7th generation igbts improve thermal …...profile, thermally efficient solution with extended 11.5...

December 2017

Special Report: Electric & Hybrid Vehicles (pg 27)

7th Generation IGBTs Improve Thermal Performance of Industrial Power Conversion

VIEWpoint

Farewell to 2017 and welcome to 2018By Ally Winning, European EditorPower Systems Design

POWERline

Power Integrations Unveils a Third Generation of InnoSwitch

NOTABLE&newsworthy

Ionic ‘Solar Cell’ Could Provide On-Demand Water Desalination

MARKETwatch

Transitioning to Electric Revolution in MedTechBy Kevin Parmenter, PSD Contributor

DESIGNtips

Power Path Management in Charger ICs

By Aaron Xu, Monolithic Power Systems

COVER STORY

7th-Generation “X Series” RC-IGBT

Module for Industrial Applications

By Akio Yamano, Misaki Takahashi, and Hiroaki Ichikawa, Fuji Electric

TECHNICAL FEATURES

Wireless Sensors

How to Maximize Your Wireless Sensor’s RuntimeBy Ajay Kuckreja, Andrew Brierly-Green, Nazzareno (Reno) Rossetti, & Ole

Dreesen, Maxim Integrated

Data Security

Security is a Complex Topic and Requires Expertise and Solutions from all Product Segments By Bernd Hantsche, Managing Director Embedded & Wireless, Rutronik

Packaging

Packages Enable Full SiC Performance With Limited ParasiticsBy Courtney R. Furnival, President, Semiconductor Power Solutions

Power Supplies

Lighting - from stadiums to vegetables

By Patrick Le Fevre, Powerbox

SPECIAL REPORT:ELECTRIC & HYBRID VEHICLES

Maximize Run Time in Automotive

Battery Stacks Even as Cells Age

By Samuel Nork. GM, Battery Chargers, ASSPs & PMICs & Tony Armstrong, Director of Product Marketing, Power Products, Analog Devices Inc.

Automakers Shift to 48v Mild

Hybrid Systems

By Ed Kohler, Intersil Corporation, a Renesas company

Advances in Substrate Technology

Enable New Power Modules for

Automotive Application

By Matthew Tyler, ON Semiconductor

2

Power Film Capacitor Design for

EV and HEV Applications

By John Gallipeau, Technical Marketing Manager for Power Capacitors,

AVX Corporation

FINALthought

The Philosophical Battle for the InternetBy Ally Winning, European EditorPower Systems Design

Dilbert

44

6

5

4

Highlighted Products News, Industry News and

more web-only content, to:

www.powersystemsdesign.com

33

POWER SYSTEMS DESIGN 2017DECEMBER

1WWW.POWERSYSTEMSDESIGN.COM

37

40

7th Generation IGBTs Improve Thermal Performance of Industrial Power Conversion

13

7

44

COVER STORY

7th-Generation “X Series” RC-IGBT Module for Industrial Applications (pg 9)

28

9

14

17

20

24

REDEXPERT

The world’s most accurate AC loss model Filter settings for over 20 electrical and

mechanical parameters Inductor simulation and selection for

DC/DC converters Ability to compare inductance/current and

temperature rise/DC current using interactive measurement curves

#REDEXPERTREDEXPERT. Würth Elektronik‘s online platform for simple component selection and performance simulation.

www.we-online.com/redexpert

Available in seven languages Online platform based on measured

values No login required Order free samples directly Direct access to product datasheets

2

VIEWpoint

WWW.POWERSYSTEMSDESIGN.COM

POWER SYSTEMS DESIGN

Welcome to the December issue of PSD. This is the last issue in what has been an eventful year. This month our special focus returns to the automotive market, where power electronics is playing a bigger part than ever before.

The articles this month explore various aspects of that market, Initially, Intersil will tell us why 48V hybrid technology is gaining traction for automotive applications around the world. Although the technology may add a little cost to vehicles, the benefits outweigh that cost by a considerable margin. Legislation may make the changes necessary, but lower CO2 emissions, improved fuel economy, and vehicle drivability enhancements will make the changes worthwhile.

Next we have On Semiconductor detailing how direct bonded copper (DBC) substrates will bring more advanced power modules to the automotive industry. The technology has the potential to reduce the physical size and weight of the system and shortens the thermal path without any compromises in electrical isolation.

Of course, there is plenty of variety in the articles inside the issue. Powerbox gives an interesting overview on the future of LEDs and how power technology can bring LEDs alive for more applications than purely lighting. Rutronik will also contribute to the issue with an article on security. With internal and external data being used to control the power in systems, security is often an issue that is overlooked. Anything connected could allow hackers access to the system and take control. Because power control can be hidden on many applications, it doesn’t mean it is secure from intrusion. Not considering security could have repercussions throughout the whole system.

Finally, the holidays are traditionally a time to spend with family and recharge the batteries, and that is what I intend to do. In my home country of Scotland, New Year was, up until fairly recently, a bigger celebration than Christmas, which wasn’t even a holiday in Scotland until 1958. It is a time to reflect on the past and look to the future. In Scotland, the New Year break gives us a chance to put the past behind us and free our minds to concentrate on what is ahead.

The future in power looks very bright. Next year is set to be a really interesting as electric vehicles make further inroads into the automotive market. Who will bring an electric truck to the market first and will it cause real disruption in a short space of time? At the other end of the scale, nanopower will become more important as the IoT starts reaching its full potential. Will we find a long term solution to store power that will enable renewable energy to shine? I’m sure there will be further breakthroughs in areas that many of us haven’t see coming, and that’s what keeps the industry interesting.

