For the PDF version of this article, click here.
In 1776, Scottish economist Adam Smith introduced the concept of division of labor and specialization, concluding that it is in the economic interest of both micro and macro organizations to complement internal capabilities, where practical, with those of outside sources. This basic economic principle still rings true, especially for today's electronics equipment designers and OEMs who are pressured by global market forces to supply increasingly cost-effective, power-minimizing “green” products and power systems with faster development cycles.
Designing and deploying the new breed of green power devices requires an increased level of sophistication that generally exceeds the accuracy, precision (repeatability) and resolution capability of many in-house labs in terms of testing and measuring harmonics, power factor, ultralow wattage and other design parameters. The good news is that designers no longer need to rely on home-brew setups, given the widespread availability of commercial test instrumentation with a dazzling array of functions that make these devices as sophisticated as the application demands.
Green Power Management
Green can mean many things when applied to electrical equipment, including any product containing no lead or other RoHS substance.
Green also refers to the operational efficiency of a device in either the on or standby modes. For the latter state, this includes many copiers, fax machines, monitors, computers and consumer devices. This circuitry typically requires an off-line power supply, which also may be used for battery charging, distributing system power or other functions.
Measuring the power of a product when in the off-line (standby) mode can be more complex than might be anticipated. The wide dynamic range needed to measure the equipment poses the initial challenge. Inrush or operational current may consume hundreds of watts, and yet standby power can be well under 1 W. Unfortunately, not all ac current meters provide sufficient accuracy, precision or resolution to extract such low-level signals from the noise floor for these measurements. Therefore, using conventional test equipment in an attempt to avoid the expense of using the proper specialized instrumentation often results in errors.
Measurements for Green Compliance
A broad selection of commercial instruments is now available to meet the measurement criteria of various countries' energy programs, including Energy Star, Blue Angel, 1-Watt Initiative and Top Runner. But what exactly is needed to achieve accurate and precise measurements with enough resolution to meet the requirements for measurements below 1 W? To begin with, it is highly recommended to obtain evaluation units from potential suppliers to determine which instrument offers the optimum combination of price, user friendliness, accuracy, precision, resolution, suitability of performance and functionality for the desired application.
Commercial power analyzers and digital power scopes use the following method to calculate power:
where v(t), the instantaneous voltage, is multiplied by I(t), the instantaneous current, and then integrated over a period of time.
Some of the newer oscilloscopes incorporating built-in PCs also offer ac power-measurement capability and standard or optional power-analysis software. This configuration allows the mathematical and analytical functions of the PC/oscilloscope combination to analyze and average the voltage and current data. In this way, one channel of the scope functions as a current input using a differential probe placed across a current shunt resistor, with the other channel acting as a voltage measurement input. The input bandwidth and sampling times of the oscilloscope will be fast enough to capture the complex waveform. However, an oscilloscope, even the fancy PC-based variety, will not offer the accuracy, precision and resolution of a dedicated analyzer. Nor will it offer the same cost-effectiveness.
Another thing to be mindful of is oscilloscope and chassis grounding when probing the power line. It will be necessary to use an isolation transformer for the unit undergoing testing and possibly the oscilloscope itself to avoid damage to either. The drawbacks of this technique are setup time, programming and the cost of the scope. Obviously, this scenario takes more time than simply hooking up an instrument and pushing a few buttons for a fast, accurate result. However, if the budget allows, it makes good sense to add a power-measuring capability to a high-performance oscilloscope.
Defining Basic Terms
At this point, the terms accuracy, precision and resolution need to be revisited. It is essential to clearly understand how these terms differ in order to correctly evaluate and interpret test results. Low-level ac power measurements not only require accuracy to known standards, but also must be precise and provide sufficient resolution below 1-W levels.
Though basic, these terms often are confused or used interchangeably. Accuracy quantifies the uncertainty in the measurements made with an instrument, which is defined as the agreement between a measured quantity and the true value of that quantity. Precision quantifies the expected scatter or spread of a group of repeated measurements of a fixed quantity. Resolution refers to the smallest change or increment in the measured quantity that the measuring instrument can detect with certainty.
These concepts are better understood when applied to the simple exercise of hitting a target. Good precision is said to exist if the same spot is hit with repeated attempts. For example, if repeated measurement of a 1-V standard consistently produces a 0.9-V result, then a high degree of precision is present, but not good accuracy. Using the same example, repeatedly reading a true value of 1 V from a 1-V standard source results in a high degree of both accuracy and precision.
Resolution is concerned with the smallest perceptible change that can be measured. Extending the analogy of targets, the target rings provide a means for judging the distance between the holes in a target, and it would be difficult to judge the distance between two closely spaced holes in the same ring. Likewise, using the millivolt range of a 3.5-digit digital multimeter (DMM) to read a dc voltage to microvolt resolution may be inadequate for the task.
Today's modern energy-conserving power systems use burst-mode pulse- or cycle-skipping techniques to lower the average power drawn from the line. This nonclock-based signal produces an aperiodic, complex and essentially random current waveform that varies widely over time, line voltage and load. Obtaining high-integrity results in this scenario demands fast sampling rates and long-term, time-averaged measurements. This, in turn, requires the right type of test equipment.
Survey of Measurement Techniques
Let's look at a few pitfalls of the various techniques in which low-level ac measurements may be made.
