HFSS port definition problem

[abhi_kolkata]

hi,

i am trying to simulate a 30GHz inset (50 Ohm line)fed microstrip patch antenna over 0.35mm thick substrate(permittivity = 12.9) and finite gnd and substrate (15x15 mm) using HFSS. wat should be the actual port definition with all its parameters(such as : wave port or lumped port with solution freq.,no. of passes, mesh refinement, sweep type) for this structure.i also want to know wat will be the deviation in that case from fabricated one(if i fabricate the same).

plz share ur views.

with regards

abhi

 

[adel_48]

hi,

for microstrip port definition you need to draw your structure inside an air box with width of around 10 * microstrip line width and height about 5 * substrate thickness. then you define the port as the whole airbox side, using a wave port.

the results should be very near to the measurements, I made some microstrip filters ad the measurements results were near the simulation, except that there might be some extra losses.

Best Regards,
Adel_48

 

[savedadogs]

I never make my waveport the same size as the side of the airbox. I always make it about 6-8 times substrate height and 4-6 times trace width. But it all depends on the substrate you are using (higher permittivity = room for smaller waveport).

 

[flyhigh]

Hi,

for low frequencies fields are dispersed so wide and high ports are needed. For high frequency, EM field in microstrip is concentrated in the vincinity of the track and in the substrate. It is a rough approximation but at high frequencies it might be needed to make ports smaller to avoid multimoding that is not present in reality.

flyhigh

 

[adel_48]

Dear flyhigh,

I think for both high frequencies and low frequencies the port fields has approximately the same distribution, and are concentrated near microstrip. but concerning the fields inside the structure they change with frequency exactly as you said.

best regards,
Adel_48

 

[costox]

This is a tutorial for HFSS Port (CPW)

 

[abhi_kolkata]

hi all,

i have tried the above mentioned geometry with a waveport of width=5*trace_width and height = 5*substrate_height. and keeping other specifications as solution type->modal, integration line with zvi and no normalization and no de-embedding and delta s = 0.02 and meshing refinement = 30%(in the solution setup).

whether that is ok ? and will it give reasonably good match with the fabricated circuit??

with regards

abhi

 

[masashisi]

i defined the striple port width as H and length as 5*W.
on high freg. result are close to planar EM, but on low freq. i got very deferent results. how to define a port for stripline. thanks