Thank you for reading the magazine, and thanks to those who support it with advertising. I wish you all a very merry Christmas and a happy and prosperous New Year. See you after the break!

Best Regards,

Ally Winning European Editor, [email protected]

Farewell to 2017 and

welcome to 2018Power Systems Corporation 146 Charles Street Annapolis, MD 21401 USA Tel: +410.295.0177Fax: +510.217.3608 www.powersystemsdesign.com Editorial Director Jim Graham [email protected]

Editor - EuropeAlly [email protected]

Editor - North AmericaJason [email protected]

Editor - ChinaLiu [email protected]

Contributing Editors Kevin Parmenter, [email protected]

Publishing DirectorJulia [email protected]

Creative Director Chris [email protected]

Circulation Management Sarah [email protected]

Sales Team Marcus Plantenberg, [email protected]

Ruben Gomez, North America [email protected]

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

Power Systems Corporation and Power Systems Design Magazine assume and hereby disclaim any liability to any person for any loss or damage by errors or ommissions in the material contained herein regardless of whether such errors result from negligence, accident or any other cause whatsoever.

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

Volume 14, Issue 10

INTRODUCING POWER BY LINEAR. THE POWER TO SOLVE ANY PROBLEM.

ANALOG DEVICES’ NEW POWER BY LINEAR

PORTFOLIO ENABLES YOU TO TACKLE YOUR

TOUGHEST POWER CHALLENGES.

Now you can get to market faster with everything you need to

deliver breakthrough solutions that hold a true competitive edge:

Industry-leading performance

Proven reliability

Consistent quality

No-obsolescence policy

Premier service and support

LEARN HOW THE INDUSTRY’S LEADING PORTFOLIOJUST GOT MORE POWERFUL. linear.com/POWERBYLINEAR#PowerByLinear

54

POWERline

WWW.POWERSYSTEMSDESIGN.COM WWW.POWERSYSTEMSDESIGN.COM

POWER SYSTEMS DESIGN 2012OCTOBER

Efficiency is always the key in power designs. As well as getting the most use out of the supplied power,

higher efficiency means less is heat given off, so therefore less cooling is required. Less cooling means that components can be placed closer together or packages can be smaller, which in turn means overall solutions are smaller. This has prompted an ongoing chal-lenge between suppliers to deliver the highest efficiency products.

Power Integrations (PI) is one of those suppliers, and the company has laid out its own claim to pro-vide the highest efficiency offline CV/CC flyback converter with its recently launched InnoSwitch3 range. The company claims the In-noSwitch3 devices can achieve up to 94% efficiency across line and load conditions, which cuts losses by 25% over its previous range and enables the production of power supplies with an output of up to 65W without heatsinks. The 65W limit is enough to power devices up to the size of a laptop. Any higher and expensive power factor correction would be required.

The InnoSwitch3 range has a unique architecture that eliminates the need for an optocoupler. PI’s FluxLink technology communi-cates across the isolation barrier

Power Integrations Unveils a Third Generation of InnoSwitch

between the primary and secondary side without using magnetic material. This method of commu-nication is reliable, with the upside that perfor-mance doesn’t degrade over time like photo-coupler-based systems, making the output more reliable over the prod-uct’s lifetime. Fluxlink offers a high bandwidth link that provides a fast load-transit response. The lack of an optocoupler in the design makes InnoSwitch designs more compact and provides a higher power density.

There are currently three series of InnoSwitch3 products. The CE or Current External range features accurate CC/CV regulation with external output current sense to allow designers the maximum amount of flexibility. The CP or Constant Power range has been for designs that require a dynamic output voltage, for example USB Power Delivery (PD) systems. The EP or Embedded Power range has a 725 V MOSFET, and has line and load protection and multi-output cross-regulation.

The Innoswitch3 flyback switcher ICs also offer synchronous rectifi-cation, quasi-resonant switching and a precise secondary-side feed-

back sensing and control circuit. InnoSwitch3 devices are CCC, UL and VDE safety-certified to bridge the isolation barrier, and the InSOP-24 package provides a low-profile, thermally efficient solution with extended 11.5 mm creepage and clearance between primary and secondary sides for high reli-ability, surge and ESD robustness. The new devices also incorporate lossless line overvoltage and under-voltage, output overvoltage, over-power, over-current and over-temperature protection, as well as output rectifier short-circuit protection. Device sub-families are provided with either latching or au-to-recovery capability, according to the typical demands of each target application space. All InnoSwitch3 ICs feature on-board high-voltage MOSFETs (rated at 650 V for the CP and CE series and 725 V for the EP series).

Power Integrationswww.power.com

Ionic 'Solar Cell' Could Provide On-Demand Water Desalination

Modern solar cells, which use energy from light to gener-ate electrons and

holes that are then transported out of semiconducting materi-als and into external circuits for human use, have existed in one form or another for over 60 years. Little attention has been paid, however, to the promise of using light to drive another electricity-generating process -- the transport of oppositely charged protons and hydroxides obtained by dissociat-ing water molecules. Researchers in America report such a design, which has promising application in producing electricity to turn brack-ish water drinkable, on November 15 in the journal Joule.

The researchers, led by senior author Shane Ardo, an Assistant Professor of Chemistry, Chemical Engineering, and Materials Sci-ence at the University of California, Irvine, write that they have crafted an "ionic analog to the electronic pn-junction solar cell," harnessing light to exploit the semiconductor-like behavior of water and generate ionic electricity. They hope to use such a mechanism to manufacture a device that would directly desali-nate saltwater upon exposure to sunlight.

