February 2014
AMETEK Programmable Power
DIFFERENTIATED POWER SUPPLIES
Powering Audio Amplifiers
Shawn Smith Vice President & Business Unit Manager of AMETEK Programmable Power
All the forces in the world are not as powerful as an idea whose time has come. —Victor Hugo, 1800
Power Developer contains new ideas that come every month. —Power Developer Editors, 2013
POWERDEVELOPER
Read Power Developer, the monthly newsletter for Engineers:
h t t p : / / w w w. e m b e d d e d d e v e l o p e r. c o m / n e w s _ l e t t e r /
Power Developer
4 10
CONTENTS
TECH COLUMN
eGaN FETs for High Performance Class-D Audio Amplifiers
TECH ARTICLE
Power Factor & Power Factor Correction
18
COVER INTERVIEW
24
FEATURED ARTICLE
32
Shawn Smith - Vice President of AMETEK Programmable Power
Compensation Methods in Voltage Regulators
TECH COLUMN
Tackling Problems with Innovative Concepts: Extra Flexibility thanks to two-stage driver solutions
Perfection in Power
Tackling Problems with Innovative Conce LEDs are now widely used in many different areas. They allow for lighting solutions that were simply not possible with conventional lighting equipment. This means, however, that LED drivers need to meet ever more complex, application-specific demands. In many cases, two-stage drivers are the best solution. By separating the AC power supply from the DC LED power, they offer new possibilities that are simply not achievable with conventional LED drivers.
Thomas Rechlin Senior FAE for Europe at Gmunden, Austria
3
Power Developer
4
TECH COLUMN
Alex Lidow
CEO of Efficient Power Conversion (EPC)
5
Power Developer The quality of sound reproduced by an audio amplifier, measured by critical performance parameters such as THD (Total Harmonic Distortion), damping factor (DF), and T-IMD (Inter-modulation Distortion), is influenced by the characteristics of the switching transistors used. Class-D audio amplifiers typically use power MOSFETs, however, lower conduction losses, faster switching speed, and zero reverse recovery losses provided by enhancementmode GaN (eGaN) FETs enable a significant increase in the sonic quality, and higher efficiency that can eliminate heatsinks. The result is a system with better sound quality in a smaller form factor that can be built at a lower cost.
To read the previous installment, click the image below:
Musical sound usually contains multiple frequencies, all of which must be amplified equally to truly reproduce sound without inducing distortion and impacting the sound quality. Factors such as cross over distortion, slew rate, and current gain of transistors can make amplifiers non-linear. The signal delays with negative feedback systems make the amplifier incapable of correcting distortion when subjected to fast transient signals commonly found in audio applications [1, 3]. These items cause harmonic distortion and intermodulation distortion in the reproduced signals. As we will show, the improved performance of eGaN FETs compared with MOSFETs significantly improve the overall sonic performance of Class-D amplifiers.
OPEN LOOP DISTORTION & EFFICIENCY PWM generators do an excellent job of synthesizing the low level analog waveform in digital format. Of course a higher PWM frequency will generate a higher quality digital representation of the analog signal. The primary source of open loop distortion in a class-D audio power amplifier is the deviation of the amplified digital signal from the low-
6
TECH COLUMN level digital signal presented to the power stage of the amplifier. This distortion is due to excessive propagation delays and dead-times based on (a) gate-to-source charge required to charge gate to the threshold voltage (QGTH), (b) slow slew rates based on turn on charge (QGS2 + QGD), common source inductance (CSI), (c) resistive drops based on RDS(on), (d) reverse recovery delays and ringing based on diode reverse recovery charge (QRR), and (e) ringing of the open loop power stage.
Figure 1: Comparison of MOSFET and eGaN FET switching waveforms
Power stages using eGaN FETs come much closer to synthesizing the ideal power digital signal compared with power stages using MOSFETs because they are far superior to silicon MOSFETs in every characteristic that contributes to the distortion of the class-D power stage. The same factors that contribute to deviation from the ideal waveform are responsible for power losses in the amplifier, so the eGaN FET-based class-D audio amplifier will naturally be a much more efficient amplifier.
TOTAL HARMONIC DISTORTION (THD) Total Harmonic Distortion is a primary measurement of the quality of an audio amplifier. THD is measured at a constant audio frequency (typically 1 kHz) over a range of power levels. With power MOSFETs in a class-D amplifier, a fairly long period of dead-time is required to prevent “shoot-through” between the high-side and low-side FETs, as well as to account for the body diode reverse recovery time. These reverse recovery oscillations add additional noise and distortion. Since eGaN FETs switch many times faster than power MOSFETs and do not have any reverse recovery charge, the dead-time – typically 25 ns for silicon power MOSFETs – can be reduced by 80%, to 5 ns or less. THD is also influenced by the switching and delay times of the FET. eGaN FETs have much lower gate charge, QGS, compared to a silicon MOSFET of similar RDS(on). This means that the turn-on and turn-off delay, as well as the rise and fall times for eGaN FETs, will be much faster and contribute less to signal distortion. Figure 1 shows a typical switching waveform of a power MOSFET and eGaN FET illustrating the faster switching speed and absence of diode recovery.
Figure 2: RDS(on) and QGS1 comparison : eGaN FETs vs. Si FETs (100 V)
Figure 3: RDS(on) and (QGS2 + QGD) comparison : eGaN FETs vs. Si FETs (100 V)
7
Power Developer
Figure 4: THD+N of 0.003% at 8 立, 1 kHz, with the eGaN FET class-D amplifier
Figure 5: eGaN FET Class-D amplifier efficiency - 4 立 and 8 立 ohms speakers
8
TECH ARTICLES TRANSIENT INTERMODULATION DISTORTION (T-IMD) A second type of distortion is Transient Intermodulation Distortion (T-IMD), which is the result of introducing frequencies not present in the original audio [1]. With power MOSFETs, higher open-loop impedance from the higher RDS(on), QGD, and QGS of the switching device, as well as the longer dead-time, requires excessive feedback to improve audio performance. Excessive feedback limits bandwidth, and introduces intermodulation distortion [3]. The low onresistance and low capacitances as well as reduced dead time of eGaN FETs enable simpler and lighter feedback providing substantial reduction in T-IMD as well as higher efficiency in Class-D systems.
