Measuring S-Parameters of High-Q Caps
The S-Parameters for our SMT capacitors found in our MLCSoft design program are gathered as follows.
A Vector Network Analyzer (HP8720C) is used to measure S11 magnitude and phase and S21 magnitude and phase. The S22 and S12 values are assumed to be identical to the S11 and S21 measured values since a capacitor is a symmetrical 2-port device.
Fixturing is critical to accurate S-parameter measurements, as well as a good method for calibrating (de-embedding) the fixture. Along with the above mentioned Network Analyzer, we at Johanson Technology (JTI) use an Intercontinental Microwave (ICM) test fixture (see their "total solution" fixtures at www.icmicrowave.com) optimized for measuring microwave chip capacitors as well as an ICM calibration kit for de-embedding. The calibration kit uses the TRL* (Through-Reflect-Line*) type of calibration instead of the more common OSTL (Open-Short-Through-Load) calibration method. In a case such as this where measurements are being made in a microstrip line environment to greater than 20 GHz, the TRL method is the better choice for de-embedding.
After de-embedding, the physical length of the capacitor under test, as well as the discontinuity impedance at the 2 points from the microstripline to the capacitor terminations are included in the raw S-parameter data. Measurement including these two effects is desirable because these two effects will be encountered when the capacitor is mounted in actual design situations. In other words, the capacitor is de-embedded to its terminations during measurement (which will cause the capacitor's physical length to be included in the S21 phase). Nevertheless, take note that the S-Parameters reported in MLCSoft has the capacitor's physical length removed mathematically so that the reported S-Parameters can be a good reference value for various mounting configurations (See the JTI application note "Mounting"). The S-parameters reported will thus be accurate for a given capacitor mounted in a 50 ohm microstrip environment that is identical to that in our ICM series mount fixture. This 50 ohm microstrip environment is on alumina substrate (used during calibration, and also for launching to/from the DUT), except for directly under the DUT where the substrate is air. Putting the alumina substrate continuously under the DUT presents the problem of adding shunt capacitance (mostly due to the high epsilon making these shunt capacitances larger) at both discontinuities which greatly distorts the S11 results. These S-parameters can, therefore, be put into a microwave design CAD program, and accurately simulate what the mounted capacitor will act like in the actual design. Nevertheless, the method of mounting the capacitor in the final application could alter the discontinuity impedance and cause the reported S-parameters to differ significantly from those created in a particular application (again, see the "Mounting" application note. Issues such as quality and quantity of solder, series versus shunt mounting, width of transmission lines, or the dielectric constant of the PC board the capacitor is mounted on will all have an effect on this.
The JTI S-parameter data is provided on CD-ROM in .S2P format, and is available from the factory. For immediate needs of S-parameter data as measured on our fixture, preliminary engineering data for select capacitor values, could on request, be e-mailed to you in Excel 5.0 format.
Below are some sample S-parameter charts for a "C" series mutilayer capacitor of 0603 size and capacitance value of 10 pF. Note that the Series Resonant Frequency (which is about 3 GHz for this capacitor) is defined as the S11 magnitude dip when measured in a series configuration as we used here, and includes the mounting effects as described above.
Last Updated: October 1999
For more information please contact our Application Engineers.