"There had been other experiments

dating back to the 1980s that photoexcited materi-als so as to pass an ionic current through them, and theoretical studies said that those currents should be able to reach the same levels as their electronic analogs, but none of them worked all that well," says first author William White, a graduate student in Ardo's research group.

In this case, the research-ers attained more suc-cess by allowing water to permeate through two ion-exchange membranes, one that mostly transported positively charged ions (cations) like protons and one that mostly transported negatively charged ions (anions) like hydroxides, functioning as a pair of chemical gates to attain charge separation. Shining a laser on the system prompted light-sensitive organic dye molecules bound to the membrane to liberate protons, which then transported to the more acidic side of the mem-brane and produced a measurable ionic current and voltages of over 100 mV in some instances (60 mV on average).

Despite crossing the 100 mV photovoltage threshold at times, the level of electric current that

the double-membrane system can achieve remains its chief limita-tion. The photovoltage would need to be magnified by more than another factor of two to reach the ~200 mV mark necessary to de-salinate seawater, a target that the researchers are optimistic about hitting.

"It all comes down to the funda-mental physics of how long the charge-carriers persist before recombining to form water," Ardo says. "Knowing the properties of water, we are able to more intelli-gently design one of these bipolar-membrane interfaces so that we can maximize the voltage and the current."

NOTABLENewsworthy

6

MARKETwatch


Globally, the electric ve-hicle market is poised for growth, estimated to grow 21 percent

over the next ten years. Automak-ers from A to Z are announcing dates when they expect produc-tion of their vehicles with internal combustion engines to transition to Hybrid Electric Vehicles (HEVs) and then completely to electric vehicles (EVs). Meanwhile, the infrastructure required to support this transition is building out at an increasing pace.

For instance, home charging sta-tions are being installed or aug-mented and public stations are being positioned in locations to enable seamless transportation. Phone apps and websites, showing the location of charging stations along a given route, are becoming vitally important, as the recent hur-ricanes in Texas and Florida make clear. We see what happens when a mass exodus of people-filled vehicles want to move from point A to point B at the same time.

The good news for us in power electronics is that EV applications drive greater electronics content overall. The trajectory of the EV and HEV electronics market ap-

Transitioning to Electric

By: Kevin Parmenter, PSD Contributor

pears to be taking the same path as the solar cell and inverter mar-ket: China is the driver. According to McKinsey & Company, China produced 43 percent of the world’s 873,000 electric vehicles in 2016 and currently enjoys having the largest fleet of EVs on the road. China also leads the world in the components needed for enabling EVs, which includes the appropri-ate Lithium-Ion battery cells and packs, EV motors and the power supplies required for the re-charg-ing stations.

As in many other alternative en-ergy areas the popularity of the implementation is proportional to the incentives being given to the countries investing in the technol-ogy; and the more the incentives and economics make sense, the more likely it is that people will invest in the technology and adopt it as a part of their lifestyle.

After China, Northern Europe, namely Norway and the Nether-lands, offer the greatest incen-tives, so it’s no surprise that the adoption rates in these countries compared to the other European countries is much higher. Over-all, the growth numbers are very high for EVs in developed coun-

tries with existing deployments of combustion-engine vehicles, since they have reasonable incentives for converting to electric.

Even with hybrid EVs as a stopgap towards complete electrification of the transportation market, the pro-jection is that by 2040, 54 percent of vehicle sales will be fully electric. The high growth percentages are to be expected in many cases, the rela-tive numbers were low to start with, the incentives are improving, costs are coming down, the infrastructure continues to expand and the tech-nology keeps getting better.

At the same time, battery and charging technology is improving and the price to obtain them are dropping. And thermal manage-ment and other power electronics technologies continue to expand and improve in inverters and motor drives.

This creates both a large market opportunity as well as a disruptor of markets as we replace oil as our primary transportation fuel. Mak-ing the right moves to position organizations for the future will be critical.

PSD

Power Path Management in Charger ICs By: Aaron Xu , Monolithic Power Systems

A commonly used power path manage scheme, dynamic power path management (DPPM),

is discussed in this article. The DPPM control loop adjusts the charge current dynamically based on the input source current capability and load current level to achieve a minimum charge time for a given source and system load. With DPPM, the system can obtain power immediately once the input source is applied, even with a deeply discharged battery. The system voltage regulation method is also discussed. In mobile devices with a rechargeable battery, a charger IC is needed to charge the battery when an external power source is applied. The system load inside the mobile device could be provided by the battery, the input source, or both, depending on the connection of the battery and system load. A power path management scheme is needed to handle this kind of power source selection.

Dynamic power path management (DPPM) is the most popular scheme for power path management in mobile applications. The basic power

stage structure for DPPM is shown in Figure 1.

In the DPPM system, the system load is connected to the system bus (VSYS). VSYS can be powered from the battery through the battery FET, or from the input source through a DC/DC convertor or LDO. When the input source is not available, the battery FET is fully on, so the battery provides power to the system load.

When the input source is applied, VSYS is regulated by the input DC/DC converter or LDO. Simultaneously, VSYS provides a charge current to the battery through the battery FET. In this charging mode, priority is given to the system load, and the remaining power is used for charging. The charge current is adjusted dynamically based

on input source capability and system load level, achieving a minimum charge time.

During the above charging process, if the system load is over the power capability of the input source, VSYS will drop. Once VSYS drops to a DPPM threshold, the DPPM control loop activates and reduces the charge current automatically to prevent VSYS from dropping further. This process is also called DPPM mode.