A SIMPLE FIGURE OF MERIT TO COMPARE TECHNOLOGIES Given their impact on performance and efficiency described above, perhaps the best way to gauge the relative performance of the an eGaN FET and a MOSFET in a Class-D audio system is by calculating two figures of merit (FOM) based on, (a) the product of the on-resistance, RDS(on), and the gate-to-drain turn -on charge, (QGS2 + QGD), and (b) RDS(on) and the gate-to-source charge (QGS1) [2]. These comparisons are illustrated in Figures 2 and 3, where it can be seen that eGaN FETs are significantly better than state-of-the-art MOSFETs in both FOMs.
CLASS-D AMPLIFIER EXAMPLE To demonstrate the advantages of eGaN FETs in class-D applications, a Bridgetied Load (BTL) output class-D amplifier demonstration system was built and tested [4]. Each channel was designed to deliver 150 W into 8 ohms, or 250 W into 4 ohms load, with less than 0.1% THD, when powered from a ±27 V nominal power supply. Faster switching speeds, shorter dead-times, and the absence of the diode recovery enable very low THD+N as shown in figure 4, while minimizing T-IMD and the EMI emissions from the amplifier.
In addition, the losses in the eGaN FETs were so small that heatsinks could be completely eliminated, while delivering 150 W into an 8-ohm speaker (or 250 W into a 4-ohm speaker). Figure 5 shows the efficiency of the EPC2016 [5] eGaN FETbased class-D power converter.
SUMMARY Lower conduction losses, faster switching speed and zero reverse recovery losses provided by high switching speed enhancement-mode GaN (eGaN) FETs enables a step forward in designing class-D audio amplifiers with superior sonic performance and lower cost.
REFERENCES [1] Non-linear distortion and intermodulation distortion, white paper, Klippel, GmbH http:// www.klippel.de/measurements/nonlineardistortion/intermodulation-distortion.html [2] J. Strydom: “Comparing Figure of Merit,” http://powerelectronics.com/discrete-powersemis/egantm-silicon-power-shoot-out-part1-comparing-figure-merit-fom [3] L. Butler, “Intermodulation performance and measurement of intermodulation components,” VKSBR, Australia. http://users.tpg.com.au/users/ ldbutler/Intermodulation.htm [4] Efficient Power Conversion, “EPC9106 – Demo Circuit Print,” data sheet [5] Efficient Power Conversion, “EPC2016 – Enhancement-mode Power Transistor”, data sheet eGaN® FET is a registered trademark of Efficient Power Conversion Corporation. ■
» CLICK HERE to comment on the article.
9
Power Developer
10
TECH ARTICLES
Long gone are the days when only engineers that worked with large electric motors and high power electric loads need worry about power factor. The introduction of switching power supplies into electronic systems has led to increasing international legislation, moving power factor up the list as a key concern for most engineers developing systems that run on ac mains power.
11
Power Developer
WHAT IS POWER FACTOR? Defining Power Factor Power factor (pf) is the ratio of real power (P) flowing to the load, to the apparent power in the circuit (S). It is a sinusoidal waveform and therefore expressed as dimensionless number between -1 and 1: pf = P/S For a purely resistive load, the two figures are identical; for a reactive load the arithmetic for the apparent power produces the same figure, that is, the product of the RMS values of voltage and current. However, to find the actual (real) power delivered to the load, the instantaneous product of voltage and current must be integrated over the complete sine-wave cycle.
Adding Reactive Power into the Equation
power, measured in volt amperes (VA), is the product of the current and voltage of the circuit. When current is leading or lagging voltage, the value of that integral will always be less than the value for the in-phase case over the same interval. This reflects the attribute of an inductor or a capacitor to act as energy stores; at various points through the ac cycle the reactive component (Q) is either storing energy, or returning it to the system. Perfectly sinusoidal waveforms follow Pythagoras, where the square of the apparent power is equal to the squares of the active power and reactive power measured in reactive volt amperes (var): S2 = P2 + Q2 or P = S cos θ The relationship is conventionally visualized in a right-angled triangle vector diagram;
Real (or active) power, measured in watts (W), is defined as the circuit’s capacity to perform work at a given time. Apparent
Figure 2: Right angled triangle vector diagram
Figure 1: Instantaneous and average power calculated from ac voltage and current with a zero power factor.
12
Apparent power is always greater than or equal to real power and a negative power factor can occur when the device starts to generate power that then flows back to the generator.
TECH ARTICLES Ideal vs. Real World Ideal power factor occurs when the current and voltage waveforms are in phase: pf = 1 (i.e. cos 0) When the power factor is not equal to 1.0, power losses, and potentially harmonics that disrupt other devices, occur. Once again, this basic definition is for pure sinusoids. Non-sinusoidal waveforms are more complex, but as they can be represented by a series of harmonic sinusoids, the same basic principles apply. Problems occur, however, when a nonlinear load on a power system, typically in the form of a power supply, create current harmonics not present in the voltage; in effect, these current harmonics become part of the reactive power. They do not represent true power delivered to the load, thus degrading the power factor.
THE NEED FOR POWER FACTOR CORRECTION Importance of Power Factor and Power Factor Correction Power factor is on the list of design concerns for designers of virtually every device that draws significant power from a mains socket, as well as for engineers in heavy-electrical sectors. Not only are there principles of good engineering practice at stake, there is also legislation to enforce conformance with power-factor norms.
Figure 3: Instantaneous and average power calculated from ac voltage and current with a lagging power factor of 0.71 Across the Atlantic, the EU legislates the EN 61000-3-2 standard for electronic equipment, setting limits to the 39th harmonic for equipment with input currents less than or equal to 16 A per phase. The standard is split into four classes: A, B, C for appliances, power tools, and lighting respectively, and the most stringent class, D, for PCs, computer monitors, and TVs rated between 75 and 600 W. Following Europe’s lead, a host of countries are developing increasingly stringent legislation. Most notably China, Japan and Australia have government regulations for power-line harmonics that have evolved from EN 61000-3-2.
