Learning Forum

[kagankarakusx]

Hello,

I have a reference PCB EDB file exported from Altium. I imported this layout into HFSS 3D Layout.

Now, I want to assign geometric parameters to the via properties and perform a parametric sweep using Optimetrics(Tuning). My goal is to add sliders (bars) for these variables, manually change them, and observe how the TDR graph changes.

The specific via properties I want to parameterize are:

• Via plating thickness (to be assigned as a variable)
• Hole size
• Annular ring (for all unused and used pads)
• Antipad (distance to GND pour on all layers)

Initially, I want to manually adjust these parameters and try to optimize them.

How can I achieve this? Is there any tutorial or can you explain the workflow step by step?

Thank you!

[jeffrey.w.acuff]

I'm not aware of a way to implement sliders bars to change variable values, but someone else may know more about that.

My method is to:

1. Assign variables to your geometries in the padstack editor, as well as non-padstack variables (like a square connecting two differential antipads creating a common antipad).
2. I believe padstack variables must be project-wide, that start with a $. I use $AP_R for anti-pad radius.
3. Assign starting, ending values and step size to a new sweep in parametric analysis.
4. View the results of the sweep in a TDR plot and choose the value set that aligns closest to your target impedance, or further refine the sweep.

There is a way to do it in OptiSlang as well i believe, but i'm not sure of the specifics in setting that up. I would be interested in further advice as well.

[kagankarakusx]

Hello, yes, I learned the method you mentioned; I can assign variables internally. Is there a method for creating slot-shaped antipads for differential connections? How do you create this frame? And do you bind the frame to this parameter as the antipads change?

You're creating the existing parametric sweep in optimetrics, right?

I can see the S parameters changing instantly when I adjust other variables as tuning, but there's no instantaneous change in the 3D model.

I'd also like to add a movement command for stitching vias for single-ended vias. It would be great to test the same way by changing the distance in addition to the normal trace transition via.

What I'm curious about is that when creating a rectangular plot in the results section, there's a transition from the frequency domain to the time domain. When I do this, I can see the TDR analysis result. Does this result assume there's a terminal at the output (I didn't add a terminal resistor in the 3D model)?
Is this where you're observing the results?

The next step is to create a side-by-side GIF of the s parameters and TDR results as the 3D geometry changes.

The next step is to apply all these parameters using optislang.