Design TipDesign Tips s€¦ · AGS Media Group and Power Systems Design Europe ... and market...

ISSN: 1613-6365

E m p o w e r i n g G l o b a l I n n o v a t i o n O c t o b e r 2 0 0 9

Design Tips Design TipsSpecial Report – Powering Freight & Transportation

http://www.linear.com/3507

http://www.digikey.com/europe

Viewpoint Long-Haul Readiness, By Cliff Keys, Editor-in-Chief, PSDE .....................................................................................................................................4

Industry NewsTI Opens World’s Most Advanced Analog Manufacturing Facility in Richardson, Texas ..........................................................................................6Digi-Key Corporation and Cooper Bussmann Expand Distribution Agreement .......................................................................................................6ABB Sees Energy, Climate Change & Emerging-Markets Driving Demand ..............................................................................................................7Farnell Increases Focus on Design for Transportation Applications .........................................................................................................................7

High Power 1200A 4500V IGBTs for Transportation .................................................................................................................................................8

The Challenges and Opportunities for Power, By Thomas Grasshoff, Semikron ....................................................................................................10

Power Semiconductors Hit the Bottom, Reported By Ash Sharma, IMS Research ...............................................................................................11

Design Tips Frequency Response of Switching Power Supplies - Part 8, By Dr. Ray Ridley, Ridley Engineering .....................................................................12

Long-Term LED Maintenance, Reported By Cliff Keys, Editor-in-Chief, PSDE .......................................................................................................16

Redefining the Digital Power Market, Reported By Cliff Keys, Editor-in-Chief, PSDE ............................................................................................19

Cover Story Latest Generation IGBT Gate Drivers, By Michael Hornkamp, Sascha Pawel and Olivier Garcia, CT-Concept Technologie ................................21

Power Supplies No Load Power without Compromise, By Neil Massey, CamSemi .........................................................................................................................24

Special Report: Supplying the Power Grid Designing for Railway Applications, By Marco Panizza, Vicor ................................................................................................................................28Thermal Gap Filling Revolution and Evolution, By Eoin O’Riordan, Parker Hannifin Ltd. .......................................................................................32High Efficiency Rectifier Technology, By Morten Schoyen, Eltek ...........................................................................................................................35The Art of Motor Design, By Dr. Stephan Chmielus, Fairchild Semiconductor .......................................................................................................37Automotive Power, By Henning Hauenstein, International Rectifier .......................................................................................................................40EV/HEV Battery Management, By Greg Zimmer, Linear Technology ......................................................................................................................42Mastering Power Modules, By Werner Obermaier, Vincotech ................................................................................................................................45

Green: It Doesn’t Happen by Itself? Reported By Cliff Keys, Editor-in-Chief, PSDE ..............................................................................................48

Dilbert – 48

Power Systems Design Europe Advisory Board

MemberArnold AldermanHeinz RüediMarion LimmerEric LidowDr. Leo LorenzDavin LeeTony ArmstrongHans D. HuberAndrew Cowell

RepresentingAnagenesisCT-Concept TechnologyFairchild SemiconductorIndustry LuminaryInfineon TechnologiesIntersilLinear TechnologyLEMMicrel

MemberMichele SclocchiMike NoonenKirk SchwiebertChristophe BassoBalu BalakrishnanPaul GreenlandUwe MengelkampPeter Sontheimer

RepresentingNational SemiconductorNXP SemiconductorsOhmiteOn SemiconductorPower IntegrationsSemtechTexas InstrumentsTyco Electronics

http://www.IGBT-Driver.com

VIEWPOINT

Power Systems Design Europe October 2009�

a major positive for the PC industry, prompting a huge advertising and media campaign from Microsoft that will put PCs in high-profile positions using the strength of mainstream electronic and print media.

Powering signageAccording to IMS Research, there

remains a healthy medium term growth outlook in the world market for LED driver ICs in signs, signals, outdoor displays and traffic lights.

Large outdoor LED displays are an

increasingly important business area, powered by LEDs from suppliers such as Cree and Nichia, primarily used in areas such as sports, concerts and outdoor advertising due to their high quality, full colour images in much larger sizes than can be achieved by other technologies.

The signage segment represents a strong addition to the market with good growth potential. Leading suppliers of LED driver ICs for signage include Macroblock, Toshiba and Texas Instruments.

Times are still tough, but for the longer-haul there are great opportunities to be had in the power industry. I truly believe the only way forward for power companies is to invest in the engineering of innovative, value-adding future products that fit the fast-changing needs of our society, rather than looking at the kind of short term gains we’ve witnessed in the banking industry.

Enjoy the issue. Please keep your design projects and feedback coming, and check out our fun-site, Dilbert, at the back of the magazine.

All the best!

Editor-in-Chief, [email protected]

Welcome to this issue of PSDE, where we have explored the theme of ‘Powering Freight and Transportation’ here in the print magazine, with further reporting in PSD’s increasingly popular online magazines.

There are, as can be seen from the reports in this issue, growing opportunities to bring a brighter future for our colleagues as power technology paves the way for better and more energy efficient systems. Power now plays a key role in the majority of new designs and even in the medium-term there is much good work to do for creative power engineers.

It is interesting to note that some firms are now ‘talking-up’ the economy and releasing details of better profit-margin performance to their shareholders. Many of these firms have simply slashed ‘headcount’ and dumped valuable engineering talent into the unwelcome hands of the state, damaging their own credibility and future readiness. There are also longer-sighted companies who have benefited from these individuals’ fate and are indeed ready for the recovery, while the others knee-jerk back into costly panic-recruitment and struggle to hold onto their remaining but demotivated staff.

PCs power-upThe PC market also made one small

step away from the abyss in the second quarter as it achieved sequential growth in unit shipments for the first time in six months, according to iSuppli Corp.

The launch of Windows 7 should be

Long-Haul Readiness

Registration of copyright: January 2004ISSN number: 1613-6365

AGS Media Group and Power Systems Design Europe

magazine assume and hereby disclaim any liability to

any person for any loss or damage by errors or omis-

sions in the material contained herein regardless of

whether such errors result from negligence, accident

or any other cause whatsoever.

Cover graphics courtesey of Analog Devices, Inc.

All rights reserved and respective partiies.

Send address changes to:[email protected]

Free Magazine Subscription, go to: www.powersystemsdesign.com

Volume 6, Issue 8

AGS Media Group146 Charles Street

Annapolis, Maryland 21401 USA Tel: +410-295-0177 Fax: +510-217-3608 www.powersystemsdesign.com

Editorial Director, Power Systems Design China, Europe & North America Editor-in-Chief, Power Systems Design Europe & North America

Cliff [email protected]

Contributing EditorsLiu HongEditor-in-Chief, Power Systems Design China

[email protected]

Ash Sharma, IMS [email protected]

Dr. Ray Ridley, Ridley [email protected]

Publishing DirectorJim Graham

[email protected]

PublisherJulia Stocks

[email protected]

Circulation ManagementKathryn [email protected]

Research DirectorMeghan [email protected]

Magazine DesignBeata Rasmus, Eyemotive [email protected]

Production ManagerAbby [email protected]

http://www.bergquistcompany.com/screenprint

INDUSTRY NEWS INDUSTRY NEWS

Power Systems Design Europe October 2009� �www.powersystemsdesign.com

mies, and that this trend is likely to continue. “Despite the current economic crisis, the

long-term growth factors for ABB remain unchanged: Steadily rising electricity demand, the rapid rebound of the emerging-market economies and the need to feed renewable power into existing grids require solutions for more and better power infrastructure, energy efficiency and industrial productivity,” Hogan said. “As a high-tech infrastructure company and market leader for power and automation technologies, ABB is ideally positioned to offer solutions to these challenges, and we are continuing to invest aggressively in new technology.”

An ABB analysis of industrial energy consumption carried out this year has shown that $180 billion could be saved annually by improving energy efficiency and productivity with the best-performing technologies on the market today.

ABB has asserted the economic crisis is re-inforcing three trends that have been driving demand for the company; the battle against climate change, the search for energy and process efficiency, and the strength of emerg-ing markets.

The trends will extend beyond the current economic cycle and will increasingly shape ABB’s research and development activities, as well as decisions on where to locate new facilities, CEO Joe Hogan said recently in Zurich.

Efforts to lower greenhouse-gas emissions and especially to promote renewable energy have gained new momentum under the stimu-lus plans of many governments, Hogan said. In addition, awareness that energy prices will rise over the long term, combined with the need to cut costs to weather the current downturn, are driving industries to improve the efficiency of their energy consumption and processes.

ABB also said that approximately 75 per-cent of the growth it has seen in the past five years has come from the emerging econo-

www.abb.com

ABB Sees Energy, Climate Change & Emerging-Markets Driving Demand

Joe Hogan, CEO, ABB

of news, product and market information. Technical articles and FAQs designed to

Farnell has extended its support for electronic engineers working on automotive, rail and other transportation applications by adding a large number of transportation-ori-entated electronic components to its product portfolio. At the same time the company has made available a wide variety of technical collateral that will help these engineers speed product selection and application develop-ment.

The recent addition of over 800 new parts means that Farnell’s dedicated Transportation portfolio now features almost 4000 products from 80 industry-leading suppliers includ-ing Altera, Analog Devices, Cree, Freescale Semiconductor and STMicroelectronics. Tech-nical documentation for electronic engineers working in the Transportation field comprises over 60 new documents ranging from applica-tion notes and selector guides to design tips and reference guides.

To coincide with Farnell’s increased trans-portation focus, the latest Farnell Technology First Journal also has transportation as its theme. The Journal brings together a variety

Farnell Increases Focus on Design for Transportation Applications

help engineers working on applications as diverse as motorcycles and heavy agricultural vehicles are also included.

Downloadable PDF and interactive online versions of Technology First can be accessed via the Transportation link on the Technology First area of the Farnell website. This link also provides access to information on the new product introductions, allows visitors to search products by category, and incorporates ‘hot-links’ to the technical collateral.

Jamie Furness, Global Head of Strategy & Community Development, states: “From comfort and safety to infotainment, security and drive-trains, innovative electronic design is fundamental to delivering more features, making vehicles safer and more efficient and, critically, reducing cost of manufacture. The key to future success for manufacturers lies in innovation so, despite the short-term challeng-es facing the Transportation industry in general and automotive manufacturers in particular, this is the perfect time to increase our focus on this sector.”

www.farnell.co.uk

TI Opens World’s Most Advanced Analog Manufacturing Facility in Richardson, TexasIncreased volume of energy-efficient chips and new jobs in North Texas

City of Richardson, Collin County, the Plano Independent School District and the Collin County Community College District.

“Texas Instruments’ decision to again invest in Texas is yet another example of how, even during these economic times, the Lone Star State remains the top choice for companies looking to expand,” Gov. Rick Perry said. “Our combination of a predictable regulatory climate, low taxation and a world-class workforce that is well prepared to fill the hundreds of new high-technology jobs TI will

Texas Instruments Inc. has announced the opening of its new manufacturing facility in Richardson, Texas. The company expects to begin moving equipment into the facility in October.

Known as RFAB, (‘R’ for Richardson, ’FAB’ for Fabrication), the fab will be the world’s only production facility to use 300-millimeter (12-inch) silicon wafers to manufacture ana-log chips, which are essential components in virtually all electronics. The facility will give TI a strategic advantage in high-volume pro-duction because thousands of analog chips can be etched onto each of these wafers, more than double the number on the more commonly used and smaller 200-millimeter wafers.

“The time is right for this investment,” said Rich Templeton, TI’s chairman, president and CEO. “Customer demand for analog chips is growing, and there’s tremendous desire to save energy and protect the environment. The chips produced here will help our customers make thousands of electronic products that are more energy-efficient. It is significant that these devices will be made here, in North Texas, in one of the industry’s most environ-mentally responsible fabs.”

The facility will produce analog integrated circuits based on TI’s proprietary process. Customers will use these chips in electronics ranging from smartphones and netbooks, to telecom and computing systems.

At a ribbon-cutting ceremony with local and state officials, Templeton said TI plans to ship the first chips from this facility by the end of 2010. When the first phase of equipment is ramped and producing at full capacity, the facility will be capable of shipping more than $1 billion worth of analog chips per year.

Hiring will begin immediately for 250 jobs in RFAB. “These are high-quality, well-paying engineering, manufacturing and administra-tive jobs for our North Texas region. The in-frastructure that a facility like this requires will create other indirect jobs with suppliers and support services,” said Templeton. “We want to thank our great partners in Richardson, Dallas, Plano and Austin who helped us make this happen,” Templeton said, referring to the

be bringing to the area.”

Recent expansionsThe opening of RFAB is the most recent in

a series of manufacturing expansions by TI. Earlier this year, TI opened Clark, an assem-bly and test facility in the Philippines. TI also has been installing new test equipment at several other locations, and is in the process of installing newly acquired 200-mm manufac-turing equipment for analog chip production at sites around the world, including Dallas.

Local education benefits

Local education has benefitted from TI’s decision to build the fab in Richardson. As part of the original agreement between community and state partners, the nearby University of Texas at Dallas will receive a total of $300 million from the Texas Enterprise Fund, the Texas General Land Office, the UT System, and private donors for improvement of its engineering and research programs.

“Texas Instruments has been a remarkable partner with education at all levels,” said Dr. David E. Daniel, president of UT Dallas. “The impact of the RFAB’s creation on UT Dallas has been dramatic in terms of recognition and research activity. This project’s high profile played a direct role in our most recent efforts to reach for what many call Tier One status—our aspiration to become a nationally recog-nized research university. We look forward to partnering with Texas Instruments and the rest of our community as we grow toward this stature in service to the Dallas area and the region.”

Green model

RFAB has been an important model of green construction. It was the first semicon-ductor facility to achieve Gold certification with the U.S. Green Building Council’s Leadership in Energy and Environmental Design (LEED) program. TI has applied knowledge from the RFAB designs to other facilities all over the world.

For more info, visit: www.ti.com/rfab

TI’s manufacturing facility (RFAB) in Richardson, Texas. RFAB is 1.1 million square feet of building space that includes administration, mechanical, support and fabrication buildings on 92 acres of land.

www.ti.com

Cooper Bussmann products offered by Digi-Key include Cooper Bussmann circuit protection fuses and ESD suppressors, Coil-tronics inductors and transformers, and Pow-erStor supercapacitor lines. Stocked products

Electronic component distributor Digi-Key Corporation has announced today the expan-sion of its distribution agreement with Cooper Bussmann from a North American agreement to a global agreement.

Digi-Key Corporation and Cooper Bussmann Expand Distribution Agreement

are featured in Digi-Key’s print and online catalogs and are available in prototype/design and production quantities on Digi-Key’s global websites.

“We at Digi-Key are very pleased to extend

www.digikey.com

our distribution contract with Cooper Buss-mann and offer its quality products to custom-ers worldwide,” said Jeff Shafer, Digi-Key vice president of interconnect, passive, and electromechanical products.

Cooper Bussmann is a global manufacturer offering a broad range of products, including magnetics, overvoltage and overcurrent cir-cuit protection devices, and supercapacitors. Through innovative technology and custom-

engineering designs, Cooper Bussmann provides unique solutions for today’s market-driven requirements.

