2
\$\begingroup\$

I am trying to simulate the filter impedances on LTspice. There are two images of the simulation results of the same pi-filter on LTspice and Redexpert. Although I am not sure the parameters are correct on the graph in LTspice, the C1 and C2 positions are reversed on purpose. But I may have made a mistake there. Do you spot any mistakes that I may have made to have these relatively different results?Insertion Loss as 4MHz

Insertion Loss as ~200kHz

\$\endgroup\$
3
  • \$\begingroup\$ Do your models of C1, C2 inherently include parasitic series inductance and/or series resistance (ESR)? \$\endgroup\$
    – glen_geek
    Commented Jul 3 at 13:51
  • \$\begingroup\$ They are selected as X7R caps from Würth as well, but not exactly the same components. ESR and ESH is given for C1 and C2. \$\endgroup\$
    – mocal_381
    Commented Jul 3 at 15:17
  • \$\begingroup\$ @mocal_381 I found the models for those caps on their website, so I was able to sim it with the cap and inductor models for those exact parts. The shape of the response in LTspice looks similar to the one in Redexpert but the frequencies are still off, using their models vs. LTspice default models made little if any difference in that. \$\endgroup\$
    – GodJihyo
    Commented Jul 3 at 16:08

1 Answer 1

Reset to default
3
\$\begingroup\$

You need to use Wurth's SPICE models, however, the peak response I get using Wurth's models is different from the Wurth modeler. Perhaps Wurth has better models for their modeler. The general form of the LTspice simulation with the Wurth models is close to what the Wurth modeler predicts.

You can download SPICE models from Wurth's web site.

LTspice simulation

\$\endgroup\$
2
  • \$\begingroup\$ So the main reason that creates the difference is series and parallel capacitance/resistance/inductance. I found the values in the .lib files of Würth model library, as you said. I opened them as a .txt file. Thank you very much. Do you know how to plot the output impedance correctly on this filter circuit? Or is plotting the current on the load resistor correct for AC analysis? \$\endgroup\$
    – mocal_381
    Commented Jul 4 at 6:04
  • 2
    \$\begingroup\$ @mocal_381 To measure output impedance, set V1 to zero and drive the output with a AC current source set to 1 A. The impedance is merely the voltage across the current source (you may need to a a minus sign depending on the direction of the current source). You can set the graph from Bode to Cartesian (right-click on the y-axis label) to get sensible units. Remove the 100 ohm load if you are only interested in the filter's output Z. \$\endgroup\$
    – qrk
    Commented Jul 4 at 16:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.