Legislation
Supply Implications
The first attempt to legislate for ac mains power interference came over 100 years ago (1899), to prevent incandescent lamps from flickering, but one of the key regulations came in 1978, with IEC 555-2 requiring power factor correction be incorporated into consumer products.
Why is this cause for concern? Electric utilities and generating bodies require their customers to present a load to the power grid that is as nearly unity power-factor as possible. The main, but not the only reason, is fiscal. The customer expects to pay for the “real” work done on his premises – in other words, the value of W, above. But a PF of less than 1.0 means that the VA value must be higher than W.
Today, the U.S. Department of Energy’s voluntary Energy Star guidelines call for computing equipment to have a power factor of ≥ 0.9 at 100 percent of rated output in the application’s power supply, meaning designs with internal power supplies require the use of active power factor correction.
The generators and transmission companies must provision to deliver the peak voltage and current values in the waveform at any time. A power factor of less than 1.0 is effectively an
13
Power Developer
increase in their costs, and one that they pass back to customers by imposing an increased tariff for customers with low power-factor loads. Achieving maximum power factor is therefore a “win-win� for all concerned.
Engineering Problems of Low PF There are further effects that power utilities must contend with, that makes a unity-powerfactor load highly preferable. Rotating plantgenerated power is more difficult to manage and to keep stable when supplying a low power factor, and there can be heating or overload hazards for transformers and transmission equipment in the supply grid. Grid stability is also more difficult to maintain with low-power-factor loads attached to the system. Low power factor also tends to be associated with other negative attributes for a well-behaved electrical load. Highlydistorted current waveforms drawn from the mains can inject high-order harmonics back into the supply grid.
14
Transmission equipment has higher losses at higher frequencies leading to heating problems, and, if the higher frequencies are present in the load placed directly on the generating plant, they can manifest themselves as destructive vibrations leading to excessive wear on components such as bearings. Current distortion can lead to out-ofbalance currents in the neutral lines of 3-phase distribution networks, which in turn can take the neutral away from ground (voltage) and give rise to multiplicity of problems.
Poor Current Waveforms and Harmonic Effects Waveform distortion is not, strictly, the same phenomenon as power factor but because of the close connection between the two (as noted above, current harmonics degrade PF), they are governed by the same standards that place limits on the load that may be drawn by mains-supplied equipment; for example EN 60555 in Europe, and IEC 555-2 internationally.
TECH ARTICLES Power Supplies for Electronic Systems Even when most electronic equipment was supplied by power supplies that used linear regulation, power factor (and waveform distortion) was often less than ideal, but was rarely addressed for anything other than the largest supplies. The typical, conventional offline arrangement was that of a transformer followed by a bridge rectifier, feeding a reservoir capacitor. Conduction through the rectifier would take place when the dc voltage on the output line had sagged below the instantaneous value of the transformed ac supply, which could be for the complete cycle at full load, or only at the peak of the ac waveform under light load.
Switching Power Supplies Switching power supplies can significantly worsen the situation. The off-line part of the design may not change, still comprising a transformer/rectifier, and capacitor, but now one or more switching regulators have been added to the equation. The input rectifier continues to generate poorly-shaped current waveforms with the added challenge that some of the higher-frequency switching noise from the regulation stage can find its way back into current drawn from the wall socket.
Phase and Harmonic Effects Not only does this shift the effective current peak away from that of the voltage waveform in time, it also introduces high-harmonic-content switching waveforms that potentially worsen the distortion of the current waveform. The arrival of this class of supply broadly coincided with the widespread deployment of PCs and other IT products in great numbers. Such trends led directly to today’s legislative environment.
of an uncorrected switching power supply. This corrected waveform minimizes losses as well as interference with other devices being powered from the same source. Compensation for low power factor can be by passive or active devices. The simplest case is that of improving the power factor presented by electric motors. Naturally, being wound machines, they are highly inductive loads, and adding capacitors to the supply network has long been standard practice. Even this case may not be entirely simple; the designer of such a network has to take care not to create unwanted resonant effects, for example. Variable power factor in the load may be accommodated by an adaptive scheme to connect reactive elements as required and in high-power contexts (MW scale) rotatingmachinery solutions can be applied. Power supply designers must–subject to their designs falling into specific power rating bands– take into account regulatory constraints, but they are invariably also under pressure to meet space, component bill-of-materials and efficiency targets.
Passive Power Factor Correction Passive power factor correction in the form of filtering can be effective, within limits, and has the effect of reducing the higher-order current harmonics that, as noted above, contribute to degraded power factor. Such techniques involve putting a low-pass filter in the input side of the power supply to suppress higher- order harmonic components, and then compensating lead/lag characteristics as with conventional power factor.
METHODS FOR POWER FACTOR CORRECTION Passive vs. Active Power Factor Correction The solution to excess harmonics is to use power factor correction (PFC), which shapes the input current of the power supply to maximize the real power level from the mains and minimize harmonic distortion. Ideally, the electrical appliance should present a load that resembles a linear load, such as a simple resistor, rather than the reactive load
Figure 4: Effect of power factor correction
15
Power Developer supply’s main converter. However, the added pre-regulator introduces a potential source of EMI (electromagnetic interference) noise that must be taken into account in the final design.
Design Constraints and Measurement The design is therefore constrained by a power factor target, which will come from legislation and is non- negotiable, plus all of the usual factors of efficiency, component cost, and volume/board space occupied. To design to those targets, especially power factor, demands that it can be measured.
Figure 5: Circuit diagrams of passive/active filtering The downside to this scheme is that large (both by value, and physically) inductors and/or capacitors may be required. Additionally, there are limitations to the input range and power rating when implementing this scheme; passive PFC circuits are generally able to achieve a power factor in the range of 0.70-0.75.
