X band RF resistor for stub isolation

[yefj]

Hello,In the links of the photos below, we have a 10GHz isolation stub.
So in order to make a good matching i put a 38Ohm idial resistor.
However in 10Ghz there is no ideal 38Oh, resistor at 10Ghz.
Is there a manufacturer you can recommend me ?
Thanks.

 

[volker@muehlhaus]

This is another strange yefi design.

It seems that your line width is wrong, and you have some impedance transformation going on. The proper design will work with 50 Ohm resistor at the radial stub.

 

[yefj]

This is another strange yefi design.

It seems that your line width is wrong, and you have some impedance transformation going on. The proper design will work with 50 Ohm resistor at the radial stub.

Hello Volker , this is a starting point .
my logic says to use 0805 and put 50 Ohm .
My substrate is 0.508 er=3.5 in the link shown below.trace thickness is 2oZ =0.07mm
In cst I get 50 ohm at 1.05 .
I use 0805 and 50 Ohm it got me -19db


Could you recommend me a component to use?
With S2P model.
Thanks.
https://www.rogerscorp.com/advanced...-duroid-laminates/rt-duroid-6035htc-laminates

 

[volker@muehlhaus]

In cst I get 50 ohm at 1.05 .

That looks good. Your problem might be the large (and wide) pads, try using a smaller resistor, size 0603.

One problem with your CST model is that it places the resistor terminals at the inner edges of the pads, this is not realistic. I don't know if CST offers a better way to model that. My Empire 3D solver has 3D resistor models that are more realistic.

 

[yefj]

Hello Volker , I will define the pads as 0603 .
And put the artificial resistor from the middle of the pads .( and ask CST how properly to do that)
Could you recommend me an accurate resistor with S2P model I could use ?
Thanks .

 

[volker@muehlhaus]

Could you recommend me an accurate resistor with S2P model I could use ?

For 3D modelling, I use the built-in models of my 3D tool. They model the internal structure and parasitics nicely.

 

[yefj]

Hello,given the model in the attached article.
when i have my resistor 50 ohms in parralel to 0.04pF and 0.0267nH it totally ruined the result. i get almost 0dB

i also tried to put S2p of the following manufacturer it shifted the center frequency of the isolation by 2GHz.
in real life if i manufacture something and it will be shifted by 2Ghz then my whole coupler will be ruined.
What do we do in real life to handle the parasitics problem?
Thanks.

Ultra Broadband Resistors (UBR) | KYOCERA AVX

The UBR Series provides the best combination of thin film resistor technology in a proprietary SMT package.

www.kyocera-avx.com

where did i go wrong?

https://www.vishay.com/docs/49427/vse-tn00.pdf

 

[volker@muehlhaus]

when i have my resistor 50 ohms in parralel to 0.04pF and 0.0267nH it totally ruined the result. i get almost 0dB

The inductor must be in series to the resistor, not in parallel. Only the capacitor is in parallel.

i also tried to put S2p of the following manufacturer it shifted the center frequency of the isolation by 2GHz.

I think this is caused by physical size (length) of the resistor. You need to be careful that you include that length once. If you model the physical size in CST (port size) and then insert the S2P, you might have included the effect twice. Not sure what port parasitics the lumped port in CST gives here.

Some EM solvers like Momentum or Sonnet offer port calibration to care about these important details. 3D solvers like HFSS or CST are not ideal for this work, because they ignore these details.

 

[yefj]

Hello Volker ,In simulation we have the ability to change everything .
In real life what are the common ways to fix frequency offset , or if even worst I got my matching ruined ?
Thanks .

 

[volker@muehlhaus]

In real life what are the common ways to fix frequency offset

Increase or decrease the length (radius) of the radial stub.

 

[yefj]

Hello Volker , I have found here 0406 resistor series it’s power rating very very low 0.125W
Do you know 0603 resistor I could order which has S2P models ?
Thanks .

https://datasheets.kyocera-avx.com/AVX-UBR-Series.pdf

 

[volker@muehlhaus]

No, I haven't searched for resistor manufacturers with S2P library. Long ago, I used Vishay resistors for a 24 GHz termination with radial stub, but I did not look for S2P models at that time.