hfss lumped port

A lumped port can be assigned to a rectangle. Does the size of the rectangle affect the simulation results? Should it be as small as possible?

Hi waltzsu,

You can refer to following link and find the attachments name: "HFSS_full_book_v10" and "Port_tutorial"

https://www.edaboard.com/ftopic82440-0.html

They will tell you how to define a lumped port and size of the port.

Hope it helps you

Thank you very much. I have downloaded the port tutorial. It is helpful. But I did not have enough points to download the full book.
anyway, Thank you very much.

hi waltzsu,

Try with HFSS_V10_full_book at:
http://www.megaupload.com/?d=LRDGNI0G

have fun!

thank you so much.

waltzu:

The larger the rectangle, the bigger the electromagnetic discontinuity to your lumped port. The key is that the lumped port is on an ideal wire that connects one metal body to another body or PEC boundary condition. So you will always have that discontinuity as current flows from the big body down to the connection point of the ideal wire.

There is no way to calibrate out, or de-embed that discontinuity in HFSS. So the technique works well as long as your frequency is low enough so that the discontinuity doesn't throw your simulation results.

--Max

in hfss size doesnt matter because lumped port just define the type of excitation....b/w the ports so not to worry abt the size

I'm agree fully.
Also, with the Lumped Port you can execute one-mode simulation only.

akedar:

Your statement really depends on your application. If you want an excitation for a small body at low frequency, then I would agree that the error introduced by using a lumped port is small enough that it can be neglected for practical purposes.

But for most practical microwave design, I believe this is not true.

For example, try taking a 50 ohm transmission line on 100 um GaAs substrate. Simulate this example as a reference result.
Next, slice a 100um long section from the middle, and put a lumped port on each side of your gap. Simulate this example (1 waveport on each end, and then the 2 lumped ports in the middle). Simulate a 100um section of the transmission line in another example (you will need to use reference planes from the larger original problem so that you don't get your waveports too close together).
Cascade the results of your 100um transmission line with the lumped ports from the 4-port example, so you get a model that should be equivalent to the longer, non-segmented transmission line.

This example is similar to the case on an IC where you want to simulate circuitry on both sides of the transistor.

I think you will agree that lumped ports do indeed have limitations if you use them in situations where the error that they contribute to the problem negates their value in modeling.

--Max

If the size of the lumped size is too big, you may have to change the dimension of the simulated structure in order to leave the space to the port. In this case, the size of the lumped size does matter.