In DPPM mode, if the charge current is reduced to zero, and the system load is still over the input power capability, VSYS continues dropping. Once VSYS drops below the battery voltage (VBAT) level, the battery provides power to VSYS through the battery FET. This is called supplement mode. In supple-

Figure 1: NVDC Power Path Management Structure


DESIGN t ips

ment mode, the input source and battery provide power to the system simultaneously.

Before entering supplement mode, if the battery FET is in linear mode (not fully on, for ex-ample when VBAT < VSYS_MIN + DV, or during a start-up transient), to ensure a smooth transition in and out of supplement mode, an ideal diode mode is preferred to control the battery FET, such as the one in the MP2624A.

During the ideal diode mode, the battery FET operates as an ideal diode. When the system voltage is 40mV below the battery voltage, the battery FET turns on and regu-lates the gate driver of the battery FET. The voltage drop (VDS) of the battery FET is about 20mV. As the discharge current increases, the battery FET obtains a stron-ger gate drive and smaller on-state resistance (RDS) until the battery FET is completely turned on. When the discharge current goes lower, the ideal diode loop generates a weaker gate drive and larger RDS(ON) to keep a 20mV difference between the battery and system until the battery FET is turned off.

VSYS regulation in DPPM mode can be flexible depending on the system requirement. If the front-end converter from the input to the system is an LDO, VSYS can be set at a level to specially ben-efit the system requirement. For example, VSYS is 4.65V for the MP2661 and 5.0V for the MP2660.

If the front-end converter from the input to the system is a DC/DC converter, VSYS is usually set to follow the battery voltage to im-prove efficiency. This is commonly referred to as a narrow voltage DC (NVDC).

There are several advantages for DPPM control. First, the system gains power immediately once the input source is applied, regardless of whether the battery is depleted or not. Second, the charge current is adjusted dynamically based on the input source and system load to achieve a minimum charge time.

The limitation for DPPM control lies on the fact that it is complicat-ed to ensure a smooth transition between the different operation modes. Usually, a VSYS loop, ideal diode loop, charge voltage, and charge current loop are required for battery FET control.

Conclusion With DPPM control, the system can obtain power as soon as the input source is applied, even if the battery is depleted. The charger IC with DPPM control can also op-timize the charge current to fully utilize the input source current capability. Although the control for DPPM is complicated, DPPM is widely used in charger ICs that re-quire power source selection, such as in MPS’s MP2624A, MP2660, and so on.

Monolithic Power Systemswww.monolithicpower.com

LOWER REQUIREMENTS, SAME QUALITY, SAME ADVANTAGE.

Our new CMx series is specially designed for lower performance requirements, meaning that we now lead the competi tive environment in both high and low performance ranges for the first time.

It features, for example, an outstanding power value of 2 watts at a contact point temperature of 100°C.

NOW ALSO LEADING THE LOW RANGE!

LOW OHMIC PRECISION AND POWER RESISTORS

Isabellenhütte Heusler GmbH & Co. KGEibacher Weg 3 – 5 · 35683 Dillenburg, Germany

Phone +49 (0) 2771 934-0 · Fax +49 (0) 2771 [email protected] · www.isabellenhuette.de

Innovation by Tradition

7th-Generation “X Series” RC-IGBT Module for Industrial Applications

By: Akio Yamano, Misaki Takahashi, Hiroaki Ichikawa, Fuji Electric

Innovations miniaturize, reduce loss and improve reliability

Ever since it commercial-ized insulated gate bi-polar transistor (IGBT) modules in 1988, Fuji

Electric has contributed to minia-turization, cost reduction and per-formance improvement of power conversion equipment. It has done so through many technology in-novations to miniaturize, reduce loss and improve the reliability of IGBT modules. However, any further miniaturization of IGBT modules increases power density, which may lead to lower reliability due to an increase in operating temperatures of IGBTs and free-wheeling diodes (FWDs). Accord-ingly, to miniaturize IGBT modules while maintaining high reliability, technology innovation of chips and packages is essential.

Fuji Electric has carried out tech-nology innovation of chips and packages to commercialize the 7th-generation “X Series” IGBT module(1)(2). In addition, we have developed a reverse-conducting IGBT (RC-IGBT), which integrates an IGBT and a FWD into one chip, and in turn the 7th-generation “X Series” RC-IGBT module for in-dustrial use that incorporates the chip(3)(4). By applying the chip

technology of the 7th-generation X Series to optimize the chip structure, we have success-fully reduced the number of chips and the total chip area. At the same time, we have maintained the same generated loss as that of a combination of the X Series IGBT and X Series FWD. Furthermore, by combining the package technol-ogy of the 7th-generation X Series with the RC-IGBT, we have reduced thermal resistance and improved reliability. These technology inno-vations have led to a further power density increase and miniaturiza-tion of IGBT modules, which were impossible through conventional combination of IGBT and FWD.

2. Features

2.1 Features of the “X Series” RC-IGBT for industrial applicationsWith a conventional IGBT, a cur-rent is run only in the direction from the collector to the emitter by applying a voltage to the gate. An

inductor, which is used as a load of inverters widely in use as power conversion equipment, generates induced electromotive force in the direction to prevent any current change caused by the self-induc-tion effect. Accordingly, the current has a behavior of flowing in the same direction even if the IGBT is turned off, and running a current in the reverse direction required an FWD to be connected in antiparal-lel with the IGBT. Meanwhile, the “X Series” RC-IGBTs for industrial applications (X Series RC-IGBTs) achieve the same purpose with one element by using an RC-IGBT (see Figure 1).

