohms port that drives all the dipoles combined?

For my suggestion, instead of spending time simulating, you go on a little design journey.
Just from imagining the dipole layout, first dream ahead. Pretend you already know the drive impedances of the dipoles, each in the presence of all the others. Start to think of quarter-wave matching lines, and the space they need. Can they be in the array semi-circle?

Now try to exploit symmetry. You know already that the end dipoles have the same impedance, and that the two dipoles next in from the ends have the same impedance, and that the centre dipole might have an impedance all of its own.

You can see that the array is what remains if you had started with a full circle array of 8 dipoles, and then suddenly removed 3. One thing you do know about a full circle array is that the dipole impedances must all be equal.

Next is to model a single dipole in that structure, with the dielectric there, and no other dipoles present.
Don't bother to extensively simulate. Use a single frequency, or a small range. Then, add in all the other 7 dipoles, each with a terminator having the impedance you just found. What has happened to the driven dipole? You can set some optimize strategy to find (approximately) the impedance they all should have. Take care to keep it real! Do not allow silly divergent answers.

Next step is to remove 3 of the dipoles, see what has happened to the the centre dipole of the 5 you have left.
Hopefully, the impedance change will be minimal. Check also the dipole next to the centre. You have started to zero in on the final design.

Move the excitation to an end dipole and find what it should be. By now, you will have a good idea of where the design is going. You can try an open circuit, and then a short-circuit on an end dipole, and see what happens to the impedance of the centre dipole. If it does not move much, then we get more confident.

Perhaps there is a shorter trick to finding the impedances of the dipoles, but once you have these, the business of matching to them, possibly combining also, moves on to laying out matching sections, quarter-wave transformers, Wilkinson-style combiners, etc. as may suit.

 

[Shehata]

Hi

You can sweep the phases offsets and find the value set that corresponds to the beam direction you want. It is a post processing sweep so you do not need to re-simulate,

Cheers,
Shehata