Power Systems Design Europe October 2009�

Hitachi’s two new high current 4500V IGBT module options address the distinctive needs of two high

power field applications; Heavy Rolling Stock in particular often requires low conduction losses without

the penalty of de-rating system output power due to a lack of device robustness. Hauling freight and

passengers across long and difficult terrain, often compounded by high altitudes, requires careful

consideration of both the IGBT manufacturer and the system designers.

duction losses. Simply reducing Vce(sat) is one satisfactory method of reducing dominant sys-tem losses which will allow for an output current increase. Hitachi’s fine pattern E2 IGBT process can effectively manage this demand for higher power density (allow-ing a 33% reduction in Vce(sat) and up to 45% for diode VF), but critical to locomotive

applications will become other often forgotten stresses that will indirectly limit benefits gained elsewhere. These are likely to multiply with increased output currents and increased maximum operating junction temperatures; both of which are already employed in lower voltage IGBTs.

Using ideal shape processing the improvement in thermal contact, particularly directly under the diode chip centres, allows for more efficient heat dissipation during operation and under fault conditions. With a reduc-tion of almost 20% in the guaranteed Rth(c-f), thermal stress experienced by the diodes can be reduced. With lower maximum junction temperatures under repetitive diode surge current conditions there is less need to de-rate system out-put current to compensate rare failure mode events like cosmic ray strikes.

ented super high reverse recovery (Super HiRC) processing Hitachi’s dual diodes, rated �00A and 1200A can be selected with an emphasis on low VF or low Err operation, optimising total power losses and application durability. Additional product options of �.4kV and 10.2kV isolation will support air cooled, deion-ised and non deionised water cooling.

Modern locomotives will typically em-ploy both an input converter to manage the catinerary input power source and output inverter for traction. This will lead to separate demands of the device char-acteristics, taking into account the region of operation and driving frequency. For this reason HITACHI supports standard and –H options, allowing best perfor-mance optimisation for both 16.67Hz and 50Hz high voltage supplies.

Operating at relatively low switching frequencies in the inverter side (<400Hz) will lead to a dominance of system con-

Building upon the successful 3.3kV E2 series,

Hitachi Europe has in-troduced the two new high current 4500V IGBT modules offer-ing low conduction loss and high speed. Rated 1200A the MBN1200H45E2 and MBN1200E45E2-H modules are favour-ably positioned to sup-port high power 2 or 3 level drives solutions for transportation, heavy industrial and renewable power generation applica-tions.

Using E2 fine pattern silicon process-ing IGBT collector-emitter voltage is reduced 24-33% when compared to ex-isting 900A technologies on the general market. Depending on the anticipated application and operating frequency the designer may consider either: H45E2 with 3.7V Vce(sat) (125ºC); or H45E2-H 4.2V Vce(sat) (125ºC). Optimised IGBT vertical design and Super HiRC FWD ensure robust short circuit and diode surge capabilities, paramount for use in demanding locomotive propulsion sys-tems. High speed H45E2-H will support higher frequency applications, notably grid connection inverters in renewable energy generators.

Partnering the IGBTs, input rectifiers and clamping diodes provide a total system solution. Using Hitachi’s pat-

High Power 1200A 4500V IGBTs for Transportation

www.hitachi.eu/pdd

http://www.ridleyengineering.com

Power Systems Design Europe October 200910 11www.powersystemsdesign.com

of the 15 largest suppliers of power semiconductors experienced a decline in their business in 2008, with all of them likely to see further drops this year.

One of the largest gains in market share was achieved by Texas Instruments. Although it only held a 4.5% share of the $26 billion power semiconductor market, it was again ranked #1 for power ICs, by growing its share to 9.1% in 2008. As well as penetrating new markets with its power IC products, the company also re-entered the power discrete market last year with its acquisition of CICLON Semiconductor.

Outside of the top 15, one of the largest revenue and market share gains was seen by Hitachi, driven by its success in the growing traction market. Traction is often overlooked within the power industry, particularly by those companies focusing on higher volume, faster moving sectors like computing and consumer. Although very small in comparison, traction was in fact the fastest growing sector for power discretes and modules in 2008, growing by around 25% - not insignificant given this sector actually generated more revenues than the closely watched renewable energy sector.

Even with the help of relatively buoyant sectors like renewable energy and traction, a strong rebound in growth for power semiconductors is not predicted and instead a gradual return to growth is anticipated. Despite the cautious sentiment within the industry that the bottom has been hit and that growth will soon return, annual growth of just 8% is forecast for 2010. Furthermore, IMS Research predicts that the power semiconductor market will not return to its 2008 size until 2012.

ments in semiconductor technology now allow for higher operation temperatures.

Innovative technologies provide solutions

These problems are all interdepen-dent factors. It therefore makes sense to search for an integral solution rather than looking at the problems as isolated matters.

The challenge faced by the designer concerning the base plate and solder connections can be resolved with Semi-kron’s SKiiP technology, where the base plate and large, fatigue-prone solder connection to the substrate are removed entirely. A patent-protected pressure contact system is used instead. In the pressure contact system, the substrate is pressed onto the heat sink by way of mechanical pressure. As the ceramic substrate is relatively flexible and the pressure applied by way of mechani-cal “fingers” located at several points, very close contact between the ceramic substrate and the heat sink is reliably achieved. As a result, the thermal paste layer can be reduced to a minimum of just 20-30µm. By way of comparison, the thermal paste layer in conventional modules with base plates is in the order of 100µm.

But despite the current economic situ-ation, the drives and renewable energy sectors will continue to play a vital role in boosting industrial production and employment rates for the future. The power semiconductor industry has already taken on the challenges ahead. The technological demands posed by hybrid and electric vehicles, as well as new materials such as SiC and GaN will certainly pave the way for new develop-ments.

pactness of the system are essential for customers. For manufacturers of power semiconductors, this translates to a particularly difficult challenge in meet-ing these, in some respects, conflicting requirements. Furthermore, as inverter power increases, parallel module con-nection and heat management become increasingly important.

New challenges for power module manufacturers

In a conventional soldered power module with a base plate, the solder connections often constitute one of the weak points of the module. The base plates themselves, for modules with large dimensions and consequently high power output, can only be optimised with some difficulty in view of best thermal and mechanical performance. For modules of 200A and above, sev-eral semiconductor chips have to be connected to the DCB (Direct Copper Bonding) ceramic in parallel in order to achieve modules with increased current ratings. Chip temperatures have to be carefully considered in the light of reli-ability and here, the significant improve-

The Challenges and Opportunities for Power

By Thomas Grasshoff, Head of Product Management, SEMIKRON International

Environmental policies and stron-ger consumer preferences in the deployment of energy in for exam-

ple, industrial, automotive and transpor-tation applications means that products and equipment must meet new levels of efficiency, service life and compactness. Power electronics is a key technology to assure future mobility with electric and hybrid vehicles, the answer to increasing emissions and limited natural resources. Manufacturers are developing new assembly and connection technology, offering higher current densities and reli-able chip temperatures, as well as using new semiconductor materials.

Profitable to the power semiconductor industry

The power semiconductor industry will profit from the forthcoming growth in the HEV and renewable energies market in two respects. Firstly, power semicon-ductors are needed for energy conver-sion itself – for instance in inverters in wind power plants. Secondly, semicon-ductors are the core element of variable-speed drives.

Owing to their technical superiority as well as for reasons of user-friendliness, modules are used predominantly as electronic switches. A module compris-es a silicon chip, an insulated ceramic substrate and a module case, housing the necessary power connections. For the designer, there are many issues to consider regarding assembly and connection technology, as well as the integration stage, for example includ-ing integrated driver, current sensor and heat sink.

In the world of industrial drives, HEVs and renewable energy generation, reli-ability is top priority simply because this is what maximises economic operation. In addition, high efficiency and com-

www.semikron.comwww.imsresearch.com

Power Semiconductors Hit the Bottom

Now that global economic prospects finally appear to be on a rising trend again, it seemed

like an opportune moment to review the state of the power semiconductor market, and check out which suppliers suffered most during the downturn, and which are best placed to take advantage of any recovery.

Latest projections from IMS Research’s Power Semiconductor Intelligence Service show that the market will fall by 18.5% in 2009. This includes all power discretes, power modules and power ICs. This comes a year after the market grew by just 3.2% - a poor result given historical results and largely due to the dismal fourth quarter when the global economy nose-dived.

According to the latest data, Infineon Technologies retained its position as the world’s largest supplier of

power semiconductors, increasing its market share to 9%. It was one of the few leading suppliers to record revenue growth in 2008. In total, eight

By Ash Sharma, Research Director, Power & Energy Group, IMS Research

Global Power Semiconductor Market Rankings

2008 Rank Supplier

23456789

1011121314

1

15

STMicroelectronicsFairchildTexas InstrumentsInternational RectifierVishayNational SemiconductorToshibaMitsubishiMaximLinear TechnologyON SemiconductorFuji ElectricRenesas

Infineon

Rohm

Others

Total Market Size:Source: IMS Research Aug-09

$25.5 Bn $26.3 Bn 3.2%

6.5%4.5%4.0%4.0%3.5%4.0%4.0%3.0%3.5%3.0%3.0%3.0%2.5%

9.0%

2007 Share 2008 Share Change

2.0%

40.5%

6.5%4.5%4.5%3.5%3.5%3.5%3.5%3.5%3.5%3.0%3.0%2.5%2.5%

9.0%

2.0%

41.5%

0.0%0.0%0.5%-0.5%0.0%-0.5%-0.5%0.5%0.0%0.0%0.0%-0.5%0.0%

0.0%

0.0%

1.0%

DESIGN TIPS

Power Systems Design Europe October 200912

By Dr. Ray Ridley, Ridley Engineering

Step Load TestingThe seventh article of this series

showed three different converter step-load responses. For the first response, it was shown how a simple gain reduc-tion stabilized the system. However, the other two responses do not improve with a reduction in gain, as we will see, and the step-load test offers no guid-ance on how to improve the design.

Buck Converter Transient Response – too Little Gain

Figure 1a shows a step-load transient response for a buck converter with voltage-mode control. The converter has a 4 kHz oscillatory response, indicating insufficient phase margin.

In the previous article, reduction of loop gain was the proper solution to

improve the response. However, in the buck converter case shown here, reduc-ing the gain makes the stability problem worse. The step load response is even more undamped, as shown in Figure 1b

The proper solution in this case is to increase the gain of the feedback loop, resulting in the response of Figure 1c.

Looking at the loop gains of Figure 2 gives us insight into what is happening. The green curve shows the original gain, the red curve shows the decreased gain, and the blue curve shows increased gain. The phase margins at the cross-over frequencies give us the characteris-tic transient response. For the red curve,

Frequency Response of Switching Power Supplies –

Part 8In this article, Dr. Ridley continues the topic of frequency response measurements for switching power

supplies. This eighth and final article in the series shows how step-load testing cannot provide sufficient

insight into how to redesign systems that are near instability.

More Step-Load Transient Testing

Figure 1: Transient Load Response of Buck Convert-er (a) Before (b) After Gain Reduction and (c) After Gain Increase.

http://www.intersil.com/power

DESIGN TIPS DESIGN TIPS


we can see that the loop crossover has dropped close to the LC filter resonant frequency, resulting in only 10 degrees phase margin.

For the blue curve, the crossover is moved to a much higher frequency, away from the LC filter resonance, where the phase delay is much less. This results in a stable system with 60 degrees phase mar-gin. Further changes can be made to stop the phase dropping down close to -180 degrees, and the loop gain information provides clear information on how to proceed with this.

The example given in the last article had a loop gain which crossed over in a region where the phase mar-gin was decreasing with frequency. For the buck con-verter with voltage-mode control, it is common for the opposite to be true. Just above the filter resonance, the phase improves as the frequency increases. This information is only available from the loop gain, and cannot be seen from the step-load response.

Flyback Transient Response – Incorrect Compensation

Figure 3a shows the transient response of a flyback converter, with a 5 kHz oscillation. A reduction in gain of the loop results in the waveform of Figure 3b, and the system is still only just stable. The result of an increase in gain is shown in Figure 3c, and the system remains undamped. For this system, the ringing can-not be eliminated by changing the gain alone, and more complex adjustments must be made.

The loop gains of Figure 4 give much more insight. The green curve shows the original gain, the red curve shows the decreased gain, and the blue curve shows the increased gain. The phase margins at the differ-ent crossover frequencies are 15, 35, and 40 degrees respectively. This is not enough to guarantee stability for a product, and additional changes must be made to improve the phase margin.

The gold curve shows the loop gain with the first compensation zero moved to a lower frequency. With this change, the phase margin is increased to 60 de-grees, and the design is optimized.

Trying to arrive at this new compensation with only step-load testing would be extremely difficult. There are so many degrees of freedom in changing the com-pensator, it is unlikely that an optimal design would be reached. Only by looking at the loop gain can we see which design approach is correct.

Summary This article shows how step-load testing is not

useful for correcting systems that have insufficient stability margin. In general, power supplies have very complex control loops, especially when the effects of additional filters, multiple outputs, and different modes

Figure 2: Loop Gain and Phase of the Buck Converter.

Figure 3: Transient Load Response of Flyback Converter (a) Before (b) After Gain Reduction (c) After Gain Increase (d) After Compen-sation Shaping.

Figure 4: Loop Gain and Phase of the Flyback Converter.

of operation are considered. Relying on the very limited data available from a step-load test to guarantee stability over the lifetime of a product is not recom-mended.

The last eight articles in this series have addressed the issue of frequency response measurement, a powerful tool for development of power supplies and power electronics systems. There are many more topics in this area, and I will return to some of the more complex is-sues in future articles. www.ridleyengineering.com

To Receive Your Own FREE Subscription to Power Systems Design Europe, go to:

www.powersystemsdesign.com/psd/subslogn.htm

References1. “Frequency Response of Switch-

ing Power Supplies, Parts 1-7”, Power Systems Design Magazine, Design Tips Archive. http://www.powersystemsde-sign.com

2. “AP Instruments AP300 User Manu-al”, http://www.apinstruments.com/files/Model300.pdf

3. Assorted articles on frequency response, Switching Power Magazine, www.switchingpowermagazine.com.

http://www.intersil.com/ibuy

http://www.powersystemsdesign.com


directly and precisely controlled up to 150°C.

Long-Term Testing Observations The LED industry currently examines

junction temperature and drive current to predict an LED lamp’s long-term lu-men maintenance. However, Cree has observed that high ambient air tem-peratures also play an important role in the long-term lumen maintenance of silicone-encapsulated LED lamps.

Conventional wisdom dictates that the (TAIR=25°C, TSP=90°C, IF=1.000A) case should have the worst lumen maintenance of the three cases be-cause it has the highest junction temperature and drive current. Cree has observed repeatedly that in fact, higher ambient air temperatures will acceler-ate lumen depreciation to a degree that cannot be observed through TAIR = 25°C testing.

Therefore, there is a third critical factor that affects the rate of lumen

trolled by the drive current – the higher the drive current, the higher the result-ing TSP.

