Help with circuit simulation in HFSS

[ZincBear]

Hi all!

Im having trouble with HFSS. Im not sure what i am doing wrong, but the results differ wayyyy too much from momentum simulation. Attached is my simulation file... pls enlighten me of any mistakes i could have made.

Here i am using a 10mil thick alumina substrate. Dimensions of circuit have been double checked already. I have also declared a 0.2mil thick ground on the bottom. Im not sure if thats the right thing to do. But, something is very very wrong...... Pls advise...

Thank you!

 

[Ivan_]

Hi,
there are several incorrect settings in your HFSS design. First of all, use a driven modal solution type. Moreover airbox, ports and dielectric are not correctly set: use wave ports, set the dimensions the way the height of the port is 4-5 times the substrate height, starting from the ground plane. The port width must be 6 to 8 times the width of metal line touching it. This allows to keep account for fringing effects. Do not set a too wide or high ports since higher modes could be excited. This concept is the same for each port (1 & 2). A surface of the port must face the interior volume, while the other one must face the non-material background outside the airbox. This means that X dimension of substrate, gold ground plane and air box must have a length equal to the X distance between the two terminations of the line. Let the airbox to start from the ground and to be at least lambda0/4+dielectric thickness with radiation boundaries over all the faces except the one overlapping the gnd plane itself.
About the Y size, in oder to reduce border effects , set the airbox, dielectric and gnd lengths to be at least 3 times the free space wavelength. These object must be centered at port's middle point location along Y. I suggest to fix the middle point of port 1 at 0,0,0 location in the 3D coordinate system, in order to simplify the drawing, thanks to a reference point for it. Furthermore, learn to use project variables: define a variable for the dimensions of each object by putting $VariableName and definying the desired value.

Hope this helps.

I.

Added after 8 minutes:

By the way, if you want you can draw a rectangle instead of the box in place of the ground plane and assign 'finite conductivity boundary', choosing gold as material and assigning the desired thickness. You can do the same with signal line. Simulation sweep cannot be 'fast' but must be 'interpolated' if you use a non-zero thickness for finite cond. boundaries.

 

[ZincBear]

Thank you for the quick reply... it has been very helpful! However, i have further questions.

Since one side of the port should be facing the non-material background outside the box, does this mean that the substrate should terminate where the circuit terminates?

Additionally, for complex 3D shapes, is there any other way to "draw it into HFSS"? What is the usual practice? Does most professional users use the coordinate system, or some other software, and then importing the drawing into HFSS?


Thank you so much for the advice!

 

[Ivan_]

The substrate terminates where the circuit terminates with ports, the other sides must be larger than the circuit the way border effects can be reduced.
For drawing complex structures you can try importing a dxf file, or try obtain complex structures from the combination of simple blocks, using new custom system coordinate systems to be more facilitated in drawing them (see HFSS help, it'sn quite simple). Or you can draw polylines and create objects from them.
Bye