# Measure PWM motor efficiency

-November 01, 2001

Last spring, in response to California's recent electricity shortage, Governor Gray Davis signed a bill into law that provides financial incentives for upgrades that improve the efficiency of motor and lighting systems (Ref. 1). As a result, motor designers must strive to design more efficient motors, and designers of electromechanical systems need to place a higher priority on efficiency in new and upgraded systems.

Measurements play a key role in both new designs and system upgrades. Whether you design or specify variable-speed motors, you need to calculate a motor's efficiency by measuring both its electrical power input and its mechanical power output. Unfortunately, there is no industry standard covering the measurement of a motor's pulsed electrical input power. You'll have to choose from one of four measurement options, each of which has limitations.

## How PWM motors work

Most variable-speed motors get their power from pulse-width-modulated (PWM) signals. These motors often replace ones that rely on varying-amplitude DC or AC signals to control their speed. PWM motor-drive controllers chop sine waves or DC voltages into signals of varying frequency and duty cycle. A motor's speed results from the signal's average amplitude of a chopped voltage.

Figure 1 shows a simplified PWM drive controller schematic for a DC motor. The controller's input converter rectifies AC line voltage and filters it into DC. (The distribution network that carries the DC is called the "DC link" or "DC bus.") Then, the inverter circuit pulses the DC voltage based on the PWM input signal that comes from a digital motion controller, often called a "servo controller." Some drive controllers don't contain rectifiers and filter capacitors, leaving you to supply the DC voltage. Those controllers often are called "amplifiers."

 Figure 1. PWM motor controllers for DC motors convert AC line voltage into DC, then chop the DC voltages based on the PWM input signal from a servo controller.

A motor's PWM input voltage, which can reach to about 50 kHz, contains sharp rising edges. Those edges contain high-frequency energy from harmonics of the PWM signal's frequency. Because a motor presents an inductive load to the inverter circuits, its inductance filters much of the high-frequency energy. The high frequencies do little to rotate the motor, but the energy in those frequencies must go somewhere, and the high-frequency energy dissipates as heat.

 Figure 2. Current in PWM motors resembles DC with ripple, because a motor's inductance filters high-frequency energy from the PWM drive voltage.

Because of its inductance, a DC motor's current resembles DC with ripple (Figure 2). As the PWM signal's duty cycle increases, the average voltage increases as does the motor current. The motor's rotational speed increases.

Measuring efficiency

To calculate a motor's efficiency, you must measure its mechanical output power and divide it by the electrical input power. Measuring mechanical output power is fairly easy: You can use a torque meter to find the mechanical power based on the motor's speed and load.

Measuring input power is not as straightforward. The motor's input power (the drive controller's output) is an electrical quantity. Because PWM drive controllers chop their output voltages, measuring power in PWM signals is more difficult than measuring power in DC and unmodulated sinusoidal signals.

To measure PWM power, you can use a DMM, a power meter, a harmonics analyzer, or a DSO, although none of these is fully satisfactory. And, because the industry does not have a measurement standard, it's possible for the test results obtained by motor and controller manufacturers to differ from those obtained by customers.

The lack of a standard causes confusion. Customers often need to verify motor efficiency against published specs, but unless they use the same measurement method used by the manufacturer, they can't make fair analyses. Similarly, a motor manufacturer's efficiency specifications lack significance unless customers can replicate the measurements.

The easiest and least expensive, and also the least accurate, measurement method uses a true-rms DMM. With this method, you set an amplifier to drive a motor at a constant speed under a constant load. Measure the rms voltage across the motor with the DMM and record the result. Then, break the motor-drive signal circuit and insert the DMM (set to measure current) in series with the motor. Measure the rms current and multiply by the voltage measurement to calculate power.

Unfortunately, this method lacks a crucial element: phase angle. As in all changing signals, the power in a PWM signal can be expressed as Power=V*I*cosè. Because the DMM method doesn't take phase into account, you can't use it to make absolute measurements. You can use the results only to compare motors to motors or amplifiers to amplifiers.

By using a power meter, you can come closer to measuring the actual power output from a drive circuit. Power meters simultaneously measure voltage and current; they also measure phase angle and use it to calculate power. A good power meter will display measurements with about 1% accuracy, assuming it can measure the power in most of the PWM signal's higher harmonics. If your meter lacks the bandwidth to measure a PWM signal's third and fifth harmonics, it will fail to measure some of the power that the motor uses. As a result, a power meter can still fall short of providing accurate measurements.