High Junction Temperature Testing In the High Junction Temperature

Testing configuration, TAIR is held at 25°C, while the LEDs are attached to thermo-electric coolers (TECs) so that the LED TSP (and thus the TJ) can be

cuit boards (MCPCBs). A set typically contains thirty in¬dividual XLamp LED lamps. The boards are then attached to heat sinks in environmental test chambers. The TSP of each LED lamp is actively monitored and controlled by continuously regulating the tem-perature of the heat sinks. Ambient air temperature (TAIR) in the chamber is also actively monitored and controlled by regulating the temperature of the air flowing through the chamber. Per LM-80 4.4.2, TAIR in the environmental chambers is controlled to be held within 5°C of TSP.

Reported by Cliff Keys, Editor-in-Chief, PSDE

this information to develop methods of predicting L70 lifetimes for LED lamps under a wide range of operating param-eters.

TJ = TSP + ([Rth j-sp] x [VF] x [IF])

Cree LM-80 Compliant Lumen Maintenance Test Configuration

Cree tests XLamp LED lamps for long-term lumen maintenance consis-tent with LM-80 methods. Specifically, sets of XLamp LED lamps are first mounted onto metal core printed cir-

Cree’s long-term testing meth-odology provides guidance on mean L70 lifetimes for XLamp

XR-E LED lamps in a wide range of op-erating conditions. Using this method, LED luminaire designers can predict the expected mean L70 lifetime of XLamp XR-E LEDs in a specific LED system based on critical parameters.

High-power LED lamps typically do not fail catastrophically but will slowly decrease in light output over time. To characterize this gradual light loss, Cree uses both IES LM-80 compliant and other test methods. LM-80 is the indus-try standard that defines the method for testing LED lamps, arrays and modules to determine their lumen depreciation characteristics and report the results. The goal of LM-80 is to allow a reliable comparison of test results among labo-ratories by establishing uni¬form test methods.

The US Department of Energy recog-nizes LM-80 and requires it to be used for testing LED lamps in lumi¬naires that are submitted for Energy Star ap-proval.

Many high-power LED lamps do not reach L70 even after thousands of hours of testing. Therefore, Cree uses the available data to project the L70 lifetimes for the LED lamps under those operating conditions and uses

Long-Term LED Maintenance

I had the opportunity to speak with Paul Scheidt, Product Marketing Manager for Cree LED

Components. He is based in Durham, North Carolina and gave me an insight into the hitherto uncharted

territories of LED lifetime testing. LEDs have be used extensively in interior applications but now with

the emergence of their high brightness ‘big brothers’, the applications are escalating. But how much do

we know about their reliability other than the usual 50,000 hour argument that has driven many cost-of-

ownership discussions around the industry?

The luminous flux and chromaticity of the XLamp LED lamps are initially measured in an integrating sphere before the testing begins (at t=0). The LED lamp sets are then placed into the environmental chambers – with various sets of lamps being operated at various drive currents (from nominal to maxi-mum as specified in the XLamp data sheets). At regular intervals, the LED lamps are removed from the environ-mental test chambers and the luminous flux and chroma¬ticity are re-measured in an integrating sphere.

Additional Lumen Maintenance Test Configurations

Additional test conditions are also being used at Cree to study the long-term behavior of XLamp LED lamps. As with Cree’s standard long-term testing configuration, LED lamps are tested in an integrating sphere for luminous flux and chromaticity at t=0 and at regular intervals during the testing period.

Room Temperature Testing In the Room Temperature Testing

configuration, TAIR is actively controlled at 25°C. Each test set uses identical heat sinks. The TSP is passively con-

Note: Rth j-sp is the thermal resistance between the LED junction and the solder point of the LED lamp. This parameter is provided on all LED data sheets. Tem-perature, Solder Point (TSP) is the temperature of the thermal pad on the bottom of the LED lamp. Cree shows the recommended TSP for all XLamp LED lamps in each applicable Soldering & Handling document (Also called case tem¬perature (TC)).

Lens

Reflector

LED ChipSubmountBond Wire

Substate

Encapsulant

As an example, the graph shows the lumen maintenance results of XLamp XR-E white LED lamps under three dif¬ferent test conditions.

Cross-Section Diagram of a High-Power LED.


POWER ELECTRONICS

& MOTOR DRIVES

www.psim-europe.com

Co-Simulation:JMAG® & Matlab/Simulink®

AccurateCustomizableEasy to UseFast & RobustInteractive Simulation

PSIM-JMAG User ConferenceSeptember 3-4, 2009

Aix-en-Provence - Francewww.psim-europe.com/UC2009/

Reg

iste

red

trad

emar

ks:

JMA

Gof

JRI-

Sol

Inc.

,M

atla

ban

dS

imul

ink

ofM

athw

orks

Inc.

AC AnalysisAutomatic Code GenerationHarmonic AnalysisSupport Custom C CodeMagnetics ModelingMotor Drive AnalysisParametric SimulationParametric SimulationSwitch Losses Calculation

Renewable Energy Applications

Simulation SoftwarePSIM

ing. “These ICs deliver on the promise of Digital Power by providing engineers with a cost effective configurable power solution. Additionally, Exar’s Digital Power Studio™, an easy-to-use soft-ware environment, enables the creation of complex sequencing schemes and power systems, theoretically pos-sible with analog, but too expensive to implement. We are encouraged by early market adoption – clearly our custom-ers understand the value.”

Key Product FeaturesBoth devices offer a wide input volt-

age range (6.5V to 20V), and output range (0.9V to 5.0V), with a built-in


“The XRP7704 and the XRP7740 are the first 5-channel power-system ICs to combine high performance analog technology with advanced, high-speed, proprietary digital control circuitry,” said Tim Maloney, Senior Director, Market-

Exar, a company that delivers high-ly differentiated silicon, software and subsystem solutions for in-

dustrial, datacom and storage applica-tions, has now entered the digital power market with two PowerXR system-level solutions comprised of highly integrated field programmable power regulator ICs that change the design mindset for creating power solutions. These two ICs provide programmable digital power at the same low cost and robustness as analog, with many additional power management features.

Exar has recently launched the XRP7704, a 5Amp/channel regulator, and the XRP7740, a 16 Amp/chan-nel version. These power ICs integrate the best of both worlds – the low cost and flexibility of digital power man-agement and control, and the robust power capabilities of analog switching power supplies. No specialized com-ponents are required; simple off-the-shelf capacitors and inductors can be used. PowerXR products will reduce development time from weeks to hours enabling a significant time-to-market advantage. Both PowerXR devices per-mit real-time power system adjustments during design, in response to changing requirements or even ‘on-the-fly’ after field-deployment, thereby minimizing design risk and uncertainty which can delay a preferred launch date and miss a marketing window of opportunity.

Redefining the Digital Power Market

I talked with Tim Maloney, Senior Director of Marketing at Exar Corporation; about a new family of

programmable power system ICs that help designers significantly reduce development time, cost, and are

easily reconfigured for real-time design changes.

Exar introduces PowerXR high-performance system solutions

depreciation for LED lamps, the temperature of the air surround-ing the LED lamp (TA). Most high power LED lamps, including XLamp LED lamps, use silicone materials in the package. When exposed to high temperatures these silicone materials will de-grade, reducing the light that is transmitted through them from the LED chip. Just as the indus¬try has observed that higher TJ and TSP accelerates the rate of lumen depreciation, higher TAIR also accelerates the rate of lumen depreciation.

Functional Model for XLamp XR-E White Lumen Maintenance

Cree has accumulated as much as 50,000 hours of XLamp LED lamp test data under both LM-80-compliant conditions and other conditions described in earlier sections. The effects of TAIR, TJ, TSP and IF on long-term lumen maintenance have been closely studied and are well understood.

Cree has observed that the lumen maintenance character-istics of XLamp XR-E white LED lamps are different in the first 5,000 hours of test (Period A) than in the time period following 5,000 hours (Period B). A best-fit regression model for Period A was developed from 16 individual long-term data sets and includes variables such as TSP, IF, TJ and TAIR. Cree has ob-served that lumen maintenance in Period B is linear, so Period B uses a linear model that includes the same variables as Period A.

The lumen depreciation rates for both Period A and Period B for any combination of critical operating parameters (TSP, IF, TJ, TAIR) can be calculated using these models. With both Period A and Period B lumen depreciation rates known, the final L70 lifetime can be determined. This is the method used to create the lifetime prediction graphs.

Cree’s investment in this innovative research, which is briefly outlined in this interview report, has resulted in the accumulation of a great deal of useful data which could be extremely valuable to designers of LED lighting systems.

For further information visit: www.cree.com

www.cree.comXRP7740 Evaluation board efficiency curves.

http://www.psim-europe.com


integrated gate drivers for the high-current outputs and up to 6 General Purpose Input/Output (GPIO) pins. Exar’s Digital Power Studio enables design-ers to intelligently configure the power supply’s voltage setting and current thresholds, fault monitoring and re-sponse, soft start and active shut-down timing, and channel sequencing, phase shift management, and loop response, amongst other features.

The XRP7704 and XRP7740 use a digital PID (proportional, integral, differ-ential) control algorithm which performs full-digital loop control at switching frequencies to 1.5MHz. The controller supports I2C interface commands for control, configuration, and management of the power supply.

Exar’s PowerXR products fuse digital power-supply control and monitor-ing technology with high-performance analog circuitry in a new generation of digital power-management products that enable system architects to create products with advanced, integrated, switching power supplies that signifi-cantly reduce wasted energy compared to more commonly-used linear power supply regulators.

Availability Samples of the XRP7704 and

XRP7740 are available now and prod-ucts are offered in a RoHS compliant, “green”/ halogen-free, 40-pin QFN package operating within the -40ºC to +85ºC range.

Exar’s comprehensive experience of end-user markets along with the com-pany’s underlying analog, mixed signal and digital technology has enabled truly innovative solutions to meet the needs of the evolving connected world. With a product portfolio of power management and interface components, communi-cations products, storage optimization solutions, network security and applied service processors, Exar utilizes its worldwide locations and staff to provide fast and effective support for its cus-tomers.

For more information, visit: http://www.exar.com

Low-Dropout (LDO) for standby power, power sequencing capability, and integrated gate drivers. The 5-output, XRP7704 and XRP7740 digital power system controller ICs each contain four digital pulse width modulator (DPWM) controlled power supplies with an effec-tive 12-bit resolution.

www.exar.com

Both devices also contain an inte-grated LDO regulator that provides a fifth voltage supply, which can also be employed as a standby-voltage source. Both chips are fully configurable via an I2C interface for monitoring, control and management of DC/DC point-of-load power conversion. The devices contain

XRP 77XX Block diagram.

COVER STORY

21www.powersystemsdesign.com

Latest Generation IGBT Gate Drivers

discrete components and off-the-shelf ICs.

The SCALE-2 ASIC platform has been specifically developed to address this problem. Custom integration shows its strength in the new IGBT gate driver cores 2SC0108T and 2SC0435T.

Reinforced insulation and high dv/dt stability

The operating range of today’s 1700V IGBT modules can be fully exploited with these drivers, as clearance and creepage are coordinated to EN50178 under pollution degree 2.

Moreover, both the signal barrier and the DC/DC transformer between the pri-mary and secondary sides provide rein-forced insulation to EN50178, protection class 2. The combined coupling capaci-tance for DC/DC and pulse transform-ers between the primary and secondary sides is less than 18pF. Thanks to these outstanding performance figures, drive systems built with these cores withstand dv/dt transients as high as 75kV/μs.

Power and speed The SCALE-2 ASIC chipset used in

the 2SC0108T and 2SC0435T intro-duces a delay of less than 100ns in the

For high-performance cost-effective solutions

With the uncompromising integration of IGBT drive and protection functions as well as DC/DC

control and signal transition, a new generation of reliable and cost-effective IGBT driver solutions is born.

The resulting smart IGBT gate driver construction helps to economize price-sensitive external components.

The total system cost is thus significantly reduced. The IGBT gate driver cores 2SC0108T and 2SC0435T

constitute a new standard in cost, functionality, reliability, and user-friendly design.

By Michael Hornkamp, Sascha Pawel and Olivier Garcia, CT-Concept Technologie AG, Switzerland

It is well known that the number of individual components interacting in a system is important for its overall reli-

ability. Fewer components mean higher reliability in non-redundant systems, as long as component failure rates are comparable. This principle is success-fully applied to the monolithic integra-tion of electronic functions into ICs for almost all aspects of everyday life, sci-

ence and industry. In power electronics, however, design cycles are considerably slower and designers are more cautious in adopting technical advances. This risk-minimizing attitude has brought to-day’s power systems to a very high level of reliability. However, as the number of electronic functions increases, the reli-ability limitation is becoming more and more severe due to the large number of

Figure 1: SCALE-2 drive and protection principle.

http://www.electrocube.com

COVER STORY


COVER STORY


module has its own driver with full gal-vanic isolation, thus blocking any offset currents. Easy scaling of the system output power by adding more paral-lel legs is another plus. The IGBTs are shown in Fig.3. When IGBT driver cores are paralleled, an electrically insulated auxiliary emitter avoids offset currents in them. Higher accuracy and uniform IGBT switching are simpler and more predictable. Thanks to the stable and fast driver core, a running time of less than 100ns and a jitter of 4ns, synchro-nization of the IGBT gate driver can be obviated. Offset currents in the auxiliary emitter due to an asymmetrical AC star-point connection are inhibited. The low cost of these driver cores means that there is no economic limitation to this system topology.

Low cost half-bridge driverThe cost-saving capability of ASIC

integration is fully exploited towards the lower end of the driver power spectrum with the 2SC0108T low cost driver. The PCB is assembled with as few as 23 components, including transformer and ICs, to form a complete IGBT and MOS-FET driver core with most of the func-tions familiar from the SCALE-2 plat-form. The output power rating is 1W per channel at up to 8A maximum output current within an ambient temperature range of –20°C and 85°C.

Where there is a need for more power and current, the 2SC0435T provides the full set of SCALE-2 functions, highest power density and high output current together with advanced active clamping (AAC). The ambient temperature range of the 2SC0435T is –40°C to +85°C.

IGBT gate driver for 10USD per channel

Available now, the pricing of the 2SC0108T is very competitive thanks to the low production cost of the SCALE-2 platform. At quantities of 10,000 items, the driver will be priced at US$20 (US$10-per channel). It thus compares very favorably with discrete solutions for bidirectional signal transmission, isolated DC-to-DC power and gate drive output. The benefits of high reliability and tried and tested SCALE technology are also included.

The gate drive ASIC evaluates the col-lector voltage feedback and adapts the output strength of the gate drive to the situation. The closed-loop control main-tains an optimized IGBT gate voltage to combine low switching losses and inherently safe IGBT protection.

The switching and protection behavior

can be easily adjusted by changing the external components Racl and Cacl in Fig. 2. More advanced options such as di/dt feedback and partial du/dt feed-back are also supported.

