HFSS: How do I connect a port to coax?

I'm a HFSS newbie and are having troubles with something I think should be quite simple. I'm trying to connect a 50 Ohm port to a piece of coax, which forms part of a coaxial colinear antenna. This sort of thing:

**broken link removed**

The coax has an inner, dielectric, braid and sheath. The dielectric properties of the sheath are possibly important in this application, so I gave that in my model. See attached file.

I want to connect the coax connector across the bottom of the antenna (at z=9 mm). But I'm puzzled how to do this. I've done it before, but get a warning that the port only has one connection.

The big cylinder which surrounds the antenna is my radiation boundary. An Ansof engineer said it was better to use a cylinder than a box for my application, as it is essentially a cylindrical antenna.

Hi,

You should close the end of the coax with a cylinder made of PEC and select the face of tshis cylinder which is towards the coax, then you should use waveport (Modal solution) and draw the integration line between the coax shield inner metal to the central pin outer metal.
Hope this helps.

bya036 said:

Hi,

You should close the end of the coax with a cylinder made of PEC and select the face of tshis cylinder which is towards the coax, then you should use waveport (Modal solution) and draw the integration line between the coax shield inner metal to the central pin outer metal.
Hope this helps.

Click to expand...

What do you mean by "close the end of the coax"? At the minute my model has all parts of the coax starting at z=9 mm. Should the PEC cover just the inner conductor? The inner is currently copper, so are you saying I should add a short length of PEC on the end of that? That will then mean the inner protudes outside the dielectric and braid, which does not seem too logical to me unless I also add the same thickness of PEC on the braid.

See attached file!

bya036 said:

See attached file!

Click to expand...

Thank you. Did HFSS -> Analyse All (or whatever it was called), actually run properly? If so, I'd be intersted in any results you got.

I've had HFSS on a trial, and the trial has now expired, so I can't actually look at the file you posted, although I believe the trial will be extended next week, so I'll be in a position to look then.

HFSS sure is one complex bit of software to use. I'm in my 40s and can honestly say its the most complex bit of software I've ever tried using. (This reminds me of something funny. I once saw a CV for a post-doc position where someone had put their software skills. They put "Windows Paintbrush". I can understand putting Photoshop or HFSS on ones CV, but not Paintbrush! Needless to say the guy was not interviewed!)

Dave

drkirkby said:

Thank you. Did HFSS -> Analyse All (or whatever it was called), actually run properly? If so, I'd be intersted in any results you got.

I've had HFSS on a trial, and the trial has now expired, so I can't actually look at the file you posted, although I believe the trial will be extended next week, so I'll be in a position to look then.

HFSS sure is one complex bit of software to use. I'm in my 40s and can honestly say its the most complex bit of software I've ever tried using. (This reminds me of something funny. I once saw a CV for a post-doc position where someone had put their software skills. They put "Windows Paintbrush". I can understand putting Photoshop or HFSS on ones CV, but not Paintbrush! Needless to say the guy was not interviewed!)

Dave

Click to expand...

Dave, I ran the simulation file above submitted by the other poster and it died late in the 'game' with the following three errors:

error: Port refinement, process hf3d error: Port 1 does not have a solved inside material on either side.. (5:27:04 PM Jan xx, xxxx)

error: hfsscomengine, process hf3d died unexpectedly. (5:27:04 PM Jan xx, xxxx)

error: Simulation completed with execution error on server: Local Machine. (5:27:04 PM Jan xx, xxxx)

- - - - - - -

I got a slew of "info" messages before like these first "info" messages before the error messages above appeared:

info: Bodies "Inner_conductor" and "Dielectric" intersect. "Inner_conductor" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

info: Bodies "Inner_conductor" and "Braid" intersect. "Inner_conductor" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

info: Bodies "Inner_conductor" and "Braid_1" intersect. "Inner_conductor" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

info: Bodies "Dielectric" and "Braid" intersect. "Braid" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

info: Bodies "Braid" and "Sheath" intersect. "Braid" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

info: Bodies "Braid" and "Inner_conductor_1" intersect. "Inner_conductor_1" will take precedence in the region of intersection. (4:53:39 PM Jan xx, xxxx)

and more ...

- - - - - -

Any idea what can be done to make the simulation run 'happy'?

Jim

Hello Jim

It is Normal for hfss if you have given precedence to Pec/copper over substrate when they intersect, what happens when you have a pec running through substrate,pec takes precedence and substrate is considered as if it is drilled in by the pec. like for example you have your coaxial conductor running through the substrate to feed the copper in a patch, either you have to cut the portion of substrate where the pec coax is going through or give the precedence to it. so it is not a major issue if there is a conflict of substrate and pec intersecting.

hope that helps

Regards
Rotmanlens.

I think I found the solution for the error: "process hf3d error", although I have not tried it yet. It involves defining the feedpoint on the antenna as a waveport per the 9:38 point in this video instead of a lumped port:

Ansoft HFSS z Antenna tutorial - YouTube

Jim

I can certainly understand the comments about how difficult HFSS is to use. Since it sounds like you are using a demonstration license I suspect that you are also in touch with Ansof/Ansys personnel. There are extensive training packages and examples available. Just ask them for guidance. That is really one of the most efficient ways to get started.

There is nothing wrong with the responses above but if you are just getting started asking questions and trying to make sense of the answers is a challenge.

Thanks a lot. Yes, I have been in touch with Ansys - the technical contact in the UK is very good.

Dave