You can obtain better measurements with a harmonics analyzer. Harmonics analyzers—with accuracy as good as 0.05%—are often used for measuring AC mains harmonics emissions in EMC compliance tests. Harmonics analyzers typically resolve voltage and current to 14 bits or more, and sample rates can run up to 5 Msamples/s—high enough to measure the power in the higher harmonics.

Typically, the finer a harmonic analyzer's resolution, the slower its sample rate. An 18-bit harmonics analyzer, for example, might sample at 500 ksamples/s. That yields a usable bandwidth up to 250 kHz. For a 50-kHz PWM signal, this instrument will measure power up to the fifth harmonic only. At 20 kHz, you can measure power up to the 12th harmonic, which still isn't enough because you can get measurable power out to the 100th harmonic.

## Scopes work, almost

In contrast, most DSOs will have no trouble sampling fast enough to capture a PWM signal's harmonics. A 200-Msample/s, 100-MHz scope easily can capture a 50-kHz PWM signal out to the 100th harmonic. The DSO can multiply the two measurements to produce instantaneous power, then integrate the power measurements to get total power. Many DSOs can perform the calculations in the box and produce real-time results.

Although DSOs have no trouble with the PWM signal's harmonics, most scopes resolve signals to just 8 bits and typically have a vertical accuracy as high as 2%. Current probes can add another 2% of error.

Some engineers advocate using a scope with 12-bit resolution. That specification severely limits your choice of a scope. (See "Who makes what," above, for a listing of 12-bit oscilloscopes and scope cards with sufficient bandwidth for PWM power measurements.)

Some research applications may require finer resolution than you get from 12 bits, and they also need high sample rates and bandwidths. For those applications, some PC plug-in scope cards have the necessary specifications. Oscilloscope software usually can perform the power calculations for you, or you can write your own application.

PWM motors and controllers provide the most efficient means of converting electrical power into mechanical motion. Each of the various methods for measuring PWM power has limitations, especially when you need to measure small changes in power consumption. The motor industry would benefit from a standard measurement method, even one that accounts for those limitations. Without a standard, though, you can't resolve differences among manufacturers' measurement methods.

## For more information

Lee, Edward C., "Review of Variable Speed Drive Technology," Powertec Industrial Motors, Rock Hill, SC. www.powertecmotors.com/dlmanager.html.

"Making Measurements on PWM Drives," Power Analysis Reference Sheet 009, Voltech Instruments, Morrisville, NC, February 1998. www.voltech.com/Downloads/86165_01.pdf.

"Measurement of Adjustable Speed Drives with Fluke Meters," Fluke, Everett, WA, 1997. www.fluke.com/Application_Notes/ElectricalPower/Go416b_u.pdf.

"Modulation Analysis—PWM: Use Jitter and Timing Functions to Analyze PWM Signals," LeCroy, Chestnut Ridge, NY. www.lecroy.com/tm/library/labs/lab421/default.asp.

"Power Measurements on AC Motors and PWM Drives," Application note 105, Voltech Instruments, Morrisville, NC. www.voltech.com.

The Small Motor & Motion Association (SMMA) offers courses in motor design, feedback devices, and other topics. SMMA, Sherborn MA. 508-376-5360; www.smma.org.

The University of Florida's Power Quality Lab lists publications and conference papers on harmonics measurements and power quality. www.powerqualitylab.ece.ufl.edu/gen_info/publicat.html .

The following company information appeared in the original print version of this article. For up-to-date information about companies, visit the Instruments portion of our Buyer's Guide.

 Who makes what? 12-bit digital oscilloscopes and oscilloscope cards Gage Applied Lacine, QC, Canada 514-633-7007www.gage-applied.com Hi-Techniques Madison, WI 608-221-7500www.hi-techniques.com Measurement Computing Middleboro, MA 508-946-5100www.measurementcomputing.com National Instruments 512-794-0100 Austin, TXwww.ni.com Nicolet Technologies Madison, WI 608-276-5600www.niti.com

 Power meters See "Power Monitors/Analyzers" under "Stand-Alone Test Instruments" in the Test & Measurement World Buyer's Guide, July 2001 (p. 136).

 Power harmonics analyzers See "Harmonics & Flicker Testers" under "EMC Test" in the Test & Measurement World Buyer's Guide, July 2001 (p. 156).

 References "California Passes Energy Bill; Incentives For Motors and Lighting," electroindustry, National Electrical Manufacturer's Association (NEMA), Rosslyn, VA, May 15, 2001. p. 1. www.nema.org.

Loading comments...

Write a Comment

To comment please Log In

FEATURED RESOURCES