Lange Coupler Designer with MWO， Designer and HFSS

Hi,

I have been working on a Lange coupler with MWO, Ansoft Designer and HFSS. What I do is I start my initial design with MWO or Designer, then I use HFSS to verify my results. However, the results from HFSS is very different from Designer and MWO.

I understand that HFSS should give me the most accurate results since it is a 3D simulator. However, I expect reasonable accuracy from MWO and Designer too. In this case, it is not.

I am not sure if the built in bond wire model in HFSS is accurate as well. Also, the boundary condition I am using is a radiation boundary. Does this sound right? I am using this type of boundary condition because I have seen this set up in a tutorial before. I am not sure if this is correct.

Has anyone use MWO and/or HFSS for designing a Lange coupler? If so, were the simulation results very different?

Thank you in advance!

Hello,

See my thread on lange coupler using AWR MWO with example file...
https://www.edaboard.com/thread49651.html

Also an example from AWR
https://awrcorp.com/download/faq/eng...erconnect.aspx

AWR Microwave Office is good for Lange coupler design & Optimization...
MLANGE1 - EM quasi-static model (more accurate) for fingers of the coupler only & use EMSight for Interface interconnects..(Circuit & EM Co-simulation)

Hi Manju,

I basically did what said on the your post. When I try to simulate the exact same structure in HFSS, I get a significant different simulation result. At this point, I just don't know which tool I should trust. It sounds like you have experience designing Lange couplers. Can you step through what you did for your design? I really need help on this!

Thank you!

That is propaganda by the 3D solver marketing guys.

The planar MoM solvers (which they call "2.5D" to make them look bad) also solve Maxwell's equations and the full 3D fields. There is no reason why a "full 3D" solver would be more accurate for a planar PCB problem. It only solves a more general problem, giving room for more numerical problems and more user mistakes. For planar work (e.g. PCB or multi layer modules), the planar solvers are usually more accurate.