Active Power Factor Correction The existence of, and rapid progress in, highspeed, high-current capacity semiconductor switches – the same components that enable the high-efficiency switch-mode supply to be built at all – make available the option of active power factor correction which allows for a power factor close to 1.0. Because of this, it is the strategy that is most widely applied in present-day designs. A switching pre- regulator stage is placed in the input current path of the supply. That regulator is designed not only to maintain a constant DC voltage to feed the main converter stage of the power supply, but also to draw current from the input in-phase with the incoming AC voltage waveform. An additional switching stage in the supply does impose some extra losses, and some extra cost. As indicated above, there are compensating savings in the form of smaller passive filtering components, and in the
16
Simple multimeters will usually misread to some extent when presented with distorted waveforms; a good digital multimeter may make a good effort at reporting the RMS value of even a moderately distorted waveform, but will be challenged to record power factor. An oscilloscope with current and waveform probes can be used to gauge the phaseshift (of the angle θ, as above) but only if the voltage and current wave-shapes are reasonably similar. Electrical engineering has traditionally used precision electromechanical (moving coil) meters that directly measure power factor, but these are hardly suitable to track down the effects of kHz or MHz regulator switching in an advanced PSU. A small number of test & measurement suppliers offer power analyzers; by rapidly sampling current and voltage waveforms and performing a suite of calculations – and applying a Fourier transform to extract harmonic information – every detail of a load’s performance in terms of distortion and power factor are revealed.
Summary Power factor is on the list of concerns for designers of virtually every device that draws significant power from a mains socket, as well as for engineers in heavy- electrical sectors. The power factor target, based on legislation, plus efficiency, component cost, and volume/ board space need to be considered. For this reason CUI has designed active power factor correction into the vast majority of its ac-dc power supplies rated at 100W and above to help ease implementation and ensure compliance for OEMs. ■
www.cui.com
TECH ARTICLES
PURE DIGITAL POWER ZL8800 Combines Digital Control Loop Technology and PMBus™ Interface for the Industry’s Most Advanced Digital Power Solution Not only can you now track and control every aspect of your power supply without additional monitors or sequencers, the ZL8800’s ChargeMode™ control loop technology delivers best-in-class transient response and eliminates the need for complex compensation. This all adds up to a pure digital solution that lowers system cost, reduces design time and increases available board space.
►INTERSIL.COM/ZL8800 Download Datasheet Read App Note Order Demo Board Watch Videos
Stop Coding and Start Designing with Intersil’s PowerNavigatorTM Download free at intersil.com/powernavigator
New Digital Power Knowledge Center Watch and learn at intersil.com/DigiPowerTraining
©2013 Intersil Americas LLC. All rights reserved.
17
Power Developer
DIFFERENTIATED POWER SUPPLIES Interview with Shawn Smith Vice President/Business Unit Manager of AMETEK Programmable Power Shawn Smith has been in power electronics his entire career. After working for a variety of power conversion and superconductor companies throughout the U.S., Smith joined AMETEK to lead the Programmable Power Business Unit in San Diego, California. Smith has dedicated himself to ensuring AMETEK is not like other power companies out there—the company stresses highly differentiated innovation, industry leading quality, and superior customer service. We spoke with Smith about some of the key areas of AMETEK Programmable Power’s differentiation, a few areas of key innovation, and what the future holds for the programmable power industry.
18
COVER INTERVIEW
19
Power Developer
“ AMETEK excels at developing products that are highly differentiated in the market with value added features on the products. We have very deep expertise in power technology from working with customers for more than 70 years.” Can you give the readers an overview of the products you offer for the Programmable Power division of AMETEK? AMETEK’s Programmable Power division covers a broad range of products and applications. The business is built on a variety of very well-known and strong brands, including Sorensen DC Power Supplies, California Instruments and Elgar AC Power Sources and AMREL Electronic Loads. We also do some custom solutions for high performance power applications, such as Space Environment Simulation Power Supplies.
As far as differentiating yourself as a power company, what would you say is the strategic success among all your power products? One thing that sets us apart from other power supply manufacturers is that AMETEK excels at developing products that are highly differentiated in the market with value added
20
COVER INTERVIEW
features on the products. We have very deep expertise in power technology from working with customers for more than 70 years, and we apply that expertise to our new product concepts. Second is a brand name. We have a lot of well-established brand names with Sorensen, Elgar, AMREL and California Instruments that are well respected in the industry and represent a broader product portfolio than other companies that may only offer DC or AC solutions. Third, we have toptier manufacturers’ reps and distributors that are top shelf channel partners supporting our customers everywhere in the world.
What are some of AMETEK’s biggest innovations? The biggest one is the broad range of products that we offer. Most of our customers would buy a variety of different products for different applications—and we are one supplier who can meet all those needs. As just one example of how that’s important for us, we’re the only power supply company able to support the standard Department of Defense ATS system across all four military branches. Another innovation is that we are deploying more digital controls for our power supplies that give our customers greater flexibility and freedom in what they can do with their power supplies. Lastly, with every generation of product innovation, we really push the power density higher and higher.
California Instruments BPS Power Supply
What kind of feedback are you getting from your customers and how is that helping you drive things forward? There are actually a couple of things. One, customers really value service and can easily find us on a global basis. The better service you provide, the more important you become to those customers. It is not just about the products, but also the level of service you provide as well. In term of features, we offer a broad range of capabilities. Engineers can now do more with a single box that allows them to be more flexible and adaptive to their environments, rather than having to buy multiple different devices that are only designed to perform a specific application or function. Our products offer more versatility today.
Elgar ETS TerraSAS Series Photovoltaic Simulator
21
Power Developer
How global of an audience does AMETEK have? AMETEK is a global company. We are based in San Diego, but we really have business everywhere. We have business in Asia Pacific, including China and South Korea. Western Europe is obviously big for us. Latin America is growing, and we expect it to be significant in the coming years. We see the same growth prospects for Eastern Europe— Russia, in particular.
What are the biggest challenges in trying to reach these different target markets? AMETEK as a company does a really good job helping all our customers with supply chain issues by making sure we have sales and service capabilities with global reach. Entering some markets is challenging, especially if you have local competitors in those regions that can sell for a lot lower price because of fewer features, perhaps lower quality, or the advantages of being produced within that country. We compete with this not only with our products but by ensuring that we’ve extended our reach into these developing markets for us with top tier partners in the region. We’re selective about who we work with, and just finding the best partners is a big part of the challenge.
