Using Coupled Inductors and Inductor Cores

Source: OrCAD Technical Support

Revised by: Brian Hirasuna, April 1999

Edited by: David Busdeicker, March 2000

This note gives information on coupling inductors and on modeling saturable cores to add hysteresis and saturation. The inductor coupling symbols may be used to couple up to six independent inductors (with or without Core models) on a schematic. The coupling symbols K_LINEAR and KBREAK are described, as well as Core models in the MAGNETIC library.

- The K_LINEAR symbol in "analog.olb" is provided for specifying linear coupling, between inductors.
- Draw the schematic and assign the desired names (reference designators) to all of the symbols. (Reference designators are assigned automatically, but may be changed by double-clicking on them.)
- Place one coupling symbol, K_LINEAR, anywhere on the schematic, for each group of coupled inductors. These symbols have no pins; they are represented by the letter K enclosed in a box.
- Double-click on each coupling symbol (on the K-in-a-box, not the attributes) and enter the reference designators for the coupled inductors as the values for L i (i=1,2,...,6).
- Set the value of the COUPLING attribute to the value of the coupling factor, K.

- Draw the schematic and assign the desired names (reference designators) to all of the symbols. (Reference designators are assigned automatically, but may be changed by double-clicking on them.)
- Select the coupling symbol for the desired CORE model from "magnetic.olb", and place one coupling symbol, anywhere on the schematic, for each group of coupled inductors. These symbols have no pins; they are represented by the letter K enclosed in a box.
- Double-click on each coupling symbol (on the K-in-a-box, not the attributes) and enter the reference designators for the coupled inductors as the values for L i (i=1,2,...,6).
- Set the value of the COUPLING attribute to the value of the coupling factor, K.

A generic symbol, KBREAK, is provided in "breakout.olb" for specifying arbitrary nonlinear magnetic core models. KBREAK has a preassigned model attribute, but its corresponding model in "breakout.lib" has no parameters. CORE models can be created through use of the PSpice Model Editor. Example extraction of B-H loop in the PSpice Model Editor

- In the Model Editor (formerly called Parts), choose File/New. Type in a new library name. Choose Model/New, then choose CORE, then enter the name. Click on Hysteresis Loop, go get the specification entry screen.
- First set the initial permeability (at the bottom of the specificatin screen) according to datasheet specifications.
- Pick 3-4 points along both limbs in the first quadrant of the B-H loop. Pick the first point on the y-axis intercept (for example, (0, 600) ). Pick another point at the x-axis intercept (for example, (0.04,0) ). Pick a fourth point where the limbs converge. Pick 2 additional points on either limb, for the same x-axis value (for example, (0.12, 2000) and (0.12, 1500) ).
- Input initial permeability - 10000 from datasheet for W material
- Enter H, B intercept points (0, 600) (0.04,0)
- Enter point where loop closes (0.4,3000)
- Plot/Display
- Plot/X-axis settings, set Data Range 0 to 0.45
- Extract/Parameters - rough fit of loop is achieved
- Enter points (0.12, 2000) (0.12,1500), one on each limb
- Extract for final fit

After extracting the Jiles-Atherton parameters, in the Model Editor set the following parameters (in the spreadsheet) according to the datasheet specifications: AREA=<cross section area in cm^2>

PATH=<magnetic path length in cm>

The KBREAK symbol in the BREAKOUT library can be used with this new model. To use KBREAK with your new CORE model:

- In Capture, place an instance of KBREAK from the BREAKOUT library.
- Double click the symbol to change the Implementation property to the
new model name. This is the name given to the CORE model in the Model Editor.

NOTE: If you are using using Schematics v8.0, use Edit Model-->Change Model Reference to change the model name. - Check that the model library (created by the Model Editor) is configured in the Libraries tab in your Simulation Settings.

The "dot" convention for the coupling is related to the direction in which the inductors are connected. The dot is always next to the first pin to be netlisted. When the inductor symbol, L, in the ANALOG library is placed without rotation, the "dotted" pin is the left one. Edit/Rotate (<Ctrl R>) rotates the inductor +90deg, which makes this pin the one at the bottom, etc.

Certain rules must be followed when setting the properties for coupling symbols and the inductors they affect. Nonlinear CORE models may be applied to one or more inductors, so:

- The L 1 property must have a value (a reference designator of an inductor); the other L i may be blank.
- The Implementation property must reference a CORE model.
- The VALUE properties of the affected inductor symbol(s) must be set to the number of windings (turns).

- The L 1 and at least one other L i attribute must have values (reference designators); the rest may be left blank.
- The MODEL attribute must be left blank.
- The values assigned to the inductor symbols must be in Henries.

The transient analysis is run to 4 seconds (final time) and has a maximum time step of 0.01 seconds. The inductor (L1) has 20 turns around the K528T500_3C8 torroid core.

X-axis unit is Oersted; y-axis unit is Gauss. Restrictions on coupling coefficients for systems of coupled inductors When running a transient analysis, it is possible to have convergence difficulties with systems of coupled inductors. Systems of coupled inductors with physical solutions have the property that the coupling matrix (K matrix) be positive definite [1]. For three inductors this may be tested by the following inequality: K 12

[1] Yilmaz, Tokad and Myril B. Reed, "Criteria and Tests for Realizability of the Inductance Matrix," Trans. AIEEE, Part I, Communications and Electronics, Vol. 78, Jan. 1960, p.924.