Plotting Ferrite Core Impedance and S Parameter Curves

An earlier newsletter showed how to model ferrite beads using a double parallel RLC plus series resistor model. The model parameters for the seven passive components were easily found by Micro-Cap's Model program optimizer. So the question is do the S parameter curves match as well? The short answer is yes. This article will show you how to plot those S parameter curves and how to compare them with those provided by the manufacturer. To start with let's consider the BLM21AG102SH1D part from Murata. Here is its Z vs. F plot. To model this device we load the C:\MC11W\LIBRARY\FerriteBead_Murata.mdl file from the Model program. We then enter the Z and F values from the above curve. The more data points you enter the more closely the curve matches the plot above. One thing to remember is that to get a good fit you need a balance of data points over the plot. The optimizer treats each data point with equal weight so it will fit more closely where there are more data points. After entering a suitable number of points and optimizing, the plot looks like the following. The optimizer has fitted the seven RLC values to the Z vs. F curve. The next step is to produce the model. Select Model menu / Create Model for This Part. When the dialog box comes up specify a file name for where to place the model. In this example we'll use the default C:\MC11W\LIBRARY\FerriteBead_Murata.LIB. When you click on the OK button, the program will create a .SUBCKT model using the circuity name like this: *Murata BLM21AG102SH1D .SUBCKT Ferrite_Bead_ 1 2 L1 1 3 1.798u R1 1 3 800.67 C1 1 3 1.24p R2 3 4 347.92 L2 3 4 5.019u C2 3 4 4.024p RDC 4 2 1 .ENDS Ferrite_Bead_ * Copy the BLM21AG102SH1D text. Change the subcircuit name by selecting the Ferrite_Bead_ text and then pasting in the BLM21AG102SH1D text. This identifies the subcircuit by the Murata part name. Do the same to the .ENDS text name. The circuit we'll use to show the Z vs. F curve is this one, FerriteBead_Z_Plot_Murata.cir, where we've pasted the entire BLM21AG102SH1D. subckt statement into the model page. Here is its Z vs. F curve: The curve shows a plot of the magnitude of Z, the real part, and the imaginary part. All match well to the original data. How well will the generated model match the S parameter curves? We'll use this circuit to answer that question. This circuit measures the four S parameters S11, S121, S21, and S22 for both the .subckt model on the left and the measured S parameters provided by Murata as read into the circuit on the right. The .subckt model created earlier has been pasted into this circuit on the Model Page. The circuit on the right is a collection of 2-ports that model the part by reading in the provided Murata S parameters. Since the circuit is a linear two port, the S12 and S21 parameters are identical as are the S11 and S22 parameters, so we plot only S11 and S12. In this plot the blue curves are from the subcircuit model and the red curves are from the Murata S parameter data file. The match is well within 5% and very similar to the Z vs. F curve match. What this suggests is that you can match a Z vs. F curve and the resultant model reproduces the S parameter plot as well. That is a direct result of the fact that the model is a linear collection of passive devices, so nonlinearities play no role in the models.