There is such a huge range in power devices. How do you decide which products to offer? It does take a lot of work. This business was put together over the years by several acquisitions by AMETEK, so we have a mix of older products that have been part of our customers’ test protocols for years, as well as innovative newer products. We are always re-examining the markets to assess where the biggest opportunities are for AMETEK. At the same time, we also carefully consider our existing customer base to make sure that we continue to understand what they’ll need in the future. In many cases, it is not about
22
COVER INTERVIEW
chasing only the biggest segments for a particular category of products, but about how much of our installed base relies on us to maintain a product. We really try to balance both between caring for our existing customers and trying to expand to additional markets that we do not currently serve.
In terms of the development process at AMETEK, why should engineers have confidence in the reliability and performance of your products? AMETEK is a large, global company with more than $3.6-billion in revenue. It is a very well-respected and well-run company with very good processes in place to deliver innovation and maintain control of quality. Here in San Diego, for instance, our AMETEK Programmable Power operates out of an ISO9001 facility with tight controls from beginning to end. We have a lot of industry experience with long-term industry veterans from different companies. We focus on preserving that industry insight and institutional knowledge that people gained over decades, and it has been a big part of our success.
What is the culture like at AMETEK? It is a very proud culture. We are proud of the products and the brands that we have. A lot of people who work here have come from the companies we’ve acquired to create AMETEK Programmable Power, so those people have worked hard their entire careers to build value in the quality and names we carry, AMREL, Sorensen, Elgar and California Instruments. Our culture can also be considered customer-centered. Because of our long history, our people here have been working with the same customers for quite some time now, and some of those relationships even span decades. We are proud of the quality of those relationships. Lastly, AMETEK is a very driven culture. AMETEK wants to get better, wants to get bigger responsibly, wants to serve customers well, and make a good and fair profit by offering highly differentiated products.
“We’re selective about who we work with, and just finding the best partners is a big part of the challenge.” What is the most exciting thing in the future of programmable power? There is actually a great deal. One of them is the growth potential of new parts of the world; I did mention Russia earlier as a target market, and there are many emerging markets that provide numerous opportunities. Global expansion is a big AMETEK strategy. Secondly, everything now has become more dependent on electronics, and that also fuels our growth. A good example is the Boeing 787 that was the first fully electric-controlled plane in the skies. All those systems have to be tested and set to very high standards. Our customers’ need for quality really drives the growth for our industry and that emphasis on quality standards among our customers is not slowing down. There is no end to how much electronics will continue to be integrated into our daily lives. All of them have to be tested, and the only way to efficiently test is with a power supply.
Do you have anything more to share with our readers? It is an exciting industry. Customers need and want reliable products from solid companies that will be around to support them for years and decades to come. AMETEK Programmable Power is here to do that and to develop innovative technology to help our customers drive their own succes. ■
23
Power Developer
24
TECH ARTICLES Every automated test system that tests electronic circuit boards, modules or equipment needs one or more DC power supplies. DC supplies provide power to the device under test as well as test stimulus. In some cases, the supply used to provide power to the DUT also provides a test stimulus by simulating the operating environment. For example, while most automotive electronics run at a nominal 12 VDC, the maximum input voltage may be as high as 27 VDC. Because this is so, some automotive standards require margin testing up to 27 VDC on a 12 VDC device. Necessities such as these determine power supply requirements. Let’s take a look at other common power supply specifications that you need to consider when selecting a power supply for an automated test system.
25
Power Developer
Linear or Switching Supply? The first choice you must make when purchasing a DC power supply is whether to select a linear supply or switching supply. Linear power supplies offer low ripple and noise specifications and have fast transient behavior. They are, however, inefficient and generate a lot of heat. They are also quite heavy. As a result, most engineers find them desirable only at lower output power levels (typically less than 500 watts). Most linear DC power supplies are benchtop supplies. One application for which a linear benchtop supply may be the best choice is when testing communication devices such as a radio or mobile phone or the demodulator module of a radar system. These devices have very sensitive discriminator or demodulator circuitry that work best with a low noise figure. To test the true performance of these units, we need to ensure that the DC power supply does not
Fig. 1
26
add any parasitic noise to the test setup, and because linear supplies have lower output ripple and noise than switching supplies, they are a better choice for this application. Linear supplies are also a good choice when power requirements are low. The main benefits of a switching supply are only relevant at higher output power levels. It is, therefore, less expensive to use linear DC power supplies in applications that do not require more than 100 W to 200 W per DC output channel. At this point, it is also important to consider the total power output of all the DC channels in your system. If your system has four channels or less, and the power requirements are relatively low, a good solution would be to use four linear supplies in a 19-in. rack-mount kit. If your system requires more output channels, or higher output power, then using switching supplies is a better option. They provide higher power density than linear supplies. By using switching supplies, you can have 12 DC outputs, providing up to 4,000 W of power in the same rack mount. Switching supplies are easier to control than linear supplies and cost about the same per channel. AMETEK Programmable Power provides a range of DC power supplies that offer the some of the highest power density available. The Sorensen ASD Series, for example, provides up to 30 kW of power in a 3U package. The water-cooled packaging allows it to be used in environments that normally exclude aircooled power supplies.
TECH ARTICLES Even in applications where low ripple and noise output are required, switching supplies are often more than sufficient. Recent developments in power electronics, such as zero-switching, have dramatically improved the ripple and noise specifications of switching power supplies. When you also consider that switching power supplies are more flexible than linear supplies and provide higher power density, it becomes apparent they are the favored choice for all but a handful of applications.
Transient response Transient response is a measure of how well a supply copes with changes in current demand or how well the supply follows changes in the load impedance. This is an important specification for many applications. When the output current demand decreases or increases significantly over a short period of time, the output voltage may also decrease or increase significantly. The power supply’s internal voltage control-loop will try to keep the output at the set voltage, but the response is not instantaneous. To get a faster transient response, you sometimes have to settle for more ripple and noise. Within the programmable power supply, the tradeoff is between the internal voltage control-loop and the output filter. A large output filter will limit the ripple and noise, but will make the supply slower to react to fast changing loads. A very fast internal voltage control loop will shorten transient response times, but may cause overshoot or undershoot, which may damage the Device UnderTest (DUT).
