Loss and Impedance Measurements of Reactors in Operation

A reactor uses induction coils to suppress inrush current and harmonics in power conversion systems.

Introduction

A reactor is a component that uses induction coils to suppress inrush current and harmonics. In the power system industry and in the power electronics industry (inverters, converters, and similar equipment), this component is called a reactor. In the electronic components industry, the same basic device is called an inductor.

In recent years, demand has grown for better performance from the power conditioning systems used in photovoltaic and wind power generation, as well as from the inverters used in electric vehicles. Reactors used in the boost converters of these systems are critical components, and this article focuses on reactors used in power electronics applications.

Table 1: Comparison of reactors and inductors.

Types and applications of reactors

Reactors are classified by structure, connection method, and purpose, but electrically they fall into two categories: AC reactors and DC reactors.

AC reactor

Purpose: protection against inrush current, harmonic suppression, and power factor improvement.

An AC reactor is inserted on the input or output side of a power converter such as an inverter. Insertion on the input side protects against inrush current into the inverter, while insertion on the output side suppresses distortion and harmonics in AC signals. Suppressing current harmonics improves power factor, and the AC reactor is also effective at suppressing surge voltage.

DC reactor

Purpose: smoothing currents and suppressing harmonics.

A DC reactor smooths pulsed or pulsating currents flowing in a DC circuit and suppresses the harmonic components of those currents. In a step-up or step-down chopper circuit, an input DC voltage is stepped up or down according to the ON/OFF time ratio, and the DC reactor is the key component of that circuit. Current flows as the chopper switches ON and OFF, and the DC reactor provides a feedback effect that keeps the output voltage constant based on the switching time ratio.

AC reactor inserted on the output side, between an inverter and the power system.

DC reactor in a step-up chopper circuit ahead of an inverter.

Understanding reactor characteristics

For an ideal reactor (or inductor), the reactance is called inductive reactance and is expressed as XL. Inductive reactance is defined as XL = 2πfL and causes the current phase to lag the voltage phase by 90°.

E (voltage) [V], f (frequency) [Hz], I (current) [A], L (inductance) [H], XL (reactance) = 2πfL [Ω], Z (impedance) = -jXL = -j(2πfL) [Ω].

In practice, a reactor also has a resistance component and parasitic capacitance, and environmental effects such as temperature change need to be evaluated as well. Accurately capturing these characteristics requires the right measuring instruments.

Challenges / Demands

Measurement of high-frequency drive reactor in operation

LCR meters, impedance analyzers, and network analyzers are commonly used to measure the resistance, capacitance, inductance, impedance, and reactance of electronic components (Table 2). Because these instruments apply a sine wave from an internal signal source to the component under test, they cannot measure loss and harmonics under actual operating conditions. In addition, if the input signal contains a DC component, the input channel of the measuring instrument may be saturated or damaged.

For a reactor with an iron core structure, it is necessary to measure how magnetic flux density, impedance, and reactance change with the input signal level under actual operating conditions. Depending on the frequency band, power meters and power analyzers are well suited to measuring reactors in operation. Items to be measured in an operating state include:

• Loss
• Impedance, reactance, inductance, and resistance
• Harmonic (13th and higher) suppression effect

Need for wide bandwidth and high-speed sampling

A square wave used for switching includes higher harmonics as well as DC components. The level of the nth harmonic is 1/n, and the signal does not drop below -20 dB (1/10) until the 11th harmonic. For this reason, measuring a square wave requires an instrument with a bandwidth from DC to a frequency 10-20 times the input signal frequency. For example, a switching frequency of 100 kHz requires a bandwidth of 1-2 MHz or higher. The rising and falling edges of a square wave are especially steep, so accurately converting the waveform to digital values for power calculation requires wide bandwidth and high-speed sampling.

On the other hand, if the instrument's measurement bandwidth is too wide relative to the target signal's bandwidth, unwanted noise can affect the measurement. Depending on the situation, filters or averaging functions should be applied appropriately.

Measurement of high voltage and large current

Because reactors in power systems and power electronics carry high voltage and large current, the measurement setup needs a power analyzer, probe, and current sensor with a bandwidth and input range that can accept the voltage and current being measured, along with connections and wiring that eliminate noise and measurement error. When measuring a reactor driven at high frequency in particular, a sensor capable of measuring from DC is needed, since DC components are superimposed on the signal.

Measurement under low power factor conditions

The phase difference between the voltage and current of a reactor is approximately 90°, so the power factor is close to zero. When measuring with a power analyzer, even a small phase error can significantly affect the results. It is essential to select a power analyzer with a function to correct phase difference, including that of a current sensor, and with specifications that enable high-precision measurement even at a low power factor.

Table 2: Comparison of instruments that can measure reactors.

Proposal / Solution

Yokogawa's Precision Power Scope PX8000 and Precision Power Analyzer WT5000 enable measurement of a reactor's loss and impedance under actual operating conditions.

Measurement bandwidth of power analyzer

The WT5000, with a 1 MHz measurement bandwidth, is sufficient for measuring reactors used in 50/60 Hz power supply systems. For more accurate measurement of reactors used in applications with high switching frequencies, such as power conditioning systems and EV batteries, the PX8000, with its 20 MHz bandwidth, is recommended.

Observation of voltage and current waveforms

The WT5000 can sample voltage and current waveforms at 10 MS/s, and the PX8000 at 100 MS/s, and display them directly, removing the need to connect a separate oscilloscope for waveform observation. Because high voltage and high frequency voltage are being measured, it may be necessary to use a damping resistor in some cases.