Once the active clamping behavior has been set, the AAC feedback loop automatically distinguishes between normal switching, where low losses are the primary focus, and short circuit turn-off, where the main emphasis is placed on keeping control of transient over-voltages and maintaining a soft turn-off with a limited current change rate di/dt. Figure 3 shows measurements of both normal switching and short circuit turn-off for the same AAC configuration. The soft turn-off of the short circuit can be clearly seen, as can the low-loss normal switching.

Separate gate turn-on/off outputs The 2SC0108T and 2CS0435 feature

two separate gate drive outputs for turning-on (GH) and turning-off (GL), as shown in Fig.1. Each output is con-nected to a dedicated turn-on and turn-off gate resistor respectively. A parallel diode is no longer needed in the gate drive path, thus reducing costs and increasing reliability.

Flexible design with driver coresDriver cores provide all necessary

drive and protection functions with no limitation in flexibility. Any inverter de-sign and IGBT module can be accom-modated with little effort. It is easy to set different IGBT switching speeds, short-circuit thresholds or other parameters simply by adjusting a few external de-vices. A whole range of inverter designs can be served by a single SCALE-2 driver core type. Mechanical construc-tion also benefits from the high degree of freedom offered by these small and powerful cores.

A dedicated half-bridge mode allows the implementation of a hardware-set dead time between the two half-bridge channels. The dead time is user pro-grammable with a single external resis-tor for both channels.

Paralleling of IGBTs Both the 2SC0108T and the

2SC0435T are particularly well suited for direct paralleling of IGBTs. This simple and very efficient paralleling topology is made possible by the extreme delay time repeatability in SCALE-2 driver cores. Figure 3 shows the system setup for direct paralleling.

Two, three or more half-bridge legs are independently driven by a driver core. All driver cores receive the same input signal. The individual driver output signals on the secondary side show less than +/-4ns timing difference. This low value is negligible for the system.

A major benefit is that there are no longer any offset currents between the parallel inverter legs, especially if the system is connected in an asymmetri-cal star-point configuration. Each IGBT

Figure 4: Direct paralleling of IGBT and gate driver core.

www.IGBT-Driver.com

turn- on and turn-off signal paths. This delay time is delivered at a superior repeatability of +/-4ns including jitter and ageing. A comparison with other technologies such as opto-couplers shows that the propagation delay time is typically as high as 500ns with a mis-match of several 100ns. Uneven ageing of separate drive channels is a com-mon problem for fiber optics and opto-coupler systems. In contrast, SCALE-2 drivers maintain symmetrical switching of the driver channels. The clear user advantage is constant IGBT losses over the lifetime of an inverter.

Switching frequencies up to 100kHz

are allowed with no derating for an ambient temperature up to +85°C. In the same environment, SCALE-2 IGBT driver cores provide a gate drive power of 1W and 4W per channel respectively (2SC0108T/2SC0435T). The associated gate drive current is 8A/35A. Any IGBT module ranging from 600V to 1700V in this voltage class can be easily used at an inverter output current of up to 2400A.

Vce short circuit detection using no high voltage diodes

A new collector sense circuit featur-ing an externally configurable resistor network replaces the widespread high voltage sensing through diodes (Fig. 1). Its main advantage is precise, direct voltage measurement allowing any increase in IGBT saturation voltage to be detected. Direct Vce sensing is no longer influenced by parasitic capaci-tances of the high voltage diodes and their pronounced temperature depen-dence. Also, the cost of standard SMD 1206 resistors for the resistive divider is dramatically lower than for high voltage diodes. This benefit can be reaped im-

mediately with the new SCALE-2 driver cores.

Regulated gate drive voltage

SCALE-2 drivers provide a gate volt-age swing of +15V/–10V in IGBT mode. The turn-on voltage is tightly regulated to maintain a stable 15V regardless of the output power level. This feature is particularly appealing during short cir-cuit of the IGBT. The high dv/dt values occurring in this failure mode inject considerable amounts of charge into the gate node (Miller feedback). This feedback causes the gate voltage to rise during a short circuit, resulting in excessively high levels of short-circuit current. This is a dangerous situation for the IGBT module. SCALE-2 drivers use a Schottky diode clamp to limit the gate voltage to safe values (see Fig. 1). The stable 15V supply easily absorbs the Miller feedback charge and safe opera-tion of the IGBT is maintained.

Under voltage lock-outBoth primary and secondary gate

driver supplies are constantly moni-tored. If the supply voltage reaches a level of 12.5V or lower, a failure signal is generated and issued by the respec-tive failure output SOx. During a primary

side under-voltage event, the failure is shown as long as the voltage stays below the under-voltage level. The pri-mary side ASIC then blocks all incoming PWM signals. During a secondary side under-voltage event, a signal pat-tern is transmitted from the secondary side ASIC to the primary side ASIC to provide an external error feedback on the respective primary side SOx output. This status is held for 10ms (adjustable). As long as the secondary side ASIC detects an under-voltage situation, the affected gate drive channel is disabled.

Advanced active clamping (AAC)

In a standard active clamping circuit, the IGBT collector voltage is fed back to the IGBT gate via Transil diodes. As soon as the collector-emitter voltage exceeds the threshold voltage of the Transil chain, a current flows to the gate. This current is partly used to turn the IGBT 2SC0108T on again. The partly conductive IGBT effectively clamps tran-sient over-voltage peaks during turn-off.

The main drawback of this simple ac-tive clamping scheme is its speed and efficiency. The Transil chain tries to turn the IGBT on while the gate driver output stage simultaneously tries to keep the IGBT turned off. It takes some time and a lot of dissipated power for the Transil to overpower the output stage.

The SCALE-2 advanced active clamping (AAC) feature is used in the 2SC0435T to overcome these limita-tions. Figure 2 shows the operating prin-ciple of AAC. The standard feedback path through the Transil diodes to the IGBT gate starts at the IGBT collector on the right. A second feedback path senses the collector voltage and sup-plies this signal to the SCALE-2 ASIC.

Figure 2: SCALE-2 advanced active clamping (AAC).

Figure 3: AAC performance.


Power Supplies

24

Power Supplies

When regulators in the US and Europe laid out their plans for lowering no-load power dissipation in

offline chargers and similar equipment, the last thing that most power-supply designers expected was

for five major mobile phone makers to combine forces and demand a further five-fold reduction - leaving

designers with a target figure of 30 mW or less that is very hard to meet.

auxiliary winding and internally regu-lates VDD. Overall, the power necessary to start up and keep IC1 switching at around 1 kHz under no-load conditions is approximately 12 – 14mW. At around 10mW, dummy load resistor Rout becomes the next dominant power-consuming term. Its value balances the primary power level and ensures output voltage stability when no external load is present. Parasitic losses elsewhere, such as in the transformer, amount to around 2mW for a total no-load con-sumption of about 25mW.

Transient response from no-load conditions Because it examines the feedback waveform cycle-by-cycle, CamSemi’s tangent detection technique speeds transient response even at low switching frequencies. Figure 2 shows an example of the sequence of events that follows a 500mA no-load to full-load step.

The bottom trace is the load step that creates a temporary output-voltage dip in the voltage trace above it. The first, sharp part of this dip is due to IR losses that depend upon external fac-tors such as cable gauge. The second, slower negative ramp shows the droop due to the output capacitor discharg-ing. Meantime, the controller senses the voltage drop and immediately ramps up its switching frequency, as the top trace shows. Here, the effect is to constrain the dip to less than 0.8V, with the out-put voltage stabilising within around 4 msec, which is well within the 10 msec maximum recovery time for the AD-2971 reference design.

with 105oC ratings are typically accept-able. Low-value resistor Rin provides inrush current protection and—depend-ing upon safety-agency requirements—may be a fusible component. The circuit’s quasi-resonant switching technique sufficiently reduces conducted EMI that L1 can be low in inductance and dc resistance values.

The high-voltage rail supplies T1’s primary current as well as the start-up current for IC1 via resistor Rht. Because IC1 is built using a 3.3V CMOS process that requires a maximum 7µA during start up, Rht can be a very high value; even so, start-up time is less than one second. During operation, IC1 draws a maximum of 2mA from the primary-side

output power. In some cases, the circuit’s output voltage falls to zero before beginning a lengthy recovery sequence. This happens because many controller manufacturers resort to very low switching frequencies to achieve these stringent no load power require-ments which results in the controller being unable to respond quickly to load level changes. Designers have to ensure that a circuit capable of meeting the 30 mW challenge can also start up quickly and recover well from zero- to full-load conditions.

No Load Power without Compromise

Achieving the new 30mW performance requirement

By Neil Massey, Product Line Manager, AC-DC Controllers, CamSemi, Cambridge, UK

As product designers are keenly aware, regulators in the US and EC have been busy proposing

minimum efficiency requirements for offline power supplies. The first round of regulatory requirements tackles external power supplies, such as those that re-charge mobile devices. At around 70%, the minimum acceptable efficiency dur-ing device operation is not hard to meet. Similarly, today’s 250 – 300mW require-ments for no-load power dissipation are also easy to accommodate—as is the EC’s 150mW specification that comes into force in 2011. But with a 30mW no-load target for truly “green” performance, the new five-star rating system that LG Electron-ics, Motorola, Nokia, Samsung, and Sony Erics-son have agreed entirely changes this landscape, as table 1 shows.

Meeting such a small no-load target is not trivial. For instance, one effect of minimising no-load power consumption can be to create large voltage drops and lengthy recovery times when transitioning from no-load mode to full

Minimising no-load power consumption

Considering a circuit that meets the 30 mW no-load challenge illustrates some of the problems and solutions. This design employs CamSemi’s C2161PX2 or C2162PX2 controller for maximum output power levels of 4W and 8W, respectively. It exploits a primary-side-sensing flyback topology to construct a universal-input converter that achieves tight constant-current/constant-voltage operation without needing the opto-isolator, voltage reference, and associ-ated components that typify secondary-sensing flyback circuits (figure 1).

A full-wave rectifier followed by an LC filter smoothes and attenuates emis-sions for the high-voltage dc rail. For optimal performance, capacitors Cin1 and Cin2 should be low-impedance, low-leakage types, but commodity parts

Table 1: Star-rating system no-load tar-gets as set by five major mobile phone makers.

Figure 1: Basic schematic for a primary-side sensing <30 mW no-load power consumption.

Figure 2: Cycle-by-cycle sampling enables rapid transient response even from no-load states.

Figure 3: Feedback and current wave-forms and sampling points.

Figure 4: A proprietary tangent detection technique improves voltage and current regulation accuracy.


Power Supplies

26

voltage Ip that current-sense resistor Rcs generates. A short leading-edge blanking period immediately follows Q1 turning on to discard the spike that ap-pears as Ip begins to ramp up. The secondary current calcula-tion then becomes:

PCHARGE

DCHARGEOUT I

tt

KI =

where K is the transformer’s turns ratio.

Again, the patented tangent-detection technique is crucial in achieving this best-in-class current regulation performance. Figure 5 shows the typical CC/CV charger output character-istic that the C2161/2162PX2 achieves while constraining output voltage ripple to <200 mV peak-to-peak without need-ing any secondary-side sensing components:

Another desirable attribute that the tangent detection tech-nique makes possible is close-to-zero voltage switching for the transformer’s primary current, which increases efficiency and minimises the stress on power-switch components while also reducing emissions. The worst-case margin for conducted

EMI that CamSemi’s AD-2971 reference design achieves for a universal-input 4.8 W charger is >6 dB below EN 55022 limits, while its four-point average of 75% for constant-voltage mode easily exceeds Energy Star V2 and EC Code-of-Conduct V4 requirements, as Figure 6 shows.

ConclusionDesigning a five star-rated mobile

phone charger, with no load power consumption of less than 30mW can be a challenge because so many of the techniques widely used can affect key performance parameters such as transient response. CamSemi’s C2160 controller family achieves the require-ments with the best balance in all areas - without performance compromise.

Achieving tight CC/CV regulationWhile primary-side sensing circuits

are not new, the ±5% voltage and current regulation that this design achieves far exceeds the typical ±10 - 15% performance of competing de-signs. The key to its performance lies with a proprietary, patented method of measuring the circuit’s output volt-age and current levels. The controller responds to output load changes by adjusting the peak current through the primary switch Q1 and the switching frequency. The minimum switching frequency is predominantly a prod-uct of the load that the controller and dummy load present under no-load conditions, while Cosc and Rosc determine sets maximum full-load frequency. For the C2161/2162PX2, this maximum lies within the range 36 – 66kHz. The transformer’s turns ratio establishes the maximum duty cycle, which for a universal offline design is typically set to 50% at the minimum high-voltage rail value. Power conver-sion always occurs in discontinuous conduction mode As shown in figure 1, the controller switches T1 by driv-ing Q1’s emitter in cascode mode, efficiently driving a low-cost bipo-lar transistor rather than a relatively expensive MOSFET. Slew-rate limiting on the controller’s SD pin minimises conducted and radiated emissions. Resistors Rfb1 and Rfb2 scale the waveform that develops across T1’s primary sense winding, while Cfb1 provides dc blocking that allows the con-troller to centre this feedback waveform between the power s upply rail voltages. This allows the chip to examine the entire waveform rather than just the positive-going portion that competing schemes typically sample, as figure 3 shows.

In the figure, FB is the waveform at the chip’s feedback pin while Ip and Is represent the current waveforms that develop across the primary-side current-sense resistor, Rcs. It’s only possible to observe the output voltage during the flyback time tDCHARGE, which for best accuracy requires sampling at the knee point when the current falls to zero, as the Is waveform confirms.

Precisely determining the knee point is critically important. CamSemi’s propri-etary tangent-detection technique uses

a combination of analogue, high-speed sampling, and signal-processing tech-niques to assess the waveform’s rate-of-change (dV/dT) as it transitions from the tDCHARGE state to fall to the zero-crossing point. As figure 4A shows, the slope during tDCHARGE is a function of output current and circuit resistance for a given power rating, while the period in figure 4B between the knee and zero-crossing points is one-quarter of a sinewave whose frequency corresponds to the transformer’s resonant frequency. Selecting a reference dV/dT between these two known slopes establishes the sample point (figure 4C).

Because the FB waveform directly provides the tCHARGE measurement and tDCHARGE is accurately known, it is possible to determine the output cur-rent level by sensing and averaging the

www.camsemi.com

Figure 5: The C2161PX2 and C2162PX2 achieve ±5% voltage and current regulation with no second-ary-side sensing components.

Figure 6 Four-point efficiency measurements for 4.8W charger easily exceed regulatory requirements.

Koenigsegg Automotive AG – Photo by Stuart Collins

Powering Freight & Transportation

Powering Freight & Transportation

Photo Courtesy of Horizon Logistics LLC


Special Report – Powering Freight & Transportation Special Report – Powering Freight & Transportation

of 0.6(VN) for 100ms and overvoltage surges of 1.4(VN) for one second that may occur during startup.

Fast Transient Specification Equipment must be able to withstand

a direct transient of 1800V lasting 50µs. The impedance of the transient source is specified as 100Ω, with transient energy around 100mJ. To prevent damage to the DC-DC converter, the module must be connected to a suppression device, such as a transient voltage suppres-sor (TVS), directly across the input. A TVS able to withstand up to 1.5J would be suitable. The TVS clamping volt-age must be selected so that it will not exceed the high line voltage or transient limits of the converter.

