


Measurement Conditions and Notes This describes the considerations and measurement conditions used in the creation of the PPI 1111N capacitor series scattering matrices for horizontally oriented parts. All comments apply to parts in the series microstrip configuration, i.e. capacitors that span a gap in the center trace of a microstrip transmission line and that have the planes of the electrodes parallel to the substrate surface. Scattering parameters of multilayer ceramic capacitors mounted on microstrip substrates depend not only on the dielectric and electrode capacitor structure, but on a number of substraterelated parameters. These include: (1) substrate dielectric constant and thickness; (2) mounting pad and trace dimensions; and (3) capacitor orientation (internal electrode planes relative to the substrate surface), horizontal or vertical. In consideration of the above, scattering matrices have been derived from electrical models created by Modelithics, Inc., a specialty vendor contracted by PPI. The models, derived from measurements on a large number of parts disposed on several different substrates, permit users to input substrate dielectric constant and thickness, landing pad dimensions, and  for some series  capacitor orientation. Detailed information is provided on the Modelithics model data sheet (found elsewhere on the PPI website) for each series. In all cases, the models are carefully vetted against experimental data. Measurements are made using coplanar probes; the coplanartomicrostrip transitions, as well as any other structures between the probes and the capacitorundertest mounting pad edges, are deembedded using statistical TRL techniques. Measurements are typically performed on a Vector Network Analyzer from 0.1 to 8 GHz for each capacitor value on each substrate. The scattering matrix is simply a text file; a sample is shown below: # GHz S MA R 50 The first line is a header, designated by the “#” character. Succeeding characters describe respectively the frequency units, parameter, measurement format, and characteristic impedance of the measurement system in Ohms. The data is arranged as follows: The first column is the frequency, in the units specified in the header. The next two columns are S11 magnitude (again, as specified in the header) and S11 phase. Succeeding column pairs represent S12, S21, and S22. It is important to understand that the upper frequency at which the scattering matrix is valid is either the 8 GHz measurement limit or the second parallel resonance of the device. Parallel resonances in the series microstrip configuration manifest as notches (“suckouts”) in S21 or S12, the insertion loss. It should be noted that some software simulators may require that artificial data  e.g. 0 dB return losses and very large insertion losses  be added at zero frequency to avoid passivity or other issues. In general, caution is suggested when using software simulation, since many programs attempt to extrapolate scattering matrix data both below and above the frequencies where it is valid – and often arrive at grossly incorrect results. Fig. 1 depicts a typical capacitor on the mounting pads of a microstrip substrate.
Fig. 1 A multilayer ceramic capacitor in a series configuration on a microstrip substrate Reference planes are always at the pad edges, as shown. The following parameters apply: Series: 1111N The last two parameters, “50Ohm Trace Width on substrate (mils)” and “Reference plane spacing (mils)” are for reference only. To determine the intrinsic properties and capabilities of a part, it is often valuable to characterize it with minimal exogenous parasitics, in this case, the fewest additional parasitic reactances beyond those of the basic capacitor spanning a gap in a microstrip trace. Generally, two conditions are necessary to accomplish this:
Fig. 2 General equivalent circuit for MLCC on microstrip (left) approximately reduces, for DC block, to lumped equivalent circuit for short length of transmission line (right) PPI has created at least one set of scattering matrices in each series where the reference planes closely coincide with the nominal part edges, and where the 50Ohm trace width on a particular substrate is close to the nominal part width. Mounting pad dimensional limitations in the models determined to what degree these conditions could be achieved. DISCLAIMER: Passive Plus Inc. (PPI) has made every effort to make this information as accurate as possible. However, no responsibility is assumed by PPI for its use or for any infringements of rights of third parties that may result from its use. PPI reserves the right to revise the content or modify its product line without prior notice. 