Figure 2 shows a cross-section

Figure 1: X Series RC-IGBTs

COVER STORY


1110

COVER STORY



the surge voltage of the X Series RC-IGBTs is equivalent to that of combinations of the V Series IGBT and V Series FWD and of the X

Series IGBT and X Series FWD. The tail cur-rent is smaller than that of the combination of the V Series IGBT and V Series FWD and the turn-off loss Eoff is lower by 23% with no abnor-mal waveforms observed. The X Series RC-IGBTs use a thinner wafer than that of the combi-nation of the V Series IGBT and V Series FWD in order to improve the characteristics. Use of a thin wafer poses an issue of oscilla-tion at turn-off and withstand-ing voltage deg-radation. How-ever, with the X Series RC-IG-BTs, the specific resistance and the individual structures have been optimized to success-fully minimize

oscillation and withstand voltage degradation. As shown in Fig. 5 and Fig. 6, the current waveforms for the combination of the V Series

IGBT and V Series FWD have steep slopes but the X Series RC-IGBT realizes gentler current waveforms by optimizing lifetime control. Lowering the reverse recovery cur-rent peak Irrm and the tail current has reduced the reverse recovery loss Err by 20%.

2.3 Thermal characteristicsWith the X Series RC-IGBTs, an IGBT and an FWD has been inte-grated into one chip and the heat produced due to generated loss in the IGBT or FWD regions is radi-ated from the entire chip. Accord-ingly, reduction of thermal resis-tance can be expected. To further reduce thermal resistance, a new aluminum nitride (AlN) insulating substrate has been employed as the package technology of the 7th-generation X Series.

The junction-case thermal resis-tance is shown in Figure 7. The new AlN insulating substrate features approximately 45% lower thermal resistance as compared with Al2O3 insulating substrates based on the same chip size, which is a significant improve-ment. This has resolved the issue of a temperature rise caused by miniaturization of IGBT modules. Furthermore, by optimizing wire bonding and employing high-strength solder and high-thermo-stability silicone gel, high reliability has been ensured and while guar-anteeing continuous operation at 175 °C.

3. Power Density Increase

Figure 4: Turn-off waveforms of the X Series RC-IGBTs

Figure 5: Turn-on waveforms of the X Series RC-IGBTs

Figure 6: Current waveforms for the combination of the V Series IGBT

view of the X Series RC-IGBT. The X Series RC-IGBTs apply the chip technology of the 7th-generation X Series IGBTs and use a trench gate as the surface structure and a field stop (FS) layer as the back structure. As with the X Series IGBTs, the X Series RC-IGBTs have employed even smaller design rules as compared with the 6th-generation “V Series” IGBTs and optimized the surface structure. In this way, they have achieved a significant reduction of the collector-emitter saturation volt-age VCE(sat) that contributes to conduction loss. The latest thin wafer processing technology has also been applied to improve the trade-off relationship between the saturation voltage and turn-off switching loss. The X Series RC-IGBTs integrate FWD regions and have p-n junctions on the collector side. Accordingly, we have added the processes of patterning and

impurity layer formation on the back to form the p-type layer on the collector side of the IGBT and the n-type layer on the cathode side of the FWD on the back of the same chip. In addition, the trade-off relationship has been improved by lifetime control.

2.2 Electrical characteristics

Figure 3 shows the output char-acteristic of the 1,200 V X Series RC-IGBTs. The X Series RC-IGBTs are capable of outputting a cur-rent in both the forward direc-tion (IGBT) and reverse direction (FWD) with one chip. A saturation voltage lower than that of the V Series IGBTs has been realized by applying the chip technology of the 7th-generation X Series. With RC-IGBTs, electrons are injected into the cathode layer of the FWD region. This suppresses hole injec-tion from the collector layer of the IGBT and thus hinders conductiv-ity modulation. For that reason, snapback has been reported to occur (5)(6) in the low saturation voltage region. Meanwhile, with the X Series RC-IGBTs, snapback has been solved by optimizing the individual structures of the chip.

Turn-off waveforms of the X Series RC-IGBTs are shown in Figure 4, turn-on waveforms in Figure 5 and reverse recovery waveforms in Figure 6. Figure 4 indicates that

Figure 2: Cross-section view of the X Series RC-IGBT

Figure 3: Output characteristic of the 1,200 V X Series RC-IGBTs

1312

COVER STORY



ule footprint. And, by using the X Series RC-IGBT, thermal resistance Rth(jc) can be reduced by 27%. In this way, it covers the range of PrimePACK2, which uses the conventional V Series IGBT and V

Table 2: Dual XT and PrimePACK2* as lines of products with a rated voltage of 1,200 V

Table 3: Dual XT and PrimePACK2* FeaturesSeries FWD.

Figure 9 shows the calculation results of output current Io in in-verter operation and the maximum IGBT junction temperature Tjmax

for the Dual XT products re-spectively with a combination of the V Series IGBT and V Se-ries FWD, com-bination of the X Series IGBT and X Series FWD and the X Series RC-IGBT. In addition, using the X Series RC-IGBT can reduce power loss and the junction-case thermal resistance.By applying the package tech-nology of the

7th-generation X Series, the guaranteed continuous operating temperature has been increased from the conventional 150 °C to 175 °C. As a result, a high-er current density than before has been achieved with the same pack-age and even higher power density and miniaturization of IGBT mod-ules. In this way, it is possible to meet the requirements expected of IGBT modules such as miniatur-ization, loss reduction and higher reliability. In the future, we intend to continue working on technology innovation of IGBT modules and contribute to the realization of a sustainable society with improved energy management.