RIA12 Surge Protection The RIA12 standard specifies that

equipment must withstand a 20ms over-voltage surge 3.5 times greater than the nominal voltage. To meet this standard,

Designing for Railway Applications

Modular DC-DC convertersDesigning equipment for railway applications requires a specialized approach and a thorough knowledge

of the stringent regulations and standards laid down for safety of operation and reliability in service.

By Marco Panizza, European Manager, Application Engineering, Vicor, Germany

In Europe, electronic equipment in rail-way applications is governed by two international standards; the docu-

ment IEC571, “Electronic Equipment on Rail Vehicles,” also known as European Norm EN50155, “Electronic Equipment Used on Rolling Stock Equipment” and in the UK, the standard that applies is RIA12, “General Specification for Protection of Traction and Rolling Stock Equipment from Transients and Surges in DC Control Systems”, developed by the Railway Industries Association (RIA). These two standards are similar in most respects, but the RIA12 standard also requires a specific surge withstanding capability.

Vicor modular DC-DC converters can be used in designs to comply with these standards and facilitate the design of complex power supplies.

Electrical Requirements for EN50155 (IEC571)

Operating Voltage Ranges Table 1 lists the nominal input voltag-

es (VN) provided by power sources used for railway applications, and shows that with few exceptions, standard Vicor DC-DC converters cover all of the specified input ranges and transients.

Equipment powered directly from bat-teries with no voltage stabilizing device must function with input voltages that range from 0.7(VN) to 1.25(VN) during normal operation. The equipment must also withstand input voltage drops

a suppression circuit must be used with the DC-DC converter.

For each of the standard input volt-ages, the maximum input surges are as follows:

Since the source impedance of this

Table 1: Input Specifications for EN50155 vs. Vicor DC-DC Converters.

VN 3.5 (VN )

24V 84V

36V 126V

48V 168V

72V 252V

96V 336V

110V 385V

The diodes D2 and D3 can have the same breakdown voltage, in which case the output (clamped) voltage will equal D3 minus the gate source voltage.

The RIA12 standard prescribes an input voltage variation of up to 1.5(VN) for one second. For converters with an input voltage range that allows the mod-ule to withstand this, the breakdown voltage for D2 and D3 must be greater than 1.5(VN) to prevent the protection circuit from operating under this condi-tion, but less than the converter’s high line voltage.

The diode D1 clamps high voltage spikes that have low energy but ampli-tudes that may exceed 1000V. There-fore, D1 must be a TVS with a clamping voltage higher than 3.5(VN) to prevent damage by this transient.

Test ResultsThis circuit has been tested in the lab

for 48V and 110V inputs—the two most common in railway applications—sup-plying a 150W DC-DC converter. The table below lists the components used.

In general, these circuits prove effec-

voltage ranges. Only the main MOSFET Q1 and the clamping diodes D1, D2 and D3 are selected according to nominal operating voltage and power. To select Q1 it is necessary to consider the values of 3.5(VN) and the operating current (IN ) as well as the SOA limits for a single 20ms pulse. The RDS(on) must also be taken into account, since the power dis-sipation under normal operating condi-tions is calculated as:

P = RDS(on) • IN2

overvoltage pulse is 0.2Ω, a common clamping device, such as a TVS, will not provide a suitable solution because the energy dissipated would be:

E = 3.5(VN) – VZener · VN · 0.02 0.2

Instead, an active circuit as shown in Figure 1 is necessary.

Component SelectionThe components in the Figure 1 circuit

remain essentially the same for all input

Figure 1: In this circuit, D1 clamps the fast, high voltage spikes, while the active part formed by D2, D3 and Q1, Q2 limits the surge. During normal operation, Q1 is kept in full conduction by the charge pump circuit made up of the 1N4148 diodes and the 470pF capacitor. The Gate Out signal from the converter drives the charge pump. When overvoltage occurs, the diode D2 conducts, with cur-rent limited by R1, and turns Q2 on. With Q2 in conduction, the gate voltage of Q1 is held to the clamping voltage of D3. The output voltage equals the clamping voltage of D3 minus the gate source voltage of Q1. When overvoltage ceases, the system resumes normal operation.

Converter 48VN 110VN

D1 1.5KE 200A 1.5KE 440A

D2 68V /0.5W 160V /0.5W

D3 68V /0.5W 160V /0.5W

Q1 IR FP 250 IR FP 450

Figure 2: For a 48V input, a 20ms input transient with out-put voltage clamped to the converter’s high line limit.

Figure 3: For a 110V input, a 20ms input transient with output voltage clamped to the converter’s high line limit.


Special Report – Powering Freight & Transportation

shock of 50m/s2 amplitude. Because they are fully encapsulated in epoxy resin, Vicor modules can easily sustain these mechanical stresses. However, close attention must be paid to the PCB design and wiring to prevent relative movements that can stress the modules’ pins.

ConclusionThe design of transport systems car-

ries a heavy responsibility and attention to the detailed electrical and physical specifications. Vicor’s in-depth under-standing and experience in this special-ized field can help simplify significantly the task for the designer enabling a faster and safer design cycle.

In some railway applications, the equipment is bulkhead mounted on the wagon walls. This setup can take ad-vantage of the Vicor converters’ planar baseplate and use the cabinet as a large heat sink.

Vibration and shockEN50155 specifies that electronic

equipment mounted on boards and boxes fixed to the vehicle frame must be able to withstand vibration on all three axes at the levels below:

• Frequency range: 5–150Hz

• Cross-over frequency: 8.2Hz

• Displacement amplitude (below cross-over frequency): 7.5mm

• Acceleration amplitude (above cross-over frequency): 20m/s2

The equipment also must be able to withstand a 50ms, semi-sinusoidal

tive in suppressing high energy voltage surges and can be adapted to suit appli-cations besides those needing to meet RIA12 requirements.

Physical Requirements

Operating temperature rangeOperating temperatures are divided

into four classes according to the sever-ity of the environment, as shown in Table 2. When designing the power supply, it is necessary to consider overtempera-ture during start-up, as indicated in the third column.

www.vicor.com

Table 2: Operating Temperature Ranges.

OperatingTemp.Classes

Internal Cubicle Temp

Internal Cubicle Overtemp(10 min)

Air Temp Around PCB

T1 -25/+55ºC +15ºC -25/+70ºC

T2 -40/+55ºC +15ºC -40/+70ºC

T3 -25/+70ºC +15ºC -25/+85ºC

TX -40/+70ºC +15ºC -40/+85ºC

HybridPACK™ 2 - Compact Power for Your Electric Drive Train. Based on the long time experience in the development of IGBT power modules and intense research efforts of new material combinations and assembly technologies, Infineon has developed – dedicated for automotive applications – this HybridPACK™ 2 power module belonging to the HybridPACK™ family. With its pin fin base plate for

Fairchild Semiconductor

direct water cooling Infineon HybridPACK™ 2 is designed to fulfill the requirements of your electric drive train application with power ratings of up to 80kW.

www.infineon.com/cms/en/product/channel.html?channel=db3a3043136fc1dd011370e812b7043a

Minimize Power Consumption in Mobile Designs

Fairchild’s programmable LED blinker, the FAN5646, eliminates the need for a microprocessor to be in continuous operation

Infineon

mode. This results in low power consumption during operation and shutdown mode.

www.fairchildsemi.com/pf/FA/FAN5646.html

LEM

Advertisement

ITW Paktron

Features “non-shorting” operation and does not crack like large ceramic chip capacitors under temperature extremes or high vibration. There are no DC or AC voltage coefficient issues with polymer film capacitors. Capacitance values range from 0.33µF to 20µF and voltage ratings are 50 to 500 VDC. Lead time is stock or four to six weeks.

www.paktron.com

Non-Polarized Polymer Film Capacitors (CS Series) Designed for Mission Critical Applications ITW Paktron’s Multilayer Polymer (MLP) Film Capacitors (Type CS Series) feature ultra-low ESR and high ripple current capability and are designed for high frequency filtering and EMI/RFI suppression in power conversion applications. Provides mechanical and electrical stability, compared to multilayer ceramic capacitors.

International Rectifier

New 150V and 200V MOSFETsThe new 150V and 200V MOSFETs from

International Rectifier are targeted for hard switched and/or high frequency circuits such as SMPS, UPS and Inverters. These parts offer very low gate charge (Qg) and are therefore optimized for fast switching circuits where

switching losses are critical. These parts offer the best price and performance combination in the market and are available in TO-220, D2PAK, TO-262, DPAK and IPAK.

www.irf.com

Visit LEM at PCIM 2009 in Hall 12-402. This year, LEM will be highlighting a range of products, including the CAS, CASR and CKSR family of current transducers. They are suitable for industrial applications such as variable speed drives, UPS, SMPS, air conditioning, home appliances, solar inverters and also precision systems such as servo drives for wafer production and high-accuracy robots.

LEM is a leading manufacturer of current and

voltage transducers used in a broad range of industrial applications, including variable speed drives for motors and power supplies, AC/DC converters, UPS systems for computers as well as in new innovative energy applications, such as micro-turbines, wind and solar power generation.

For further information please go to

www.lem.com

PowerPackPowerPack Power Systems DesignPower Systems Design PowerPack Power Systems DesignPower Systems Design

PowerPack and ePowerPack is a new advertising program in print and online designed to promote your company’s new products, seminars, and announcements while driving traffi c to your company’s website.

Advertisers receive a 100 word listing, plus product photo and url link in print to 20, 357+ subscribers of Power Systems Design’s magazine and online through ePowerPack e-newsletter which is delivered every month to an audience of 24,000 power elec-tronic engineers in Pan Europe and 32,000 power design engineers in North America. To participate contact: [email protected]

A Powerful Combination

PowerPack

through ePowerPack e-newsletter which is delivered

A Powerful A Powerful Combination

PowefulComboAd.indd 1 3/13/08 5:58:47 PM

http://www.fairchildsemi.com/pf/Fan/FAN5646.html

http://www.irf.com

http://www.infineon.com/cms/en/product/chennel.html?channel=db3a303043136fc1dd011370e812b7043a

http://www.lem.com

http://www.paktron.com




packed circuits, simple air circulation of-ten has been found to be insufficient to cool the circuit components adequately.

Attaching components directly to a thermal dissipation member such as a part of the chassis or enclosure, or a dedicated heatsink, can improve cool-ing beyond that achievable using a fan. However, air inclusion at the interface between the two surfaces impairs thermal efficiency, since air has very low thermal conductivity of 0.024W/m’K.

Thermal Gap Filling Revolution and Evolution

Transportation, computing and consumer electronics applications

Modern system design is moving away from ensuring air circulation to maintain convective cooling,

toward diligently filling air gaps between components and the chassis or enclosure using soft, gap-filler

materials displaying high thermal conductivity.

By Eoin O'Riordan, Marketing & Technology Manager, Chomerics Division

Europe - Parker Hannifin Ltd

Miniaturisation and increased functionality have dominated the development of modern

electronic products, and have been criti-cal to satisfying customer demand in markets such as personal computing, consumer electronics, aviation and the automotive sector. Key enablers include successive shrinking of CMOS process-es, which not only increases functional density but also increases the heat dissipation of devices such as proces-sors, FPGAs and memories. Advances in power semiconductors, also, are enabling higher currents within a smaller component footprint, leading to greater power density as well as increased dis-sipation.

System designers face an increasingly difficult thermal challenge. Elevated operating temperatures impair reliability, but effective cooling of multi-layer PCBs carrying closely spaced components and packaged within tight enclosure dimensions is frequently impossible by convection alone.

The cooling fan has traditionally been a mainstay of thermal engineering. How-ever, demands to reduce size, power and audible noise are steering system design toward fanless operation. The fan also adds a potential point of failure. Moreover, in high-power or densely

Where the heatsink is clamped or bolted directly to the body of the component, microscopic air pockets exist at the in-terface since both are commercial-grade surfaces displaying imperfect flatness and smoothness. Thermal greases and, more recently, phase-change thermal materials are ideal media to eliminate this interstitial air.

Larger Gaps to FillAs modern system design has come

to require a more rigorous approach

Figure 1: Chomerics HCS10 material.

Chomerics HCS10, for example (figure 1), which has thermal conductivity of 1W/m-K, has the highest conformability among gap fillers available as sheet or pads, and compresses by 73% under pressure of 50spi. Other materials of this type offer a different combination of conductivity and compressive modu-lus, as well as formulations that can be moulded to form and retain complex shapes and provide vibration damping as well as high thermal performance. Low deflection force, combined with the vibration-damping properties of some types, makes these materials suitable for applications such as power mod-ules, telecom equipment, LED lighting systems, memory modules, or handheld devices. In particular, those applications where the sizes of air gaps cannot be predicted tend to benefit from low-mod-

sive modulus and can be formulated to operate across a temperature range up to –50°C to +200°C. Materials with low compressive modulus are able to conform to the contours of the mating surfaces without placing high stresses on components and solder joints. Hence ideal qualities for a gap-filler TIM are low compressive modulus as well as high thermal conductivity. Since con-ductivity is usually determined by filler content, higher conductivity is tradition-ally achieved alongside an increase in compressive modulus.

Gap-Filler Sheets and PadsThe latest gap filler formulas provide

a number of choices allowing system designers to prioritise conformability or thermal conductivity, or an optimal combination of both properties.

to thermal management, demand has grown for a wider variety of thermal interface materials (TIMs) allowing engineers to make better use of heat-dissipating surfaces. TIMs designed to fill larger air gaps, from around 1mm to 5mm or more are available as sheets, pads, or die-cut to custom dimensions and shapes. These can be used to fill gaps between heat-generating com-ponents and the chassis or the lid of the enclosure. In the past, such gaps may have been left unfilled so as to al-low air to circulate. In modern designs, best practice emphasises creating an efficient thermal coupling between the component and the nearest heat dis-sipating surface.

A gap filler, like a thermal grease or phase-change material, comprises a binder or carrier loaded with thermally conductive particles. These may be ce-ramic particles, or a metallic oxide such as aluminium oxide, magnesium oxide, zinc oxide, boron nitride or aluminium nitride. Factors such as the conductivity of the particle material and the ratio of particles to binder determine the overall thermal conductivity of the TIM. In general, increasing the filler content, or choosing a filler material of higher ther-mal conductivity, increases the thermal conductivity of the TIM. Gap fillers with many different levels of thermal perfor-mance are available, and thermal con-ductivity ranges from below 1 W/m-K up to around 6 W/m-K.

Advances in polymer-based binder materials have been critical to the devel-opment of gap fillers able to solve the full range of thermal management chal-lenges now facing engineers. Silicone elastomers benefit from a low compres-

Figure 3: Chomerics T630 Therm-A-Gap™ has thermal conductivity 0.7W/m-K and delivers the convenience of a pre-cured form-in-place one-component material.

Figure 2: Chomerics 976 is an example of a gap filler that combines high thermal conductivity, but also places low stresses on components when compressed.



ulus advantages of traditional thermal greases while eliminating pump-put and dry-out issues.

