question on hfss lumped port

[wufeng]

Hi Guys,
I am simulating a loop antenna with HFSS. Since the diameter of the loop antenna is very small, it does not resonant at the working frequency. I use the driven mode and lumped port. From Z11 I can see the input impedance is 818-158j. So I plan to add an inductor to cancel out this reactance. But how to do that in HFSS? Can I renormalize the terminal impedance to 818+158j? I tried, but it seems the S parameter only change with resistance not reactance. That means if I change this terminal renormalizing impedance to 818+1j, the S parameter does not change at all? It is so wired. Can anyone help me? Thanks!

 

[DavidReina]

I was in the assumption that if you want to use lumped ports you need to be in driven terminal and not in driven modal.... maybe switching the mode you can fix your problem.

 

[wufeng]

Sorry. I made a mistake. I mean driven terminal.

 

[wufeng]

Also I found if I use the driven modal, I can renormalize with <re>+<im>j. That means if I change the imaginary part, the S11 will change too. But for driven terminal mode, only real part can change the S11. The imaginary part has no effect on S11. Does anyone know why?

 

[flanello]

An other idea:
You could define an output variable MyS11Parameter and calculate here your S11.
And if you do complex renormalization check if the program does it right, i have seen some programs that have made wrong calculations for this case.

 

[tallface65]

Hi Guys,
I am simulating a loop antenna with HFSS. Since the diameter of the loop antenna is very small, it does not resonant at the working frequency. I use the driven mode and lumped port. From Z11 I can see the input impedance is 818-158j. So I plan to add an inductor to cancel out this reactance. But how to do that in HFSS? Can I renormalize the terminal impedance to 818+158j? I tried, but it seems the S parameter only change with resistance not reactance. That means if I change this terminal renormalizing impedance to 818+1j, the S parameter does not change at all? It is so wired. Can anyone help me? Thanks!

First, you need to converge on the real and imaginary part of Z via the Expression cache as delta has no real meaning for this simulation (the resonance as you mentioned is very far from the solution frequency, so the port is essentially looking into a short)

Second, by definition, S Parameters use only the real part of the port impedance as a reference. This is not an error, just a difference in convention. In the circuit world, a match is considered a complex conjugate match while in the full wave world a match is just a real match.

Third, there are two things you can do. 1) add an RLC boundary with the desired capacitance in series with the port and loop to cancel the loop inductance. 2) Use Designer to directly link a capacitor to the pin created by the lumped port.

Fourth, Driven Modal or Driven Terminal does not matter for this simulation.

Fifth, the user who mentioned the creation of an output variable that defines a complex conjugate match is correct. In the out put variable you can define your 'own' S parameter using the solved Z of the loop and a user defined complex Zref for your reference.

Sixth...have fun

 

[flanello]

Third, there are two things you can do. 1) add an RLC boundary with the desired capacitance in series with the port and loop to cancel the loop inductance. 2) Use Designer to directly link a capacitor to the pin created by the lumped port.

I think, if you use this method the S11 parameter is only valid at the design frequency at which you have calculated the value of the capacitor. Because the reactance of the capacitor depends on the frequency.

 

[tallface65]

I disagree. HFSS requires an input of the capacitance. The reactance of the load is inversely proportional to the frequency and HFSS keeps track of this relationship correctly. (You can test this for yourself using a series RLC network made up of RLC boundaries) The S Parameters will be valid over the whole frequency range of interest. The limitation of the validity of the S Parameters will come into play when the capacitor can no longer be considered as a lumped element. If the capacitor becomes larger than ~1/10 lambda then the lumped element approximation used will no longer be valid.

Have Fun

 

[flanello]

Unfortunately I can't test it at the moment, but here is an example for clarification:
If, for example, i will normalize the S-Parameter to a reactance of 30 Ohm at a frequency of 100 MHz i would use a capacitance of 53 pF.
Normally, at a frequency of 200 MHz this 53pF capacitor would lead to reactance of 15 Ohm. So the S-Parameter at this frequency point is normalized to a reactance of 15 Ohm instead of the required 30 Ohm.
So, how does HFSS model the RLC boundary at different frequencies?