Different results between Momentum uW and RF

[brunosaldana]

Hello. I'm a designing a 20dB coupler for 10GHz and I'm simulating it between 9-12GHz. However, I'm getting different results between Momentum uW and RF. Momentum RF(Blue curve) gives me excellent results but Momentum uW(Red Curve) does not.

I understand Momentum RF is suitable for electrically small RF components that do not radiate, but I get the following warnings:

-Layout is electrically large above 17GHz (space wave radiation)
-Substrate is electrically large above 7.78GHz (Surface wave radiation)
Consider: lowering the maximum frequency, simulating in mW mode.

I'm showing my port setup, substrate, coupler, and results comparison.

I have edge mesh enabled and have 80cells/wavelength. Is there something invalid with my setup for uW simulation?

Thanks!

 

[BigBoss]

-Try to simulate this coupler with sheet ideal metals first and don't try to define cover at first attempt. Use default cover of momentum.
-Try to use thinner substrate such as 0.25mm OR use different substrate because RO3010 has a high di-electric coefficient and this makes the coupler smaller. Use a different but lower di-electric coefficient with low loss substrate.
Otherwise , the momentum set-up seems correct.

 

[brunosaldana]

-Try to use thinner substrate such as 0.25mm OR use different substrate because RO3010 has a high di-electric coefficient and this makes the coupler smaller. Use a different but lower di-electric coefficient with low loss substrate.

Thank you for the suggestions. I calculated the width, length(quarter wave), and separation of the coupled lines using line calc. Shouldn't line calc have accounted for the high dielectric constant and substrate thickness?

 

[volker@muehlhaus]

I understand Momentum RF is suitable for electrically small RF components that do not radiate,

"Radiation" here also includes substrate modes, surface waves etc. Only Momentum microwave mode can model these effects, so if you see some difference it might be real!

 

[BigBoss]

Thank you for the suggestions. I calculated the width, length(quarter wave), and separation of the coupled lines using line calc. Shouldn't line calc have accounted for the high dielectric constant and substrate thickness?

No, Line calculator is a simple estimator, not real simulator.
Simulators can also not give the real results either, they are just specialized Calculators, no mystery.

RF and Microwave Engineering therefore is tough to work on. No visible voltage/current, no accurate estimation, no good prediction. Everything can happen in anytime anywhere.

But I can say that Microwave Mode is mora accurate at that frequency even tough the coupler is so small.
Try to expand the coupler using with lower di-electric then come-back.