A Note of CautionGap fillers provide a convenient and

high-performing solution for engineers seeking to optimise thermal manage-ment. However, although they have thermal conductivity many times better than that of air, they are significantly less thermally conductive than metals. The thermal conductivity of aluminium, for example, is 250W/m-K compared to 1-6W/m-K for most gap fillers. Hence, best design practice is to ensure that components are positioned close to heat dissipating surfaces to minimise the thickness of TIM necessary for ef-fective thermal coupling.

Hence, although performance is increasing with each new generation of materials, gap fillers deliver their best results when used to optimise a design that is already thermally sound.

attaching a number of components such as in a power-MOSFET array to a com-mon heatsink, for example in power-conversion equipment, motor drivers or electronic control units (ECUs). The binder may be a two-component room-temperature-vulcanising silicone mate-rial. More recently, single-component thermal gels have been developed, which can be dispensed directly without mixing and are pre-cured thereby also eliminating the cure cycle normally nec-essary with a two-part material. Little to no compressive force is required during assembly, which minimises the mechan-ical stress placed on components.

Chomerics T630 Therm-A-Gap™, which has thermal conductivity 0.7W/m-K, is an example of a gap filler delivering the convenience and ease of use of a pre-cured form-in-place one-component material (figure 3).

A variety of thermal gels are available, for gap-filling applications as well as direct replacement of thermal greases. Their cross-linked molecular structure delivers a TIM displaying the low-mod-

ulus gap fillers, which place low stress on components when compressed.

Gap filler sheets of higher thermal conductivity tend to have a higher com-pressive modulus. Chomerics 976 is an example of a gap filler that combines high thermal conductivity of 6.5W/m-K but also places low stresses on com-ponents when compressed (figure 2). It is deflected by 45% at 50psi, which is softer than materials of comparable thermal conductivity. Among the wide variety of options open to engineers, gap fillers with pre-applied pressure-sensitive adhesive can help to speed up assembly. In addition, low-outgassing and silicone-free formulations are avail-able for silicone-sensitive applications such as aerospace equipment, optical electronics and hard-disk drives.

Form-in-Place Gap FillersAs an alternative to gap fillers featur-

ing silicone elastomer carriers, form-in-place compounds can be used where the distance between the component surface and the adjacent cold surface is variable. Typical applications include

www.parker.com/chomerics

The Winner is: Company: Fairchild Semiconductor, Article: “Solar Power Shines”, Author: Alfred Hesener.

The four finalists (in alphabetical order by company) are: Company: Coilcraft, Article: “Designing for Efficiency at the Component Level”, Author: Len Crane Company: Linear Tech-nology, Article: “Solving Current Source Design Challenges”, Author: Robert Dobkin Company: Microsemi, Article: “SiC Impacts ‘Greening’ of Power”, Authors: Philip C. Zuk & Bruce Odekirk Company: Philips Lumileds, Article: “Avoiding Current Spikes with LEDs”, Author: Pat Goodman.

Educational Donation. A significant component of the 2009 GreenPower Leadership Awardsprogram is an educational donation, given to the European Engineering University of choice by the article author. This year’s donation is awarded to: The Institute of Robotics at the University of Maribor, Maribor, Slovenia. The University conducts research on motion control solutions.

2010 GreenPower Leadership Awards. Voting has already begun for our expanded 2010GreenPower Leadership Awards Program and will continue through the April 2010 issue of Power Systems Design Europe. If you want to summit editorial content on “energy efficiency” to be judged for next years program, contact Cliff Keys, Editor-in-Chief, cliff.keys@ powersystemsdesign.com For sponsorship opportunities, contact Julia Stocks, Publisher, [email protected].

2009 GreenPower Leadership AwardsWinners and Finalists Announced at PCIM Europe

For the past year the readers of Power Systems Design Europe have been voting for the best editorial contribution in the area of “energy efficiency”. The 2009 GreenPower Leadership Awards program has been made possible by the financial contributions of our two Gold Sponsors: Intersil and Linear Technology.

GreenPowerWinners09_half_hor.indd 1 5/27/09 4:31 PM



imum load. Now, these efficiencies can be spread across a wide range of loads. The Eltek Valere HE rectifier family, for

High Efficiency Rectifier Technology

‘Green’ electric vehicle battery chargingThe main focus and key success factor for the electrical vehicle has always been the energy density

of the batteries and the recharge time. As new models hit the road and volumes pick up, focus is also

diverted to other key components within the electric vehicle. One such component is the battery charger.

Often considered a non-critical component, many have chosen the first available product in the early

phase and this is not necessarily a good decision for the long term.

By Morten Schoyen, Chief Marketing Officer, Eltek Valere, Norway

Choosing the latest charger tech-nology has many advantages for designers as well as the car

owner, who will see a less impact on his electricity bill. The efficiency of rectifier technology, for example, has improved dramatically over the years. Switch mode rectifiers were introduced in the late 1970s, and were a revolution with respect to the power density and effi-ciency. Typical efficiency for these first-generation switch-mode power supplies was in the 85% range.

During the 1990s, metal-oxide semi-conductor field-effect transistor (MOS-FET) technology, together with improved soft switching topologies and control solutions, enabled significantly higher switching frequency and partially loss-less switching. This resulted in further improvements in power density and ef-ficiency. Today’s 48V rectifiers typically achieve efficiencies of 90% to 91%, with the best-in-class rectifiers at 93%.

Now, almost lossless switching tech-nology has been combined with the lat-est in digital power-based components and software to boost power conversion efficiencies well over 96%. The use of more digital controls also expands the range of voltage loads that can achieve the highest efficiencies. In the past, peak efficiencies happened only at max-

example, is rated for 96% efficiency at 50% of load, and at 95% efficiency at 100% of load.


Figure 1: Three stages of battery charging.




tion which has to be taken into account. In many cases double-sided PCBs are used. Specified areas on the PCB are very often defined where only one kind of solder process can be used. Surface mounted semiconductors are increas-ingly the preferred solution for cost-efficient solutions.

The Art of Motor Drive Design

Intelligent power modules provide rugged solution

Modern adjustable speed drives are equipped with frequency converters to adjust the output current to

the requirements of 3-phase motors. The size and form of the frequency converter is often limited by the

application. In many cases the PCB is placed close to the motor where the height of construction can

also be limited. Furthermore the physics of the high power semiconductors used and the shaping of the

chosen package demand adequate positioning on the PCB. The overlap of voltage and current during

switching operation causes losses which have to be dissipated. The capability of power dissipation can

be improved by heat sinks which can further restrict the layout of semiconductors on the PCB.

By Dr. Stephan Chmielus, Field Application Engineer, Fairchild Semiconductor, Germany

Frequency converters are the key technology to meet the require-ments of EcoDesign. Using

modern motors fed by an inverter saves up to 40% compared with non-inverter drives according to the Electric Power Research Institute. Frequency convert-ers generate sinusoidal currents to feed either induction motors, permanent magnet synchronous motors or brush-less direct current motors. For this pur-pose the switching frequency has to be some order of magnitude greater than the adjustable output frequency of the frequency converter. The pulse width modulated output voltage is applied to the inductive load. Therefore, the output current is proportional to the mean value of the voltage. High switching frequen-cies are a benefit for frequency con-verters resulting in lower torque ripple, higher dynamic and less noise from the drive. This requires fast switching opera-tion which results very often in high di/dt and dv/dt. Hence, parasitic components become an important issue and ad-ditional efforts are needed to meet the current and future standards of EMC.

Cost saving is another layout restric-

Design ConsiderationsA current trend in high power semi-

conductors such as IGBTs and MOS-FETs, is to shrink the die and to improve the performance at the same time. However, a smaller die area reduces the parasitic capacitances of the device which leads to faster switching. Hence,


Figure 2: Intelligent power modules.

Figure 1: Simplified commutation circuitry.



charger’s capabilities to cost-effectively develop a customer promotion that al-lows consumers to pay a monthly fee to have the company take full responsibil-ity for the performance of the battery including a full maintenance service agreement, insurance, carbon offset payments and, in some countries, even all electricity used. This type of promo-tion puts pressure on all components to be high performance.

In this car, the charger is placed in an IP-rated housing and mounted against the cold-plate in the car, although it could be mounted next to a water cooled plate or any part of the car with a sufficient heat transfer capacity that would not exceed the charger’s 60 degree Celsius temperature rating. The car is an ultra-subcompact “city” car measuring only 10 feet long and five feet wide. The charger’s small size (11 inches x 5 inches by 2 inches) and weight (5 lbs) was a key benefit as it gave the designers room for other engine compo-nents.

This car also exemplifies the trend of advancing the computerization of cars to the next level by building in Internet access into the auto to deliver status updates to its owner or, more practi-cally to fleet owners. Another reason the Powercharger 3000 was chosen for this

application was its integrated Control Area Network (CAN) communication capability, which facilitates reporting of a wide range of charging and bat-tery health status indicators to other systems in the car which can then be communicated across the Internet.

As electric cars enter a new phase of maturity, carmakers are looking at ways to build on their success with new improvements. By improving the battery charger, these cars can be made even more energy efficient and environmentally friendly. The key is to leverage technology improvements in virtually lossless switching and the latest in digital power controls to boost efficiency levels to 96%. The benefit of these new systems extends to customer value, to design flex-ibility, to overall cost benefits and to increased reliability.

the design. This can save automakers in their bill of materials by specifying a smaller cooling plant for their vehicles.The lower operating temperatures also mean a more reliable design because no fans are needed for the design, elimi-nating a source of unreliability. Finally, reduced heat makes it possible to user smaller components in the rectifier and can have a dramatic impact on recti-fier size. Comparing, for instance, the industry’s leading 92% efficient charger with the similar capacity high efficiency charger we see that the HE rectifier is 10 pounds lighter and occupies only 25% of the space.

Case studyOne of the first high efficiency battery

chargers available is the Eltek Valere Powercharger 3000, which leverages the company’s 96% efficient rectifier technology, originally developed for tele-communications applications.

The charger is optimized for 16A wall sockets and features 3kW of power out-put and 295-421VDC output range. The charger has been adopted for use by a leading European electric car manufac-turer that has designed everything from its plant to its car to be as low carbon as possible.

The customer company also used the

This is a very important consideration as battery charging takes place in three stages, each of which put a different load on the rectifier. At first, the bat-tery goes into a constant current stage that raises the cell voltage to a preset level that is approximately 70% of the total charge level. During stage 1, the constant high current draw keeps the charger at near 100% of load. Stage 2 involves a topping charge where the current is gradually reduced as the cell is saturated. During this stage, the load on the charger drops with the current. If the charger doesn’t have a wide peak efficiency level, then the average ef-ficiency for the entire charging process will drop dramatically. Depending on the battery type, both of these stages can take about the same amount of time – even though most of the charging is done during stage 1. (Stage 3 is the float charge stage, and in this stage, the charge current drops dramatically to a level that keeps the battery filled while it self discharges. The operation of the charger is very light during this stage and does not have a significant effect on overall efficiency.)

Battery chargers go high efficiencyApplying high efficiency technology

to electric vehicle batteries, results in a charging subsystem that mirrors the en-vironmental ethos of the car owners, as well as reducing their electricity bills.

Adopting a high efficiency charging system can cut two thirds of wasted electricity attributed to the battery charging operation. This can have a substantial impact on the consumer. If we take the example outlined above, of the average driver who trav-els between 12,000 and 20,000 miles, the use of a high efficiency charg-ing system will drop their necessary power draw to between 3,328kWh and 7,654kWh. Therefore, the amount of power wasted on conversion from AC to DC is a maximum of 294kWh per year – saving up to two-thirds of the wasted power in this application.

This increase in efficiency corre-spondingly reduces the amount of heat generated by the entire engine, which lowers the overall internal temperature of the engine compart-ment reducing the cooling needed in www.eltek.com

Eltek Valere Powercharger 3000.

Eltek Valere Powercharger 3000 in an IP-rated housing.



fastening may apply stress to inside dies which could lead to breakage or degra-dation of the module. An example of the recommended fastening order is shown in Figure 4 (partially tighten: 1 to 2; tighten: 2 to 1). In general the partially tighten torque is set to approximately 25% of maximum torque rating. Once the heat sink is cor-rectly attached, the internal thermistor of the SPM module enables an easy access to the heat sink temperature hence simpli-fying the PCB design.

A possible construction of a box converter which is optimized for com-pactness and simple manufacturing is

so critical. The pin length of discrete IGBTs should be as short as possible to minimize parasitic capacitances and inductances. The arrangement of six co-packed IGBTs and gate drivers needs to be well thought out. Additional insula-tion sheets are required to fit the devices onto the heat sink. In many cases the bulky heat sink needs to be placed at the edge of the PCB.

To overcome the above mentioned restrictions, the use of intelligent power modules (IPM) also called Smart Power Modules (SPM®) is an attrac-tive solution. Figure 2 shows typical fully isolated modules containing a complete 3-phase

VSI plus gate drivers and protection functions. The use of these modules can reduce the required space on the PCB by up to 50% compared to a discrete solution. As an example, the TinySMD module has an overall size of only 29mm by 12mm. The small pack-age and the internal arrangement of the 3 half bridges and gate drivers ensure minimal stray components. All critical loops are inside the module. The device characteristic can be perfectly adjusted during the development of the SPM module to meet the requirements of a frequency converter. In particular, these modules are designed to meet the EMC regulation with a minimum of additional external components. The peak and average EMC level of this module is sig-nificantly lower in value than the values found in conventional types.

Similar to the discrete solution of a VSI, attention has to be paid to the external arrangement in the case of in-telligent power modules. Figure 3 shows the recommendations for Motion-SPMTM modules. Because of the fast switching operation of the VSI it is mandatory to separate Signal GND and Power GND. Both grounds are interconnected at the 15V Vcc capacitance. A thin ground path between Vcc capacitance and Power GND enforces the decoupling. To prevent surge destruction, a low-inductive capacitor should be placed between the P pin and Power GND. Long wires between VSI and the motor

shaded red in Figure 1 and they are characterized by high di/dt. This also includes the corresponding gate drivers. Minimized stray inductances ensure safe operation of the gate drivers and IGBTs. In particular for the HS gate driver, a negative voltage swing beyond the mini-mum allowable voltage at VS caused by resistive and inductive voltage drops across the LS diode and current path could lead to abnormal operation.

One solution is to slow down the switching speed by increased values of the gate resistances which unfortu-nately leads to much higher switching losses. Therefore, an optimized layout is required to use the whole performance of the VSI. To decouple the power area from the signal area the respective grounds should be separated. Gate drivers should be placed as close as possible to the IGBTs without any loops or deviation. The signal path between microcontroller and gate driver is not

a closer examination of the critical loops on the PCB becomes more important. A simplified circuitry showing the two typical switching operation of a voltage source inverter (VSI) is pictured in Figure 1. IGBTs are the preferred devices for higher currents at limited switching frequencies. The figure on the left hand side shows the current commutation from the high side (HS) freewheeling diode to the low side (LS) IGBT. At the beginning, the current is flowing in the freewheeling path formed by the HS diode and by the IGBT of the corre-sponding opposite half bridge. As soon as the LS gate driver turns on the IGBT, a short circuit current will flow through the HS diode and LS IGBT reducing the current of the diode and increasing the IGBT current in the same amount (natural commutation: 1 to 2). The current of the inductive load can be considered as constant during the switching operation. Therefore, stray components are irrel-evant in this path. The switching speed is determined by the turn-on of LS IGBT and by the stray inductances of the half bridge. The reverse commutation from the LS IGBT to the freewheeling diode requires a larger voltage drop across the LS IGBT than the DC bus voltage to turn on the freewheeling diode. Hence, the IGBT has to withstand high voltage and high current at the same time before the current commutates to the diode (forced commutation: 2 to 1).