Fuji Electricwww.fujielectric.com

Figure 9: Calculation results of output current Io in inverter operation

and MiniaturizationTable 1 shows a comparison with the V Series IGBT module of 1,200 V/100 A and Figure 8 shows cal-

culation results of the power loss, junction temperature, Tj and junc-tion temperature variation ΔTjc for the respective modules. By

Figure 7: Junction-case thermal resistance

Table 1: Comparison with the V Series IGBT module of 1,200 V/100 A

Figure 8: Calculation results of the power loss, junction temperature

applying the chip technology and package technology of the 7th-gen-eration X Series, we have signifi-cantly reduced the power loss and thermal resistance as compared with the conventional combination of the V Series IGBT and V Series FWD. We have thus ensured high reliability and guaranteed continu-ous operation at 175 °C. In addi-tion, use of the X Series RC- IGBT makes it possible to reduce the number of chips and the total chip area, and miniaturization of IGBT modules can be expected. Based on these merits, applying the RC-IGBT chip technology and the chip technology and package technol-ogy of the 7th-generation X Series can produce a larger rated current than that of a conventional combi-nation of IGBT and FWD with the same package.

Table 2 illustrates Dual XT and PrimePACK2* as lines of products with a rated voltage of 1,200 V and Table 3 their features. Dual XT with a rated voltage of 1,200 V has the upper limit the rated current of 600 A for a combination of the V Series IGBT and V Series FWD. Through the use of the chip tech-nology and package technology of the 7th-generation X Series, the rated current has been increased to 800 A by combining the X Series IGBT and X Series FWD. Furthermore, adopting the X Series RC-IGBT provides a module with a rated current of 1,000 A using the same package. In comparison to PrimePACK2 that uses the V Series IGBT and V Series FWD, Dual XT offers a 40% reduction in the mod-

1514 WWW.POWERSYSTEMSDESIGN.COM WWW.POWERSYSTEMSDESIGN.COM

POWER SYSTEMS DESIGN 2017DECEMBERWIRELESS SENSORS

How to Maximize Your Wireless Sensor’s Runtime

By: Ajay Kuckreja, Andrew Brierley-Green, Nazzareno (Reno) Rossetti, & Ole Dreessen, Maxim Integrated

Battery runtime is a critical feature of remote wireless sensors

Battery runtime is a critical feature of remote wireless sensors. Wireless

sensors are becoming ubiquitous, penetrating the most diverse applications like home automation (Figure 1), activity monitors, remote sensor nodes, and tire pressure monitors, to mention a few broad categories. Runtime requirements, ranging from one to twenty years, pose a serious challenge to the design of these mobile devices. This design solution will first review a few cases of runtime challenges. Secondly, it will discuss a typical implementation of a wireless sensor system and its shortcomings, and lastly, it will present a new ultra-low-power, RF ISM transceiver that helps maximize battery runtimes for remote wireless sensors.

Runtime RequirementsAs an example, wireless magnetic window alarms and wireless locks are becoming more popular in homes and hotels. They are required to run for at least one year on a single battery. In other examples,

remotely located meter readers that monitor water and water leaks, earthquakes, gas, humidity, and temperature need to run autonomously for up to 20 years. In each case, the remote system needs to be small and inexpensive, running on very small batteries with limited capacity, often using only a fraction of an ampere hour (Ah). To meet such daunting runtime requirements, the designer must wisely use every Coulomb in the sensor battery.

Typical Sensor SystemA typical wireless magnetic window alarm sensor system (remote sensor subsystems

and central console) is shown in Figure 2. The 1.5V coin-cell battery powering the remote sensor subsystem is regulated by a boost converter. The remote sensor subsystem is in deep sleep mode most of the time, with the boost converter and the transceiver in shutdown. Whenever the magnetic sensor activates, the subsystem is awakened (pullup resistor R and transistor T) and an alarm signal is sent to the intelligent transceiver to process and transmit the alarm to the central console.

During the time the remote sensor is in deep sleep mode,

it is required to consume as little power as possible. During regular wake-up and infrequent transmission times, it can draw short bursts of current up to 100mA.

The wall-powered central console’s transceiver wirelessly communicates with each subsystem in the building or in the field. Its transceiver communicates with the main controller CPU, which is responsible for controlling the entire system, via UART.

Battery RuntimeLet’s assume that the coin-cell battery has a 150mAh capacity and the runtime requirement is 2 years. The math is very simple: the sensor operation must be managed so that its average current consumption does not exceed 8.5µA (8.5µA x 365d/y x 24h/d x 2y = 150mAh)! That’s not a lot to get by with.

Typical SolutionA typical solution available today is one that consumes a miserly 1µA in deep sleep, which still robs 12% from the overall battery runtime! It becomes evident that each component in the sensor needs to do much better to reduce the runtime loss to more acceptable levels. Ideally, if the remote sensor stays one order of magnitude below this level (100nA), then the runtime loss will be reduced

to a more acceptable 1.2%.

Ultra-Low Power TransceiverThe MAX7037 sub-1GHz, ultra-low-power, RF ISM transceiver can meet such a stringent requirement. Figure 3 shows its functional diagram.

