modelitics

Hi,

I am trying to characterize SMT components I am using for a design. The component include RF caps, inductors and resistors operating from 1-2 GHz.
These components behave poorly towards 2 GHz, so I want accuracy in my results.

My current approach has been placing the SMT components on a microstrip trace with a cut. I then obtain 2-port scattering parameters, de-embed to the device planes and use those results. Unfortunately, the results are not symmetrical which are not expected. Does anyone have any tips or ideas?

What you did is not bad, but requires careful calibration of the setup.

The best evaluation of SMT components is to use this kind of expensive stuff from Agilent:

http://cp.literature.agilent.com/lit...5952-1430E.pdf

Is expensive, but for high-volume production is mandatory to have something like this in your lab.

Thanks for the advice, unfortunately buying that equipment is not really an option.

Hi

"Modelitics" is the world famous company that does exactly what you are trying to do. They have developed a vast library of models for almost all SMD manufacturers and applied them in all major circuit simulators. Perhaps you can visit theri web site and explore some of the references how they did it.

flyhigh

Thanks for the advice. I am going to try to get the modelithics library.

Does anyone have any suggestions as why my S11 and S22 do not match when measuing a surface mount component?
I am using SOLT calibration and then de-embedding the microstrip TLs afterwards.
For example, when measuing an inductor, the equivalent inductance determined using S11 and and S22 are the same, however the equivalent series resistances are not the same. They are close, but not the same the way the inductances are.
It's my understanding that they should be for a SMT inductor at a sufficiently loq frequency (1-2GHz).

Any advice?