Fig. 2
A typical application example where transient response is important is mobile phone testing. In this application, the DC power supply simulates the mobile phone’s internal battery. When it begins to transmit, the current rise is very quick. This is not a problem for the internal phone battery, but for a programmable switching supply, it is a more difficult task. For this application, a linear supply would be a better choice than a switching supply because the power requirements are low, and the transient response of a linear supply is, in general, better than a switching supply.
Fig. 3
27
Power Developer Testing automotive relays and fuses is, however, another matter. For this application, the programmable DC power supplies must provide high currents at voltages up to 30 VDC, and typically, the power required is 5 kW to 10 kW. In this test, a large DC output voltage overshoot could damage the relay or fuse. To prevent this from happening, you want a supply that will control the DC output current instantaneously from zero to maximum output or from maximum to zero output. A practical technique to limit overshoot and undershoot is to use a pre-load. Putting a pre-load in parallel with the DUT and the DC output of the programmable power supply will now limit the percent current change, causing the DC voltage over- and undershoots to be significantly less. Imagine that 50% of the current travels through this additional pre-
load and 50% through the DUT. When the DUT creates a 100% current demand step, the power supply only sees a 50% current demand change. For the power supply to manage 50% in current demand changes, instead of 100%, it is much easier and almost eliminates the effect of high voltage overshoots and therefore eliminates any damage to the DUT. A simple inexpensive resistive load can be used in this case to function as a pre-load. Any ratio is fine. In other words, to obtain the transient response and overshoot specifications improvements it does not really matter if this load absorbs 40%, 50% or 60% of the current demand. A disadvantage of using a pre-load is that twice as much DC output current is required. Fortunately, if you use AMETEK switching power supplies, additional power is relatively inexpensive. This makes using pre-load a much cheaper and more practical approach than specialized power supply sub-systems for this specific application.
Slew Rate The next specification to consider is the DC output voltage slew rate (rise and fall time). To improve ripple and noise specifications, DC programmable power supplies have output filters that use large capacitors that store a lot of energy. It’s mainly the charge and discharge time of this filter, combined with the current demand of the DUT, that determine a supply’s voltage slew rate. The voltage slew rate is mostly independent of the connected DUT.
Fig. 4
Fig. 5
28
The DC output rise time for most AMETEK supplies is sufficiently fast for most applications. It is the DC output fall time that you must consider. The fall time depends not only on the internal LCR filter network at the DC output of the programmable power supply, but also of the connected DUT. If the current draw through the DUT is relatively low compared with the power supply current capacity, it can take many seconds before the energy stored in the output capacitors “leaks” away through the DUT. If the DUT requires a minimum current demand of at least 60% of the power supply capability, the stored energy will leak away instantaneously and the output voltage fall-time will be the shortest. Nevertheless, in most cases the DC output fall-time will be two to three times slower than the DC output rise time.
TECH ARTICLES One way to improve the DC output rise time is to choose a programmable power supply with a higher DC output range. For example, if the DUT is an automotive-related device and a 30VDC power supply would cover all test applications, choose instead a 60 VDC programmable supply, but only use up to 30VDC. The reason for this is that the output capacitor for the 60 VDC supply will be much smaller than the output capacitor for the 30 VDC programmable supply. This is to allow the voltage to rise from 0 V to full scale in the same amount of time for both supplies. In other words, when looking at the rise time in V/ms, the rise time for the 60 VDC power supply will be twice as fast. To improve the DC output fall-time, use a pre-load in parallel with the DUT or DC output of the power supply. Ensure that the total current demand of the pre-load and the DUT combined is at least 65% of the programmable power supply’s current capability. This approach requires more power from the supply as more DC output voltage range is required with the same output current demand. A typical output current slew-rate is 45A/ms. AMETEK Programmable Power also makes some DC power supplies to support solid-state laser applications. These are current sources with an output current slew-rate up to 400A/ms. Faster current slew-rates are possible by putting an electronic load in series with the power supply and using the electronic load as a current modulator. This combination allows for a current slew-rate up to 6000A/ms.
Load Regulation Another important specification of programmable power supplies is load regulation. This means some percent output voltage change from its set-point due to a change in the current demand of the DUT. Normally this effect should be very small (less than 0.01% of set output voltage).
Line Regulation Line regulation specifies the percent change of the DC output voltage or current as a function of AC input line voltage. This specification is important when the input line voltage is not stable.
Fig. 6
Fig. 7 Stability Stability is a measure of a supply’s long-term output voltage or current drift. A typical application is a magnet drive test, in which the programmable supply works as a current source in constant current mode. For this test, the user needs to be sure that the magnetic flux value is the same throughout the test. The supply must, therefore, set the DC output current to a specific value and maintain the value for the duration of the test. Stability is primarily specified in parts per million or ppm.
Parallel Operation If you need more output current than a single supply can provide, paralleling power supply outputs is generally the solution. AMETEK Programmable Power uses a dedicated control bus to connect its power supplies in parallel. The benefit of this dedicated bus
29
Power Developer unit. Any sense lines are connected only to the master unit. Realize that the total current is the sum of the current values displayed on each individual power supply. Some advanced models, such as the Sorensen SGI Series can compute and display the total system current.
Series Operation To supply a higher output voltage than is possible with a single supply, DC power supplies can be connected in series. All you have to do is to connect the positive terminal of one supply to the negative of another. This is true, but there are some limitations. Every programmable power supply has voltage isolation specifications, one for the negative to chassis isolation and one for the positive to chassis isolation. Ensure these voltages are not exceeded.
Fig. 8
A second limitation is that when operated in series, there are no master or slave units. What this means is that the power supplies need to be programmed individually. When using remote control for this, all interfaces need to be galvanically isolated through opto-couplers. Most DC power supplies offer a number of isolated interfaces, including analog, RS-232, RS-485 and Ethernet.
Analog Programming DC programmable power supplies typically provide a standard and isolated analog interface. Through the analog interface a supply’s DC output voltage, current and overvoltage-protection (OVP) can be set. These values are controlled by supplying a voltage signal, a current signal, or by connecting a resistor to the analog input.
