HFSS referring up the line

[itsjwatson]

Hello,
Is it possible to simulate away from the measuring plane towards the DUT in HFSS/Designer?

In particular I have an antenna connected to the i/p of a (known) matching line. I am measuring the reflection coefficient at the o/p of the matching line.
I am trying to refer the measured RC at the o/p of the line, through the line to it's i/p so that I can use this data to design a better matching line. This sort of involves simulating 'backwards in time' as it were.

The line and antenna are integrated in one unit so I can't simply measure to antenna impedance directly!

So far everything I've tried takes me the wrong way, going from my reflectometer away from the matching line instead of through it.

While this is easy to do with the simple transmission line equations (just change the sign of the length) it seems more difficult (or impossible)in HFSS.

 

[jeeudr]

What type of ports are you using in your design ? Have you tried to de-embed the ports ?

 

[itsjwatson]

Thanks for the reply,
I'm using wave ports as this is a simple form of coaxial line.
I've just tried your suggestion of de embedding the ports but it works as I'd expected.

HFSS models down the line from the reflectometer towards the source, then refers up the length of line specified in the de embedding (I specified the line length I'm modeling) so it ends up back where it started. I guess the is just to wind out the length of a port and get to a desired measuring plane.

Not what I'm trying to do unfortunately.

EDIT
After a bit of thinking and playing.
De embedding both input and output ports by the line length seems to do what i want, at least for a symmetrical line (plain line with a mismatch dead center). Just a simple problem so I can compare it with my analytical results.
I suppose it swaps the ports to the opposite ends of the line while the arrow of time stays unchanged.
Still working on it!

I've been using smith charts and maths for all my working life so I don't have much feel for these spiffy EM programs yet, probably looking at things in the wrong way.

 

[jeeudr]

I did not realise that you were actually trying to characterize the matching line. I was thinking that you wanted to get an idea about the impedance of the antenna after removing the effects of the feeding line (this is what I am working on currently).

After reading this: "De embedding both input and output ports by the line length seems to do what i want, at least for a symmetrical line (plain line with a mismatch dead center).",

I was thinking that maybe you could design a back to back device with two of your impedance transformers and use this to characterize the matching line? Just thinking out aloud here ...

 

[itsjwatson]

That is right. I am trying to find the impedance at the far end of the matching line after removing it's effects so that I can design a better one. I am making some progress and now think that it can be done.

EDIT
This seems to be how it works.

To refer up the line length L towards the antenna apply de embedding of +L to both ports. This will get the impedance at the far end of the line.

To come back just remove the de embedding from the ports.

I have been using an Nport model in designer to feed the HFFS model. Seemes to be working for a simple patch at the moment.