The critical current paths of a VSI are

www.fairchildsemi.com

Figure 5: Box converter equipped with SPM.

Figure 4: Exaggerated illustration of module mounting.

Figure 3: Layout recommendations.

could result in high voltage reflections. Some SPMs are equipped with external gate resistors to adjust switching speed and to minimize these reflections.

Mounting ConsiderationsThe surface of the SPM module, with

exception of the TinyDIP/SMD module, exhibits a certain warpage. An exagger-ated illustration of the warpage is shown in Figure 4. The module is pressed onto the heat sink by screws from the middle of the surface to the outside. The convex surface ensures a sufficient heat trans-fer from the module to the heat sink if it is correctly mounted. Excessive uneven

shown in Figure 5. The reverse side of the housing can be used for cooling. The module is directly attached to the heat sink which stabilizes the com-plete housing at the same time. Fur this purpose the module has to be mounted upside-down. A heat sink is not needed for TinySMD SPM modules up to an output power of about 90W at high power factor, hence simplifying the PCB layout.

ConclusionThe design of a PCB is more sensitive

to the layout of the components using of today’s fast switching devices. Among other influences, the size of current loop areas with fast variations of the current is responsible for the robustness of the chosen semiconductors in frequency converters. Intelligent power modules minimize the area of current loops and their parasitic components. Fur-thermore, they are designed for fewer external components to meet the EMC requirements and greatly simplify the design of frequency converters.

Coming in the NovemberSpecial Report – Powering Communications

The November issue of PSDE, both print & online editions, features a special report on Powering Communications.

It is just incomprehensible to imagine the way we live and work today without the richness and connectivity that modern communications provides. We can be connected by our cell phones, computers and even our home entertainment systems as never before. The communications systems today are so rich in features, all of which require power – often in short supply - and sophisticated power management for remote or portable operation.

This broad based communications feature will cover power aspects in, but not limited to, the following areas:

• Mobile communications• Cellular equipment and systems• Radio communications• Data communications

www.powersystemsdesign.com




sales for HEV type cars will escalate. The simplest form of HEV is one that is closer to a standard internal combustion engine (ICE) car with start-stop function. Here, the start-stop function switches off the ICE when it is not needed (traffic lights, traffic jams, etc.).

In addition, there are micro, mild, full and plug-in hybrids. While micro-hybrids have an additional braking energy recu-peration system in addition to the start-stop function, mild hybrids add a 10-20 kW medium power electric motor to assist the combustion engine and improve the efficiency. Likewise, the full hybrids and plug-in hybrids offer the capability to drive a certain range without the ICE, purely on their high power electric motors (often up to 100kW). Lastly, the pure electric vehicles with no ICE at all, but equipped with very strong electric traction motors

Automotive PowerThe mission to boost energy efficiency

International Rectifier is focused on developing high quality, reliable and rugged automotive solutions,

capable of addressing the new demands of the automotive industry. With a vigorous new leadership, the

company is on a mission to develop power management solutions that foster fuel or energy efficiency

with a high degree of reliability and environmental compatibility.

By Henning Hauenstein, Vice President and General Manager, Automotive Products Business Unit,

International Rectifier Corp.

For applications ranging from direct fuel injection to electric motor drives for hybrid electric vehicles

(HEVs), IR has created an automo-tive business unit over five product lines ranging from low/medium/high voltage switches to intelligent power switches (IPS) and high-voltage mixed-signal devices capable of driving power MOSFETs/IGBTs up to 1200 V (Figure 1). Implemented in proprietary silicon processes combined with in-house developed advanced packaging tech-nologies, these devices provide auto-motive certification, enhanced reliability, and operational change management complying with or exceeding automotive industry norms.

The ability to differentiate products from other suppliers has maintained IR’s leadership position in power semi-conductors. It continues to extend this lead with the introduction of the revolution-ary GaN power technology and platform. Last year IR un-veiled its proprietary CMOS compatible GaN-on-Si based high electron mobility transis-tors (HEMTs) and associated development platform that will deliver an order of mag-nitude better figure of merit (FOM) than existing state-of-the-art silicon and silicon carbide (SiC) MOSFETs.

HEV TypesGiven the soaring gas pric-

es and the worldwide pres-sure to cut CO2 emission,

with power levels beyond 100kW.

All HEV versions present their own tech-nical challenges with very different power management requirements. Governments, societies, current automotive technology standards and geopolitical or environ-mental needs and constraints will strongly influence the success of the various HEV types and favor different HEV solutions in different regions of the world.

Cost vs. Fuel EfficiencySeeing a strong focus in Europe,

IR’s biggest near term opportunities will be the start-stop function and micro-hybrids. OEMs like BMW already offer this function as standard on their mainstream platforms. The production cost adder for the entire start-stop func-tion could probably be kept as low as $300, while still offering quite impres-

sive fuel savings in the 3-10% range. In fact, in congested city traffic, the savings can go up to 25%. The low cost benefit comes from the fact that the start-stop function can run on a standard 12V power net. Unlike mild or full hybrid vehicles, the start-stop function does not need expensive higher voltage systems or energy stor-ing batteries.

However, frequent cranking cycles of the combustion en-gine using a powerful


Figure 1: IR’s five product lines deliver a variety of automotive grade mixed-signal driver ICs, low/medium/high voltage power MOSFETs and IGBTs and intelligent power-switches (IPS).

next 12 months includ-ing novel driver ICs with enhanced ruggedness and safety features and highly efficient power devices like IGBTs, MOSFETs, and Direct-FETs. To eliminate the need for microcontroller interaction in case of catastrophic failures and short circuits of the HEV traction mo-tors, IR designers have implemented protec-tion features in all motor drive ICs. IR has incorporated a negative transient safe operating

area (NTSOA) in its datasheets for nega-tive voltage spike immunity, which is a very common problem when switching high voltage IGBTs with high currents in HEV inverters (Figure 2).

On the packaging front, IR has expanded the DirectFET family. An auto-motive version has been released that is fully compatible with the AEC-qualifica-tion standards and 100% lead free. With industry leading low Rdson, superior switching performance and increased thermal capabilities (like both sided cooling); this advanced wire-less chip scale package allows drastic reduction in size and weight especially in power hungry or fast switching applications like HEV DC-DC converters.

This family will be expanded with a wide range of large power devices with solderable front metal, enabling custom-ized both sided cooling and eliminating bond wires; the root cause for mechani-cal failure over lifetime. Also, due to the high current needs of the application, this family will offer large area devices capable of unprecedented current den-sities.

The recently released automotive grade driver ICs such as AUIRS4427S, AUIR2085 and AUIRS2003, along with DirectFETs and standard packaged trench power switches exemplify IR’s commitment to the quest of boosting energy efficiency in automotive power.

kW range. Likewise, for full- and plug-in hybrids, as well as for the electric vehicles, incorporating motors up to and even above 100 kW, IR is offering IGBTs and driver ICs up to 1200 V breakdown voltage. The inverters driving those huge motors in these applications require automotive grade high voltage IGBTs.

In reality, such powerful motors op-erating at high voltages, typically from 100V up to 1200V, are probably the biggest change for engineers in the au-tomotive environment used to designing in a 12V world. So, there’s a big learn-ing curve here as industry extends its knowledge and experience to the new HEV architectures.

Since most automotive semiconduc-tor suppliers were mainly focused on the 12V systems, even 80V – 100V devices are considered as very high voltages. IR can easily leverage its high power industrial and appliance motor drives to deliver high voltage rugged power devices to these applications. The company has readied high voltage Si-processes for IGBTs and control ICs and by adding higher reliability and en-hanced ruggedness to the Si chips and packages, it is re-using and transferring proven industrial solutions to automotive applications. Consequently, IR has built a significant HEV portfolio comprising driver ICs and power devices and can also rapidly generate customized solu-tions for this market.

The company is releasing some 200 new products for this market within the

starter-alternator motor (typically 5-6kW) is the challenge for the start-stop system. To address this issue, IR is developing very rugged MOSFETs with junc-tion temperatures up to 200oC and with the abil-ity to handle very high avalanche energy from the starter-generator. Extreme ruggedness and enhanced reliability are the key features of these MOSFETs.

With start-stop on board, the 12V bat-tery net needs to provide the power for cranking the engine without suffering from the well known battery voltage drop. The design trick is to make this as smooth as possible. IR has readied fast switching components with very low EMI to enable designers to build highly efficient DC-DC converters buffering the power demand during the engine start-up. The new automotive DirectFET product line provides leadership power switches for this application.

While higher end HEV systems like the mild hybrids employ powerful electric motor to assist the ICE and provide fuel savings in the 15-35% range, they un-fortunately add often more than $1200 to the cost of the solution.

Similarly, the full hybrids, plug-in hy-brids and electric vehicles require power systems at much higher power levels (using higher currents and voltages) and correspondingly consume even more silicon content for power manage-ment functions. Additional components required to develop the electric motor drive can easily add cost in the range of $6000 to $15000 for a potential fuel efficiency gain of 30-50% in congested areas. This market segment is expected to grow a little slower. On the brighter side, the much higher silicon content needed in these systems compensates for the expected slower adoption rate.

High Voltage Power DevicesFor mild hybrids, the company has

developed advanced motor drive solu-tions for powertrain motors in the 10-15

www.irf.com

Figure 2: IR’s automotive motor drive ICs offer negative voltage spike immunity; the first automotive supplier to include negative transient safe operating area (NTSOA) in its data sheets for motor drive ICs.



a typical common-mode-rejection ratio (CMRR) of 90dB and 12-bit performance available for common-mode input volt-ages up to 50V (as with 12 Li-Ion cells). With an input voltage capability of 60V, a group of LT1991 amplifiers can be used for a practical design of twelve cell potentials. With the addition of high voltage translation or isolation, this sub-circuit can be repeated to capture as many cell readings as necessary.

Battery management electronics

should also incorporate a number of other features such as the state-of-charge determination which requires temperature and current measurements. For practicality, the BMS electronics should be powered directly from the bat-tery sense connections. For safety, the interface between data acquisition and host processor should be galvanically isolated. It is important for HEV and EV batteries to maintain cell balance. This is a critical function for high-powered battery packs because a long series of individual cells is only as reliable as the

EV/HEV Battery ManagementPractical design considerations

Until recently, NiMH was the battery chemistry of choice for electric vehicles. However, the economic

justification was marginal, and next generation Lithium Ion (Li-Ion) batteries provide a more viable

solution with an energy storage density improvement of 150%. Several lithium-ion chemistries, such as

lithium-iron phosphate, lithium-manganese, and lithium-titanate are good EV/HEV candidates. Unlike

Lithium-cobalt cells, these lithium chemistries are thermally stable and offer low equivalent series

resistance (ESR) to support high current and if carefully managed, could last 10 to 15 years.

By Greg Zimmer, Product Marketing Engineer, Signal Conditioning Products,

Linear Technology Corporation

The need for battery managementLi-Ion cells provide a typical operat-

ing potential of 4V at full charge and 2V at full discharge, specified by the cell vendor. To minimize the level of current, which allows for smaller, lighter and less expensive cables and motors, the EV/HEV battery pack is typically stacked as a group of 100 to 200 series-connected cells. Even at these high voltages, the peak charge and discharge currents of EV/HEV battery stacks can exceed 200A.

Charging any Li-Ion cell to 100% of its state-of-charge (SOC) or discharg-ing to 0% SOC will degrade its capac-ity. Therefore, only a portion of a cell’s full capacity can be used. With very accurate control of the state of charge of each Li-Ion cell, battery pack capac-ity can be maximized and degradation minimized, but controlling hundreds of series-connected cells is a challenge.

The Battery Management System (BMS) electronics quickly and ac-curately measure each battery cell’s voltage. This requires extracting rela-tively small differential voltages from very high common mode voltages (2V - >1000V), accomplished by placing high-quality, high common-mode differ-ence amplifiers at each cell to provide translated signals, ready to be digitized by an analog-to-digital converter (ADC). An example of this type of amplifier is Linear Technology’s LT1991 with built-in resistors matched to within 0.01%, with

weakest cell. It is critical that the charge level of all cells does not stray outside the recommended SOC range. The designer must design for long periods of battery pack inactivity, prior to installa-tion and during vehicle transport or stor-age. To ensure that the batteries cannot


Figure 1: Traditional Data Acquisition Approach Using a Precision Differ-ence Amplifier. Output offset is 2.5V and gain is 0.25.

Figure 3: Cell-balancing circuit for LTC6802 with protection features (circuitry for only one cell shown).

communication and operates as a stan-dard slave I/O device. This allows all BMS algorithms to be software-coded and controlled exclusively by the devel-oper. Figure 2 shows the basic configu-ration of an arbitrary cell-count battery module.

Figure 3 shows a typical battery input circuit used with the LTC6802, including a PMOSFET switch and a load resis-tance for balancing, along with other passives used for filtering and protec-tion. When using a high cell count IC, cell balancing is usually best addressed with external transistors to allow for suf-ficient current and avoid IC overheating. To maximize cell measurement accu-racy, the cell-balancing switches should be opened during ADC conversions; this ensures that I•R drops in the cell con-nections are minimized. During periods of inactivity, the LTC6802 automatically opens all balancing switches and as-sumes a minimum power consumption condition to prevent inadvertent battery discharging.

The cell-balancing switches can also be used for self-testing by adding a resistance in series with the battery inputs, as shown. If the switch is on, the cell reading will show a predictable voltage change, providing a validation of both the switch and the ADC port functionality. This capability requires, however, that the cell balancing switch is enabled during ADC conversions. The LTC6802 has anticipated this self-test capability and ADC measurements are enabled under these conditions with a simple configuration command.

Reference designFigure 4 shows a complete twelve

cell, lithium-ion battery stack schematic using the LTC6802. Higher cell count systems are easily created by replicating this circuit and cascading the SPI con-nections. Communication between ICs is accomplished with a current-mode SPI signal. Protection circuitry is includ-ed in this design for start-up surges and ESD events, which are typical during production or service of the vehicle. The protection circuitry simply consists of a surge suppressor and a small number of series resistors and Schottky diodes.

The LTC6802 also provides other

and complexity, as the number of mod-ules is increased.

Connecting a large high voltage bat-tery stack to electronics presents an-other significant design challenge. The battery to electronics interface typically consists of many contacts, over many individual connectors. Since the data acquisition electronics are typically un-powered when connected to the batter-ies and connections can occur random-ly, the BMS must be designed for hot-plugging. To prevent damage from surge currents, external protection is prudent. As an example, standard Zener diodes across each cell input will automatically distribute safe voltages across missing inputs as random contacts are made.

