How to model the PCB trace when we design RF circuit on the board?

Hi all, I have a question about how to model the PCB trace when we design RF
circuit on the board? Sometimes, we need put the passives such as RLC on board,
and the trace to make connection is necessary. But how to model it? I design the
RF circuit below 5GHz, any good estimation method or simulation tool can help this?

I wanna know is there any tool can output a lumped circuit model for PCB
trace?

in the RF simulators u can add a TL and this TL is the PCB trace which is modeled by the substrate dielectric constant , width , and length of the TL line

khouly

Hi
Use ADS library "tline multilayer"

Hi:

If it is for RF, I may not be necessary to obtain frequency independent RLC equivalent circuit. I think s-parameters should be good enough. You can use full-wave EM simulators for it. Please remember the strip thickness for the traces is critical. Try to use thickness model on it. Best regards.

But, I would like to get the information in voltge rather than power, so I need
to use it in non-50 ohms system. My platform is SpectreRF, I have no idea
if there is anything userful?

The free SonnetLite (www.sonnetsoftware.com, I work for Sonnet) will do an EM analysis of your PCB trace up to a size limit of 16 MBytes of memory, 4 ports, and 3 dielectric layers. The software has no timeout. Then, when you plot the data, you will have an Output button at the top of the window. You can select a lumped SPICE "Pi-net" model. Then, for every 2 frequencies analyzed, a lumped model is generated. The lumped model is good up to the frequency where the lumped model starts changing.

If you have access to the full version of Sonnet, you can also output a broad band SPICE model. Several customers are telling us that our broad band spice model is clearly the best in the industry, but I have not actually seen results from the others yet.

You can also output the R, L, C, and G matrices for an N-coupled line and it is ready for input to SPICE.

Probably a good idea to read the manual on any of these three options before you use them.

Hi, OMEsystem:

S, Y, Z-parameters are interchangeable. You can still get voltage information on S-parameters. Many tools can output frequency independent RLC from s-parameters. However, none of them can guarantee the passivity or the RLC be non-negative. Some SPICE can't accept non-negative RLC values and some can. However, it is quite easy to fail. Using s-parameters should be the best choice if your SPICE simulator can handle s-parameters well.

Best regards,