The MAX7037 is a high-performance, quad band multichannel transceiver with an integrated 8051 microcontroller, flash memory, and sensor interface. It supports the standard ISM bands in the 300MHz to 930MHz range with output power up to 10dBm. The supply voltage runs from 2.1V to 5.5V, enabling it to cope with a variety of energy sources, such as solar cells, electromechanical, or thermoelectrical energy. In Figure 2, the MAX7037 takes the place of each “TRANSCEIVER + µC” block. Hardware-implemented transmit-and-receive routines enable a high-efficiency transceiver system for wireless fail-safe multiband/multichannel communication

Figure 1: Wireless Magnetic Window Alarm Sensor

Figure 2: Functional Diagram - Wireless Magnetic Window Alarm Sensor System

Figure 3: MAX7037 Functional Diagram

16 WWW.POWERSYSTEMSDESIGN.COM

WIRELESS SENSORS

with advanced FSK (frequency shift keying) and ASK (amplitude shift keying) protocol features. Sleep modes allow easy implementation of low-power applications with fast reaction times.

The Wireless Unit’s Ultra-Low Power SolutionIn an event-driven wireless sensor application, the MAX7037 is in deep sleep mode most of the time. In this mode, the current consumption is 100nA (max). When an event is sensed, the sensor signals the microcontroller, typically through an interrupt. A short data packet is formed by the microcontroller and transmitted at the specified frequency. The packet needs to contain some identifier of the sensor like a flag byte indicating its status. It also needs error detection or some sort of checksum or other error-detecting code that applies to the contents of the packet. In addition, the first byte of the packet is the number of bytes in the payload of the packet. Other than the requirement that the first byte indicates the length of the payload, there is no restriction on the contents of the packet payload.

The RF frequency used can be selected from one of the standard sub-GHz ISM frequencies; for example, 315MHz or 868MHz. Depending on which part of the world the system is deployed, the ISM

band from that region should be used. If there is a choice of frequencies, the most desirable is the lowest frequency used in buildings since lower frequency signals will propagate better through walls. However, interference considerations may dictate the choice of frequency.Using our earlier remote wireless magnetic window alarm sensor example, the MAX7037 reduces the runtime loss from 12% (3 months) down to 1.2% (9 days)!

nanoPower Boost ConverterThe MAX17222 nanoPower boost converter is also ideal for this application. With its 400mV of minimum input operation, it can draw the last drop of energy out of the 1.5V battery. It also provides a 0.5A peak inductor current limit and a selectable output voltage that uses a single standard 1% resistor. Its novel True Shutdown™ mode yields leakage currents in the nanoampere range (0.5nA typical), making this a truly nanoPower device.

Central ConsoleThe central console is always on. The main controller CPU periodically polls the co-located MAX7037 to check if a packet has been received. When the packet is received, the CPU confirms the contents are error-free, and if so, broadcasts an ACK (Acknowledgement) packet. When the MAX7037 in the remote subsystem receives

an ACK with its ID, it goes back into deep sleep mode. If the packet from the subsystem has an error, the main controller CPU will instead send a Negative Acknowledgement (NAK). If an ACK is not received within a certain timeout period or a NAK is received, the subsystem controller will resend the original packet. The handshaking protocol described is one possible alternative.

When using FSK modulation, the MAX7037 can transmit data up to 125kbps.

ConclusionWe discussed the extremely tough requirements for battery runtime in many wireless sensor systems. We reviewed a typical wireless application system which wastes 12% of battery life due to excessive leakage in deep sleep mode. We showed how the use of the MAX7037 reduced the runtime loss from 12% (3 months) down to 1.2% (9 days)! The MAX7037, thanks to its ultra-low, 100nA deep sleep current, cuts the leakage dramatically and helps achieve the long battery runtimes required by this class of applications.

In turn, the MAX17222 nanoPower boost converter, with its True Shutdown mode, provides the ideal power boost for the MAX7037 transceiver.

Maxim Integratedwww.maximintrgrated.com

Security is a Complex Topic and Requires Expertise and Solutions from all Product Segments

By: Bernd Hantsche, Managing Director Embedded & Wireless, Rutronik

The newly established Rutronik GDPR Expertise Team helps customers with security concepts, components and advice when implementing the regulation

May 25, 2018 is the deadline for implementing the European General

Data Protection Regulation (EU-GDPR) and the derived data protection legislation, which is the strictest seen to date. By this time, practically all companies processing personal data are required to implement extensive measures to protect this data.

“This confronts businesses with immense challenges,” explained Bernd Hantsche, Head of Division Embedded & Wireless at Rutronik. “Not only do they have to understand the regulation and its impact on their business, processes and products, they also have to establish suitable measures and implement them – a process that is anything but trivial with such a diverse and complex topic as security.”

For hardware and software developers, and for product

managers, Article 25 “Data Protection by Design and by Default” and Article 32 “Security of Processing” are of primary importance.

“These text sections leave a lot of questions unanswered, which is why we have established an inter-departmental team of experts with which we can help our customers find answers to questions such as these and develop extensive system concepts that are compliant with the GDPR,” said Hantsche.

EncryptionOne of the requirements of the regulation is for personal data to be encrypted in line with modern technical standards and implementation costs and in consideration of the nature, scope, circumstances and purposes of processing, as well as the probability of occurrence and severity of the risk.

“But which data is directly or indirectly personal? And what does ‘in line with modern technical standards’ even mean for individual components

DATA SECURITY


1918

DATA SECURITY



the right wireless protocol and mechanisms that detect malware as early as when booting the system make a criminal’s work harder.”

Speakers – another privacy riskWhether it’s the “ding ding bah” in fruit machines, background music in the elevator or supermarket or the cigarette dispenser that informs you that “your choice is currently not available” – speakers are being used in an increasing number of different devices. And fruit machine users or shoppers usually have their smartphones with them. This combination of factors can be exploited like a spy – some apps require access to the microphone during installation, which many users barely notice. But in spring 2017, the Braunschweig University of Technology discovered 234 apps in the Google Play store alone with Silverpush’s Ultrasonic Tracking Beacon function. Even if GPS localization is deactivated, this enables a smartphone to detect where its user currently is. Here, loudspeakers send high-frequency codes of between 18kHz and 20kHz – barely perceptible to the human ear, but easily receivable by the microphone in the smartphone – a clear violation of data protection legislation!