One might think the simplest solution would be a true RMS meter used to measure ac line voltage and current. However, this yields less-than-desirable results because of time-varying loads and the typically low sampling rate of the DMM analog/digital converter. Depending on the instrument, most slow, conventional integrating-type meters simply chase the rapidly changing signal and randomly lock onto some portion of the waveform, usually producing incorrect results.
The best-case scenario is that the measurement will vary widely and will not be stable enough to calculate load power in standby mode, in which case the user will not even attempt a measurement. Worst case, the engineer may perceive the readings are sufficiently stable to be used in calculations, unknowingly producing incorrect results.
Another method, shown in Fig. 1, employs a 0-V to 600-V dc power supply set for the rectified and filtered dc voltage level on the output of a bridge rectifier supplied with 120 Vac or 220 Vac. An analog dc current meter with a bypass switch is inserted in the circuit and, after the initial charge surge, the bypass is switched out and the dc current read. The fluctuation of the input current signal causes a corresponding oscillation of the dc current, which the user then eyeballs to average the current level and log the range. However, such visual measurement methods leave much to be desired in terms of precision and accuracy.
One necessary precaution to observe before taking any measurements is to understand the internal circuitry being measured. Simply knowing that main supply lines feed a specific power supply without knowing what else may be connected to them can cause problems. For example, in the case of a transformer powering a low-voltage heater in a fax machine or copier (Fig. 2), applying a dc voltage will not react well to the low dc resistance (Rdc) of the transformer's primary, causing the fuse to blow.
The needed measurement is the ac standby power at the line frequency and voltage in question, and at an equivalent dc voltage. The test product will almost certainly include an EMI filter, such as that found in Fig. 3, and yet with dc voltage applied, it will not be possible to measure the effect on standby power of the capacitive reactance of the EMI filter or any other connected reactive components.
Furthermore, since the measurement is not being made the way the device is being used on a 50-Hz to 60-Hz ac line, potential damage may occur.
The best measurement technique involves using an ac source combined with a fast sampling and high-bandwidth capability instrument to sample current and voltage at 200 kHz. The signals must then be averaged over time, with the necessary math being performed on the current and voltage signals to achieve precision on the complex wave shape. The instrument should have enough dynamic range for both standby and running-mode measurements of efficiency and performance.
Using the metering function of a programmable ac source to make these measurements may indeed work. However, there is an attendant risk in trusting the accuracy, precision and resolution of measurements below 1 W when made on a 1500-W ac source. Meaningful data probably will not be acquired in this manner, resulting in less than accurate information being reported to a regulatory agency or printed in product specifications. The goal then is to meet or exceed global regulatory requirements.
In summary, the basic test requirement for low-level ac power measurement involves sophisticated techniques capable of fast continuous averaging on complex or distorted wave shapes. The good news is that engineers no longer have to rely on “home brew” setups for measuring voltage, current, power factor, harmonics, low-level wattage and other critical parameters because of the widespread availability of commercial instruments to meet every need.
Since all test equipment is not created equal, it is extremely important to be highly selective in purchasing the right instrument for a given application. Critical performance characteristics to look for include the accuracy, precision and resolution specifications of the instrument, in addition to these key features:
Fast simultaneous sampling of the current and voltage waveforms
Ability to handle complex random-current waveforms
Wide dynamic power-measurement range
Ability to average measurements over long time periods
Ability to use instruments either from the front panel or export the data to a PC for post-processing.
The motto “try before you buy” is appropriate here. To make a truly informed decision, the engineer must evaluate the instrument under actual working conditions. Anticipated future needs, and the instrument features and functions to meet them, should also impact the purchase decision.
It is important to remember that the same instrument will be used for any number of test activities as the product progresses from circuit design through product-level evaluation. This includes a wide range of operating-mode efficiency and standby power measurements, as well as evaluating the power factor of the design, measuring harmonic content and a host of other functions. Of course, it is difficult to project future needs with absolute certainty; however, it behooves the engineer to build as much capability as possible into the lab at the outset as a hedge against unexpected demands in the future.
Low-level ac power measurement requires highly sophisticated, commercial-grade test instruments, especially in Europe where NIST or other traceable calibration and CE certifications are important. Doing anything less will only create problems in the future as mandatory compliance with various global power-efficiency standards moves inexorably forward.
What a shame to discover that a product could not ship because someone “spared no expense to save money” in the test lab, with the end result being that product specs were off by too wide a margin. Rather than waste valuable resources reinventing a measurement instrument, engineers' talents are better spent innovating circuits and products. It is easy to far exceed a couple thousand dollars in time and effort flying blind without the right test gear. To prevent the disaster of some external testing agency or customer revealing that a product or circuit performed poorly — information that should have been known before design startup — the right test equipment needs to be installed in the lab to verify that the engineer can in fact design and ship with confidence. Just like craftsmen of any trade, design engineers are only as good as their tools.
Electronic Product Design, www.epd.com.
Voltech test and measurement, http://www.voltech.com.
Yokogawa test and measurement digital power meters and power analyzers, http://www.yokogawa.com/tm/wtpz/tm-wtpz.htm.
Hioki USA power 3332 hitester, http://www.hiokiusa.com/modules/products/index.php?op=viewproduct&proid=51.
Voltech notes, publication number VPN 104-054/1, “Standby and Low Power Measurements.”
Top Runner program, Japan's approach to energy savings and conservation, http://www.eccj.or.jp/top_runner/index.html.
Ecos consulting, http://www.ecosconsulting.com/.
EPA energystar program, http://www.energystar.gov/.
Clarke-Hess model2330, http://www.clarke-hess.com/2330.html.