Consolidated approachAn ideal solution for many battery

management functions is an integrated battery monitor. As an example, Linear Technology’s LTC6802 is a “building block” that enables construction of high performance battery modules with a minimum of components. This multicell monitor device provides direct 12-bit digitizing of up to 12 series-connected battery potentials, cell balancing con-trols and two spare ADC inputs for other signals, such as temperature sensors. Unlike the difference-amplifier approach, the LTC6802’s on-chip ADC does not depend on resistor networks, The LTC6802’s approach provides a uniform, light load on each cell and automatically assumes a low-power standby condi-tion during idle periods to reduce power. The LTC6802 uses a Serial Peripheral Interface (SPI) for command and data

be over-discharged during inactivity, the BMS idle power must be significantly less than the self-discharge of the bat-tery cells. More importantly, the idle current along the battery string must be well matched to ensure that batter-ies do not become unbalanced during storage. Once placed into operation, the batteries will experience high charge and discharge currents. In this situation, the charge level of each cell must be ac-tively balanced to derive maximum pack energy and lifetime.

A simple and cost effective technique for cell balancing, commonly used in EV/HEV designs today, is passive-balanc-ing. With passive-balancing, a resistor is placed across a cell when its state of charge exceeds its neighbors. It should be noted, however, that passive-bal-ancing wastes energy and can generate considerable heat. Future generations of battery management systems will likely incorporate active balancing, where a storage element is used to move charge between cells.

Other EV/ HEV battery design considerations

It is important to consider the physical size and weight of a large battery cell array and that significant data is gener-ated by continuously monitoring every battery cell. A module approach may be suitable where the battery management task is divided into subsets and a local processor can be placed at the module level. This also provides a common plat-form for multiple designs. The minimum size of the module will likely be dictated by the additional wiring harness cost

Figure 2: Basic topology of a high cell-count EV/HEV module.



the LTC6802 cell-monitoring platform. This battery monitoring IC simplifies the data acquisition task with a single device, while improving intrinsic reliabil-ity and adding feature-rich functionality. A well thought out design, including plenty of protection mechanisms, pro-vides robust operation in the unforgiving environment of high-energy EV and HEV systems.

www.linear.com

useful features that simplify module circuitry, such as an onboard 5V series regulator, general-purpose ADC inputs, and general purpose digital inputs and outputs (GPIO). As an example, the GPIO can be used as multiplexer con-trols to expand the two general purpose ADC inputs to an eight-channel capac-ity. To ensure proper IC operation, an open-drain output watchdog timer is provided to indicate an idle period of

communication.

ConclusionAfter years of effort and steady prog-

ress, high energy battery systems will soon be practical for everyday use, es-pecially as part of the electric and hybrid electric vehicle. The cost and complexity of high-quality Lithium-Ion battery data acquisition and cell-balancing controls are greatly reduced with the advent of

Figure 4: Complete 12-cell Battery Monitoring Circuit with cell-balancing and protection circuitry.



/ temperature cycles). The last thing a manufacturer wants in the field is an installed base of ‘potential problems’. A system that looks good on the balance sheet and performs well in the lab, but after being deployed for some time in the field, suddenly fails. This escalates the cost of the original calculation sig-nificantly and can destroy the credibility that has been built by the manufacturer, perhaps over decades.

There is also a high risk factor associ-ated with developing a discrete solution. It can be readily appreciated by any designer that these projects never run smoothly; there are the inevitable design bugs to remedy requiring re-spins, tough specifications to achieve and regulations to adhere to and get verified. All this takes up valuable time as well as a significant financial burden.

With vital impact on system design, the PCB layout needs to be given special attention. Lead inductance and resis-tance play a great part in determining the losses and heat generation, all of which –unless skillfully compensated- can impact detrimentally the reliability of the system. Down time costs for an installa-tion are painfully high and servicing faulty units in the field is difficult, expensive and in worst case, impossible. Reliability is therefore a premium priority.

One huge initial apparent cost saving which can result in failures further down the line is the use of a copper substrate. Naturally, it looks an ideal and simple

Mastering Power ModulesThe advantages over discrete solutions

With the strong upsurge in the need for power electronics in diverse applications such as solar,

wind, freight and logistics, industrial drives and UPS, system developers are becoming increasingly

aware that the selection of the power devices for these often remotely-sited applications is critical.

Such systems require the highest reliability and efficiency to minimize operating costs and build

industry integrity for the developer.

By Werner Obermaier, Head of Product Marketing, Power Modules & Hybrids,

Vincotech GmbH, Munich, Germany

There is much discussion over what actually constitutes a Power Module. A typical example of a

PIM Module for a drives application (see figure 1) would contain, in a single housing, a three-phase input rectifier, a brake chopper, a three-phase inverter plus the vital isolation material; the high prevalent voltages and levels of thermal stress are key to the ‘excellence factor’ encompassing the efficiency and most importantly, the long-term reliability, of a well designed PIM.

Power Modules have become the most convenient way of building a cost effective power supply system. Every-thing is optimized within the constraints of the particular module manufacturer. The designer can rest assured that the module will do everything electrically and mechanically that is specified, is guaranteed to work to tight specifica-tions such as EMI, efficiency, load cy-cling and reliability and it is virtually ‘off-the-shelf’ meaning that the all-important time to market pressure is minimized.

There is understandably a healthy debate on the various arguments on the topic of Module vs. Discrete component structure. Cost inevitably seems to be the driver in such arguments but then, what is meant by cost? If we simply take the sum of the costs of the individual discrete parts plus the cost of putting these parts together and shipping as the system cost, then we are delud-ing ourselves. We need to consider the impact of system reliability (load cycle

Figure 1: Typical power module show-ing the vital electrical and thermal con-nections (a), and the robust, mainte-nance-free housing (b).

Figure 2: High quality, low inductance electrical and thermal connections are vital for high efficiency and field reliability.



moved without damaging the PCB, thus allowing for the reuse of the PCB with a new module. Design flexibility is guar-anteed by the elimination of the need for soldering; the module can easily be mounted on either side of the PCB at no extra cost or effort.

Vincotech’s Press Fit Technology is DIN and IEC compliant and available upon request for all Vincotech propri-etary module designs in pin-compatible versions to the current portfolio.

Of course, there are many module vendors in the marketplace and this can be a tough call to make the right selection. Most vendors are tied to their

own proprietary brand in the choice of materials technology and semiconductors. Naturally it just would not be possible for module-maker ‘A’ to select MOS-FETs from semiconductor-maker ‘B’. Corporate rules would simply not allow it. But does this result in an optimum module? The answer is of course, no. To get the best of all worlds, one needs to go to an independent source such as Vincotech, where components and assembly can be truly opti-mized or even customized. Never has this been more important than in the power generation field where every percentage point in efficiency, reliability and therefore true cost of ownership adds to the customer’s profitability.

technology. The modules featuring this new pin technology are simply pressed in rather than soldered into the PCB, thus reducing PCB assembly time and costs.

The Vincotech press fit pin design is well established in the automotive indus-try and provides a reliable and gas-tight connection to the PCB. By eliminating the necessity of a solder process for the module, the customer reduces the PCB assembly time and thereby the produc-tion costs. A higher production output capacity is the direct consequence.

Further advantages of press fit tech-nology include reuse of the PCB and design flexibility. The module can be re-

choice due to its electrical and thermal conductivity. The problem can appear after field service and the normal in-service load and temperature cycling (see Figure 3). The thermal expansion of copper is not a good fit with silicon. A much more reliable solution is achieved with the use of ceramic. This is a solu-tion that is initially more costly, but pro-foundly more reliable. One has to invest up-front to save costs and integrity in the longer term.

Apart from the electrical design, a vital part of the overall product ‘persona’ is the mechanical design and component layout. Power systems generally operate in extremes of the environment. Poor ventilation and hostile physical and environmental conditions are generally the norm in industrial drives and renewable energy in-stallations such as solar or wind power.

A great impact on system cost is the assembly process. Dis-crete processes where solder-ing power semiconductors with heatsinks attached needs to be very precisely aligned requiring specialized alignment tooling and skilled operators in many cases or individual heatsinks for each semiconductor, using operators of similarly high skill levels.

Vincotech has introduced its new press fit technology to the market and is now shipping modules using this labour saving

www.vincotech.com

Figure 3: The layered construction of the PIM using highest quality materials (a), provides optimum long-term reliability with a high load cycling capability (b).

Figure 4: Press fit technology utilizes new pin technol-ogy (a), which eases assembly to the PCB (b), reduc-ing time and cost of manufacturing.

C M Y CM MY CY CMY K

Mes

ago

Mes

sem

anag

emen

t G

mb

H, P

ost

fach

10

32 6

1, 7

0028

Stu

ttg

art,

Ger

man

y, s

ps@

mes

ago

.co

m, p

ho

ne

+49

711

6194

6-82

8

Experience electric automationat its best!Come and see it all!Control TechnologyIPCsDrive Systems and ComponentsHuman-Machine-Interface DevicesIndustrial CommunicationIndustrial SoftwareInterface TechnologyElectromechanical Components and Peripheral EquipmentSensor Technology

www.mesago.com/sps

ElectricAutomationSystems and Components

Exhibition & Conference24 – 26 Nov. 2009Nuremberg

SPS/IPC/DRIVES/

http://www.mesago.com/sps


duce its own life-cycle carbon footprint, has quantified targets in place to do so. The company is advocating an innova-tion-driven climate agenda that utilizes technology and will include the deploy-ment of transformative low-carbon solu-tions that broadband technology can deliver.

The British Government has an-nounced plans to build the world’s largest wind turbine blades, with the company, Clipper Windpower. The US-based firm has been given a $7.15 mil-lion grant from the British government to construct the blades in the North East of England. The announcement comes at a time when the prospects for wind power and other renewable energy projects seem unsteady in the UK. Earlier this


With everything ‘apparently’ Green hitting the media at the moment, it’s a tough call to

‘sort the wheat from the chaff’ and to give a balanced report. Some compa-nies and governments do however have a firm commitment to environmental issues and do invest (a true test!) in achieving this in their organisations or regions.

Ericsson has recently asserted that the Information and Communication Technology (ICT) sector is crucial to creating a low-carbon 21st century in-frastructure. Using today’s communica-tions infrastructure in a smart way can maintain economic development while dramatically reducing emissions.

Giving a keynote address at the Broadband World Forum in Paris, Er-icsson CFO and incoming CEO Hans Vestberg said: “For too long the need for CO2 reductions has been seen as a trade-off between economic develop-ment and care for the planet. This does not have to be the case. Because the ICT sector can reduce CO2 emissions substantially, government representa-tives have an opportunity to bring ICT onto the agenda for the upcoming United Nations framework Convention on Climate Change in Copenhagen (COP15) in December.”

Ericsson, strongly committed to re-

year, the turbine manufacturer, Vestas, closed down two operations in Britain, citing a lack of demand in northern Europe. A subsequent study by the Brit-ish Wind Energy Association found that most regions in England were unlikely to meet the government’s 2010 renewable energy targets.

Contrary to the political rhetoric, prog-ress has been impeded by the rejection of applications for wind turbines by local councils citing the strict planning regula-tions for offshore and onshore wind turbines. As a result, proposed schemes have suffered extreme difficulty in ob-taining permission. Many potential sites have been blocked by a combination of opposition by residents to proposals in their local area and controversy over ‘health issues’ with councils opting to fight for the maintenance of the tradi-tional appearance of towns and cities rather than clean energy projects.

It is however, refreshing to see a more caring approach in mainland Europe where wind turbines and solar instal-lations are generally accepted as a necessary path towards environmen-tal responsibility and improvement for future generations. As always, without some pain there is normally very little gain.

Green: It Doesn’t Happen by Itself

www.powersystemsdesign.com/greenpage.htm

®

RoHSCOMPLIANT

21 Napier Place, Wardpark North, Cumbernauld UK G68 0LL+44/1236/730595 Fax +44/1236/730627

Plug in to our Power ToolsMake better, faster power inductor choices withCoilcraft’s powerful web tools.

Start with an IC, a converter topology, or alist of electrical and mechanical specs. In sec-onds you’ll get a list of every Coilcraft part that

could work for you, complete with price infor-mation. A few more clicks gives you a detailedanalysis of core and winding losses.

Check out our complete power designer’stoolbox at www.coilcraft.com/powertools

210 x 297PEE PSDE

http://www.coilcraft.com/powertools

Part NumberVDS

(V)ID

(A)

RDS(on) MaxVGS=10V

(mΩ)

Qg(nC)

Package

IRFH7921PBF 30 14 8.5 8.3 PQFN (5x6)IRFH7932PBF 30 25 3.3 34 PQFNIRFH3702TRPBF 30 16 7.1 9.6 PQFN (3x3)IRFH3707TRPBF 30 12 12.4 5.4 PQFN (3x3)IRF8721PBF 30 14 8.5 8.3 SO-8IRF8788PBF 30 24 2.8 44 SO-8

Part NumberVDS

(V)ID

(A)

RDS(on) MaxVGS=10V

(mΩ)

Qg(nC)

Package

IRFB(S)3004PBF 40 330 1.75 160 TO-220(D2-PAK)IRFB(S)3006PBF 60 270 2.5 200 TO-220(D2-PAK)IRFB3077PBF 75 210 3.3 160 TO-220IRFB4110PBF 100 180 4.5 150 TO-220IRF7853PBF 100 8.3 18 28 SO-8

Part NumberVDS

(V)ID

(A)

RDS(on) MaxVGS=10V

(mΩ)

Qg(nC)

Package

IRF3205Z(S)PBF 55 110 6.5 76 TO-220(D2-PAK)IRFB(S)3806PBF 60 43 15.8 22 TO-220(D2-PAK)IRF1018E(S)PBF 60 79 8.4 46 TO-220(D2-PAK)IRFB(S)3607PBF 75 80 9.0 56 TO-220(D2-PAK)IRFB(S)3307ZPBF 75 120 5.8 79 TO-220(D2-PAK)

IR’s latest MOSFETs offer benchmark performance and are tailored for DC/DC conversion, AC/DC synchronous rectifi cation, and Industrial Battery applications such as E-Bike and UPS systems.

IR’s latest Trench technology in a wide range of packages up to 250V enables low RDS(on), low gate charge, and high switching

capability for today’s demanding designs.

Your FIRST CHOICEfor Performance

Benchmark Point of Load: VRM, Buck Regulation

Benchmark Power Supply: Synchronous Rectifi cation

Benchmark Industrial: Industrial Battery, UPS

Benchmark Power MOSFETs forHigh Performance Applications

THE POWER MANAGEMENT LEADER For more information call +33 (0) 1 64 86 49 53 or +49 (0) 6102 884 311

or visit us at www.irf.com THE POWER MANAGEMENT LEADER

10650_10257_MOSFET_210x297mm.indd 1 07/09/2009 10:41

http://www.irf.com

Design TipDesign Tips s€¦ · AGS Media Group and Power Systems Design Europe ... and market...

Documents

Transcript of Design TipDesign Tips s€¦ · AGS Media Group and Power Systems Design Europe ... and market...