Virtual instrument determines magnetic core's B-H-loop characteristics
Inexpensive solution combines a laptop, DAQ card, and LabView.
Michael Nasab, Circuit Mentor, Boulder Creek, CA; Edited by Brad Thompson and Fran Granville -- EDN, January 5, 2006
To design an inductive component that contains a magnetic-core material, an engineer must accurately measure the material's characteristics. A magnetic core's dynamic hysteresis loop, or "B-H curve," contains valuable information about core losses and other magnetic parameters. Unfortunately, commercially available magnetic-loop-analysis instruments are expensive and thus impractical for small-scale research labs and manufacturers. This Design Idea describes a virtual instrument that uses a desktop or notebook computer with an analog data-acquisition card and National Instruments' (www.ni.com) LabView software (Version 7.1 or above). In operation, the software extracts B-H-loop information, core losses, and other magnetic parameters at a reasonable cost per measurement.
Figure 1 shows the test fixture for a magnetic-core-based device. The device, T1, comprises a sample of core material and two windings with equal numbers of turns. A precision current-sensing resistor, R1, samples the excitation current that induces a magnetic field in the core. The voltage drop across R1 is proportional to the excitation current and the magnetic field, H. A network comprising resistor R2 and capacitor C1 integrates the voltage induced in the secondary winding. The voltage across C1 is directly proportional to the flux density, B, in the core. In practice, R2's value should be much larger than capacitor C1's impedance at the operating frequency. (Textbook descriptions of the circuit suggest a ratio of 100-to-1.)
Components' tolerances and characteristics affect measurement accuracy. Use a noninductive, 1Ω, 1%-tolerance resistor of appropriate wattage rating for R1, and select a low-leakage, low-dielectric-absorption, polyester- or polypropylene-film capacitor with tight tolerance for C1. To acquire and view the data, you can use a dedicated virtual instrument using a National Instruments PCI-6024E data-acquisition card and LabView. The software features NI's Express VI (virtual-instrument) technology that greatly simplifies the creation of user-designed data-acquisition and -manipulation features. This application uses only two data-acquisition analog-input channels: Channel 0 acquires magnetic-field readings (H) for display on an x-y chart's x axis in units of ampere-turns per meter, and Channel 1 captures flux density (B) in tesla units for the y-axis display.
At low frequencies, the core's hysteresis losses predominate, whereas eddy-current losses become more apparent at higher frequencies. A wattmeter-style algorithm calculates core losses, but you can easily substitute your own mathematical expression into the VI block diagram's formula node. LabView also can save the data and export results in Microsoft's (www.microsoft.com) Excel-spreadsheet format or into other programs for further analysis.
You can use another of the data-acquisition card's eight differential analog-input channels to determine inductance. To do so, measure the voltage across the device's primary winding and calculate its rms value. The ratio of the voltage to the rms current as measured through R1 determines the magnitude of the winding's scalar impedance, XL. Then, you can calculate the inductance from the following equation: L=XL/2πf, where f denotes the frequency of the applied excitation voltage.
Figure 2 shows a hysteresis curve for a 3B7-mixture ferrite-pot core prepared with 100-turn primary and secondary windings and measured at 60 Hz. For comparison, Figure 3 displays the 60-Hz hysteresis curve for a 100W power transformer wound on a toroidal core composed of grain-oriented steel. The toroidal core's wider loop indicates greater hysteresis, a characteristic that saturable-core power inverters exploit. To apply 60-Hz excitation, you can drive the device's primary winding from a stepdown (isolation) transformer powered by an adjustable-output autotransformer, such as a GenRad (www.ietlabs.com) Variac. While observing the B-H curve display, gradually increase the primary voltage until the flattening of the hysteresis loop's upper and lower portions indicates core saturation. No calibration is necessary if you use precision. However, when evaluating core materials, you may need to experiment with different numbers of turns to obtain the windings' ampere-turns value for optimum results.
For tests at 60 Hz, use a 267-kΩ, 1%-tolerance resistor for R2 and a 1-µF polyester-dielectric capacitor for C1in the integrator network. Depending on the number of turns and the current necessary to obtain a usable output voltage, a few volts of ac excitation is usually sufficient to run the test. For core measurements at higher frequencies, use a signal generator connected to a power amplifier and alter the RC integrator's component values for proper operation at the frequency of interest. Although the application does not use an analog output from the data-acquisition card, this output can serve as a sinusoidal-signal source for the power amplifier.
Review the electrical specifications of the card you plan to use and avoid exceeding the card's peak-to-peak differential- and common-mode input voltages. If the excitation voltage approaches or exceeds the card's ratings, add a 10-to-1 resistive-voltage divider to limit the applied voltage and compensate for the attenuator's losses by adding a factor-of-10 gain multiplier in the software.
You can download a copy of the VI that this Design Idea describes from www.circuitmentor.com/services.htm. You can also obtain a trial version of LabView from NI's Web site at www.ni.com.