Example of simultaneous measurement and display of power parameters and waveforms on the PX8000, using a function generator as the input.

Measurement of reactor loss

The reactor is one of the components that reduces the efficiency of an entire system, and the first step toward improving efficiency is to accurately grasp the degree of reactor loss. Under actual operating conditions, skin effect from high current and high frequency, as well as changes in the properties of magnetic materials due to temperature rise, are also expected to affect efficiency.

Method 1: measuring directly across the reactor.

Method 2: measuring Pin and Pout on either side of the reactor.

Method 1

The reactor's power consumption P (total loss) is measured directly from the current flowing through the reactor and the voltage between the reactor's two terminals. Because the phase difference between the applied voltage and the flowing current is close to 90°, the power factor is very low. However, measurement errors of the WT5000 and PX8000 at zero power factor are very small, which allows high-precision measurement even under these conditions.

In this method, common mode voltage is applied directly to the power analyzer, which makes it easier for noise to enter the measurement. In addition, because the system under test is driven by a high-voltage switching waveform (a square wave), anti-noise measures are essential for the entire measurement system.

Iron loss can be obtained by subtracting the copper loss, calculated from the reactor's winding resistance r and the current value I, from the total loss P:

P (iron loss) = P (total loss) − r (winding resistance) × I (current)²

Iron loss in magnetic materials can also be obtained by measuring the B-H hysteresis loop. When the vertical axis is magnetic flux density B and the horizontal axis is magnetic field strength H, the area inside the hysteresis loop represents the iron loss, and the PX8000 supports this calculation as well.

Method 2

This method measures power from the voltage and current before and after the reactor, then calculates the reactor loss as the difference:

Reactor Loss = Pin − Pout

This approach can reduce the effect of noise caused by common-mode voltage. However, deviation between input channels (elements) introduces a measurement error. That error can be reduced by taking two measurements with the input channels swapped, then averaging the results.

Measurement of reactor impedance

Apply an AC signal E (frequency f) to the reactor, and measure the reactor loss along with voltage U, current I, and phase difference θ using a power analyzer. Impedance, reactance, inductance, and resistance can then be calculated using a user-defined computation function:

• Impedance Z = U / I
• Reactance X = Z × sin (θU − θI)
• Inductance L = X / 2πf
• Resistance R = Z × cos (θU − θI)

Alternatively, parameters such as impedance and reluctance can be calculated for each order of the switching frequency of a series or parallel equivalent circuit, using the harmonic measurement function. For details, refer to page App-16 of the PX8000 Precision Power Scope Features Guide.

Noise countermeasures

Protecting against noise is especially important when measuring under low power factor conditions:

• As far as possible, measure on the ground side of the circuit.
• Keep the measurement target, wiring cables, and power analyzer away from equipment that generates noise.
• Keep wiring cables as short as possible, and make them into a twisted pair.
• Wrap cables with an electrostatic shielding cover and connect it to ground.
• Cover the equipment with a metal cover.

For accurate measurements

A current sensor has a through-type structure, in which the current flowing through the primary wiring is detected by the winding of the sensor's electromagnetic core. This makes it necessary to pay attention to the following points.

1. Choose a current sensor with an appropriate rated current and frequency bandwidth, and pay attention to the current range setting on the power analyzer.
2. Wiring considerations:
   • Place the primary wiring in the center of the current sensor.
   • Make sure the primary and secondary wiring do not interfere with each other. Keep the secondary wiring as short as possible, keep it away from the primary wiring, and avoid running the two in parallel.
   • AWG24 or heavier gauge is recommended for the secondary wiring material. A twisted pair cable may be more suitable than a shielded cable for inverter measurements.
   • Secure reproducibility by fixing the current sensor and cables in place.
3. Use the amplitude and phase correction function. Because reactor measurement is performed under conditions where the power factor is close to zero, even a slight phase shift results in a measurement error. When an AC/DC current sensor or current clamp probe is used, its amplitude error and the phase difference between the voltage and current signals can be corrected to enable more accurate power measurement. The PX8000 and WT5000 include a phase difference correction (de-skew) function. For reactor measurement using a current sensor, this function corrects the phase difference (transmission time difference) of the current sensor and reduces the error factor on the time axis at the reactor's drive frequency.
4. Use individual Null to remove offsets. Null is a function that resets the offset value to zero while wired, including through an external current sensor. This is especially effective for step-up/step-down converters and chopper circuits, where DC components are superimposed. Null can be set individually to ON, HOLD, or OFF for each input. It is recommended to perform zero level compensation, which compensates for the zero level in the internal circuit, before executing the Null function, and to warm up the instrument before performing Null or zero level correction.

Improvement of reproducibility by correcting an entire measurement system

To improve reproducibility, use a function generator and a high-frequency amplifier to apply the highest possible current to a standard capacitor (or inductor) whose loss (W) is known, at the frequency at which the standard capacitor is actually switched and driven. Correction is performed by adjusting the time difference data while measuring the capacitor's loss with a WT5000 or PX8000, so that the measured capacitor loss is brought closer to the pre-specified loss of the standard capacitor using a correction function.

Conclusion

• The PX8000 or WT5000 is ideal for measuring reactors in an operating state.
• Wiring and instrument placement need to be arranged so as to eliminate noise.
• By using the phase difference correction (de-skew) function of the power analyzer, measurement error can be reduced at the reactor's drive frequency.
• Correction values can be obtained for an entire measurement system, including a current sensor and wiring, by using a standard capacitor.