Fig. 9 is that the total performance of the units in parallel still meets the original specifications for just one single power supply. The system configures itself automatically, identifying which unit is the master and which units are the slaves. With fast transient DUT’s it is sometimes recommended to use protection blocking diodes in the positive output line of each power supply. When paralleling supplies in this way, you can use supplies that have different current ranges, but all of the supplies should have the same output voltage range. All manual or remote control is done through the master
30
For example, you can use the analog output of a PLC to control the output voltage of a power supply. Or, you might use a thermistor to control the output of a supply. Also provided are voltage and current monitoring signal lines and control lines to enable or disable the power supply with millisecond reaction time. Local and remote sensing Many DC supplies can be configured for either local or remote sensing. For a more accurate output voltage setting, remote sensing should be used. In this mode, you sense the voltage where the power supply connects to the load. This method compensates for the voltage drop across the leads.
TECH ARTICLES If sense lines are long, it is recommended to use shielded cables to avoid any interference being superimposed on the main DC output. Remote sensing can often compensate for a voltage drop much larger than specified. One issue that can arise when using remote sensing is that transient response may be slower when the voltage drop across the power leads is high, but this is usually not a problem.
Constant-current Mode Although most supplies are used in constantvoltage mode, many applications call for using a DC power supply in constant current mode. When operated in constant-current mode, some features or specifications are not applicable. For example, in constant current mode, remote sensing is not a consideration. Neither is the output voltage set-point accuracy and resolution. What is important is accurate current control. Output voltage ripple and noise are not as important as output current ripple and noise. In constant current mode, the analog control can drive current changes at least 100 times faster than output voltage changes. The crossover from voltage to current mode is automatic. As soon as the current demand is larger than the set current limit, the DC power supply regulates its voltage down to match the set current limit and keeps the output current constant.
Inrush Current For some applications, inrush current is a major consideration. Inrush current is the instantaneous input current drawn by an electrical device when first turned on. Some loads, such as electric motors and power converters, draw high inrush currents. A power supply may need to be sized for the in-rush current. Digital control and measurement In general, the output voltage and current of a programmable supply is set most accurately, with the highest resolution, through its digital interface. As mentioned earlier, DC supplies typically offer many different interfaces, including RS-232, RS-485, USB, GPIB, ModbusTCP, Modbus-RTU, and Ethernet. In addition to the hardware, most DC power supply companies also supply the software you need to easily integrate your DC supply into your system. For example, AMETEK
supplies IVI drivers with each supply, and the supplies are programmed using standard SCPI commands. This makes system programming and system integration much simpler.
Accessories and Support The availability of accessories could also be an important consideration in your choice of power supply. For example, if you plan to rack mount the power supply, verify that one is available for your supply. Purchasing an off-the-shelf rack-mount kit is always less expensive than making one yourself.
Conclusion When selecting a DC programmable power supply, there are many parameters to consider. Electrical specifications are perhaps the most important, but you also need to consider the form factor, control needs, and even what accessories are available. By taking all of these into account, you’ll make the best choice for your application.
For More Information To learn more about the company’s programmable power supplies and programmable loads, contact AMETEK Programmable Power Sales toll free at 800-733-5427, or 858-458-0223, or by email at sales.ppd@ametek.com. Users can also contact an authorized AMETEK Programmable Power sales representative by visiting http://www.programmablepower. com/contact/
About AMETEK Programmable Power AMETEK Programmable Power offers the engineer’s most reliable and trusted power brands: AMREL, California Instruments, Elgar and Sorensen. AMETEK’s strong brands, broad product portfolio of AC and DC power supplies/loads, and deep application expertise across a wide array of industries make it the industry’s trusted “power partner.” AMETEK Programmable Power is a unit of AMETEK Electronic Instruments Group, a leader in advanced instruments for the process, aerospace, power and industrial markets and a division of AMETEK, Inc., a leading global manufacturer of electronic instruments and electromechanical devices with annual sales of more than $3.6 billion. ■
31
Power Developer
Perfection in P
Tackling Problems with I
LEDs are now widely used in many different areas. They allow for lightin solutions that were simply not possible with conventional lightin equipment. This means, however, that LED drivers need to meet ever mor complex, application-specific demands.
In many cases, two-stage drivers are the best solution. By separating the A power supply from the DC LED power, they offer new possibilities that ar simply not achievable with conventional LED drivers.
32
TECH COLUMN
Power
Innovative Concepts
ng ng re
Thomas Rechlin Senior FAE for Europe at RECOM Engineering Gmunden, Austria
AC re
33
Power Developer
“ While the good old incandescent light bulb will remain a common feature in our living rooms for years to come, LEDs have already overtaken the energy-saving bulbs that were introduced to replace conventional incandescent lamps.”
LEDs have simply become indispensible. It was just over 35 years ago when they first became a feature of electronic devices. We probably all remember the standard 5 mm LEDs that lit up in green or red to tell us whether a device was switched on or off. At that time, nobody could foresee that this electronic component would soon light up our homes and offices, and it took some time until LEDs were developed to serve as lighting devices. The first industry to see their potential was the automotive industry. Initially, they were installed only in luxury cars, mainly in rear lights and indicators. Car designers fell in love with LEDs and soon came up with many other lighting applications and designs. Today, there is probably no car manufacturer that does not use LEDs to create brand-specific “car faces.” Over the past few years, LEDs have conquered the indoor lighting market. While the good old incandescent light bulb will, however, remain
Fig. 1: Two-stage design with an AC/DC power supply (left) supplying a 24VDC or 48VDC lighting system, and DC LED drivers installed at, and powering, the individual lamps.
34
TECH COLUMN
a common feature in our living rooms for years to come, LEDs have already overtaken the energy-saving bulbs that were introduced to replace conventional incandescent lamps. The rapid advance of LEDs is not least due to the rapid development of reliable, easy-toinstall and versatile AC/DC LED drivers. They allow for the operation of LEDs in standard 230VAC networks. There are, however, many applications and fields where this simple conventional LED driver technology is not adequate; more about this in the following sections.
Principle Behind Two-stage Design Firstly, I would like to outline the alternatives to AC/DC LED drivers. These alternative solutions are all two-stage devices. As shown in figure 1, they consist of a power supply transforming the 230VAC power to a suitable direct current for the supply of one or more DC LED drivers. A lthough this is obviously a more complicated and generally also more expensive set-up, these disadvantages are outweighed by a number of clear advantages. First of all, two-stage driver solutions are more energyefficient. Although it sounds rather counterintuitive, many two-stage power supplies for LEDs are more efficient than single LED drivers. To illustrate this, please have a look at the example in the text box to the right. In a time of ever increasing electricity prices, it is up to end-users to decide whether they can afford the extra costs arising with singlestage installations. The second advantage is particularly relevant for lighting designers who want to devise modern and stylish LED lighting systems. The current trend is clearly towards filigree, non-obtrusive and “light� designs. LEDs are obviously the ideal lamps for such solutions. But where could the bulky AC/DC LED driver be hidden? Small DC LED drivers can easily be incorporated even into tiny luminaire bodies.
Real-World Example: A lighting installation consists of 20 LED downlights of 5W each. In a conventional system, each lamp is individually powered through a 5W AC/DC LED driver. These drivers have an efficiency rating of just about 65%. In other words, approx. 150W must be fed from the grid to achieve a lighting power of 100W. With a two-stage unit, a conventional 100W AC/DC power supply (typical efficiency rating 90%) is combined with suitable DC LED drivers at each LED downlight (efficiency rating 95%). The overall efficiency is therefore 85.5%, and the same light output can be achieved with a power consumption of only 117W. This is 33W less than with the single-stage solution! Assuming that the lamps are on for six hours per day on average, energy saving amounts to nearly 75kWh per year. This corresponds to the power consumption of a modern fridge over six months.
35
Power Developer
Fig. 2: Circuit diagram of emergency lighting system. In the event of a grid power failure, the DC LED driver is switched on, powering the lamp from the 12 V battery. The battery is recharged through the charge controller the moment the grid power returns.
Tailor-made Solutions for Special Applications
“ LEDs are a great solution for even the most demanding applications – provided that the right approach is chosen.”
36
Many LED applications can simply not be implemented with single-stage drivers. This is not least due to the numerous regulations and standards that prohibit certain installations. To further illustrate this, let’s go through a few examples. Let’s have a look at medical technical applications such as the lighting in operating theatres. Very bright lamps need to be installed above the operating table. One would think that LEDs would be the perfect solution here, as they can produce light of any hue and colour, so that surgeons could work under conditions that are similar to natural daylight. However, EN 60601 (3rd edition – Medical technical equipment and systems), conventional LED drivers are not permissible for installation above operating tables in theatres. Special LED drivers approved for use in medical environments are, however, difficult to find and very expensive. The obvious solution would consist of a separate AC/DC power supply certified according to the above standard, and LED lamps with reliable long-life DC LED drivers.
TECH ARTICLES
Another field where the new two-stage LED driver technology would be of great advantage are railways – be it reading lights in first-class carriages or signalling devices. As railway systems are generally powered with 110VDC, conventional products are not suitable and there are very few LED drivers rated for this input power available in the market. In addition, electric systems in trains are subject to frequent voltage peaks, shorttime interruptions and power pollution. These problems can be overcome by installing LED drivers powered through a power supply certified according to EN 50155 (Railway applications – Electronic equipment used on rolling stock). Such installations also have the advantages that the individual LED drivers do not need to conform to the above standard. Our third example concerns dual systems. These are lighting solutions that can be powered with direct or alternating current. A typical example would be emergency lighting systems. Figure 2 shows the general circuit design with its core components of AC power supply, charge controller, battery, and DC LED driver. Dual systems are bound to become even more widespread, as houses of the future are likely be powered from on-site renewable sources such as solar panels and wind turbines, as well as from the grid. Solar panels, for example, produce direct current. In order to make best use of the generated power, any excess energy generated during the day could be stored in buffer batteries for lighting in the evening. To do this, an internal DC onboard network would be required. Otherwise, the power loss due to transformation and rectification at the respective LED downlight would amount to a staggering 25%.
On the one hand, a power supply suitable for the actual application must be chosen. On the other, small and reliable DC LED drivers must be built into the lamps. RECOM offers a wide range of suitable products that can be combined in the above way. Choose from our AC/DC and DC/DC converter range that includes devices certified for many different applications. When it comes to LED drivers, we offer everything you need, from standard models such as the RCD-24 to the versatile RBD-12. The RBD-12 is particularly recommended for battery-powered LEDs, as its buck-boost design is tailor-made to handle fluctuating input voltages. To find out more, watch our YouTube clip: http://www.youtube.com/watch?v=0Rzvj3 QX1vI&list=PL62B0081D4254B0FA. To open up new markets and stay ahead of the competition, we need to broaden our outlook and venture into new territory, committed to innovation.
To read the previous installment, click the image below:
One-fits-all Solution or Innovative Approach? Conclusion: LEDs are a great solution for even the most demanding applications – provided that the right approach is chosen. It is, however, simply not possible to devise a “one-fits-all” solution. To tackle the problem, one needs to consider two issues.
37
M o v i n g To w a r d s a
David Elien VP of Marketing & Business Development, Cree, Inc.
Clean Energy
Let There Be
LIGHT
FUTURE
How Cree reinvented the light bulb
— Hugo van Nispen, COO of DNV KEMA
Cutting Edge
SPICE
Modeling
MCU Wars 32-bit MCU Comparison
+
Cutting Edge Flatscreen Technologies
+
New LED Filament Tower
View more EEWeb magazines— Click Here
Power Developer O ct o b er
201 3
From Concept to
Reality
Sierra Circuits:
Designing for
Durability
A Complete PCB Resource
Wolfgang Heinz-Fischer Head of Marketing & PR, TQ-Group
TQ-Group’s Comprehensive Design Process
Freescale and TI Embedded Modules
+
Ken Bahl CEO of Sierra Circuits
PLUS: The “ Ground ” Myth in Printed Circuits
+
+
PCB Resin Reactor
ARM Cortex Programming
Low-Power Design Techniques