A security concept might include the following components in this situation: a TPM (Trusted Platform Module) that detects malware during the boot process, secure and encrypted communication with valid certificates and asymmetrical or hybrid key exchange, and a range of other security measures on all vertical software and communication layers. Outside of digital communication, it may also be advisable to employ a low-pass filter that filters higher frequencies and a loudspeaker that is optimized for frequencies beneath 18kHz, such as the AS09208AR-R from PUI Audio.

Rutronikhttps://www.rutronik.com/

and systems? Is asymmetrical encryption with RSA always necessary everywhere, or is AES encryption, ECC or the hybrid SSL/TLS encryption adequate in some cases?” said Hantsche, giving food for thought. There are similar ambiguities with the other regulations.

“If you look around in various online forums on this topic, you’re more likely to find people worried about not being able to satisfy the requirements rather than answers. Many of them anticipate a rush of cease-and-desist orders. With our team, we are able to not only provide our customers with relevant answers, but also offer adequate solutions.”

Rutronik’s GDPR Expertise Team includes specialists from the Storage Media, Wireless Communication, Embedded Boards, Embedded Systems, Security Modules, Microcontrollers, Displays and Sensors product segments. It advises developers and portfolio managers on how they can design their data transfer, data storage and data processing systems securely.

“We discuss with the customer to find out which critical security aspects there are in their specific application, what the nature of the potential risks is, and how severe each risk is,” said Team Leader Hantsche, describing the process. “Once these points

have been clarified, our team can develop a suitable GDPR-compliant system concept.”

This concept includes all components and systems that are in some way essential to security for the application. These must be precisely adapted to one another, as many of them are dependent on one another and can influence each other, which is why the various experts work closely together. As a broadline distributor, Rutronik has not only the components and systems but also the expertise needed to develop such concepts in full and offer its customers complete solutions – “the all-in-one carefree package”, so to speak.

In addition to personal advice, the team of experts is also working with the relevant manufacturers to prepare a comprehensive reference book containing all the fundamental knowledge and practical information needed for

components, technologies and complete applications.

Concentrating on the traditional aspects of data transfer, data storage and data processing isn't enough. For instance, social engineering is becoming an increasingly critical topic – and device manufacturers are expressly advised to take this seriously.” Social engineering involves criminals sneaking a look at PIN codes or passwords simply with a pair of binoculars or stealing keys or RFID transponders to open doors or authorize the use of devices. This enables them to circumvent any PIN protection and any difficult-to-crack password.

“But even against social engineering, there are remedies, for instance biometric retinal scanners and fingerprint sensors, 3D camera systems for facial recognition, or at least special displays with a particularly small viewing angle,” said Hantsche. “Even choosing

TDK TechnologyDriving new mobility.

https://product.tdk.com

DC-DC converters and on-board chargers in industry-leading small size

Rare earth and ferrite magnets for high-effi ciency motors

Pressure sensors and TMR angle and motion sensors for high-precision applications

Aluminum electrolytic capacitors and MLCCs with high reliability

in industry-leading small size

DU171103_TDK_T122_AUTOMOTIVE_E_PSD_Europe_Mag_102x273.indd 1 25.09.17 11:34


POWER SYSTEMS DESIGN 2017DECEMBERPACKAGING

Packages Enable Full SiC Performance With Limited Parasitics

By: Courtney R. Furnival, President, Semiconductor Power Solutions

EV Motor Controller Inverters are limited by conventional power

Electric Vehicle (EV) power electronics offer the largest growth potential to the Power

Semiconductor Market over the next 5-10 years. Today’s EV Motor Controller Inverters are limited by conventional power silicon (Si) IGBT and diode performance, including their higher power density, higher efficiency, lower thermal resistance and smaller/lighter Motor Controllers. SiC MOSFETs can provide higher power density and efficiency, yet with SiC power density increasing faster than its efficiency, it becomes more difficult to maintain and reduce the thermal resistance in the smaller areas. Smaller Inverters built with SMD SiC devices allow simpler mechanical and thermal hardware, further reducing the complete Motor Controller size and weight. Ultra-thin double-sided (DS) surface mount (SMD) modules can provide the required SiC electrical and thermal solutions. The DS-SMD package advantages can be optimized in 650V/650A SiC Half-Bridges (HB), which can be only 29mm x 17.5mm and 0.75mm (0.030”) thick. The

smaller DS-SMD power packages can enable today’s SiC EV Motor Controllers, and ever increasing SiC power density and efficiency, and at lower costs.

Ultra-Thin DS-SMD Packaging Approach and AdvantagesThe approach of using new ultra-thin DS-SMD packages enable full SiC performance with virtually no package limiting parasitics. These packages eliminate redundant interfaces and terminals, and significantly reduce thermal resistance per mm2 with double-side cooling and high current connections and low profile packages. The DS-SMD packages are leadless and wirebondless, significantly reducing package inductance and resistance. The low current-loop inductance can be 0.1 to 0.20nH, reducing voltage over-shoot, losses and noise, and can lower required SiC MOSFET and diode break-down voltages. Lower BV SiC can be optimized for lower Rds(on). The compact DS-SMD architecture is ideal for paralleling power switches die, further reducing Rds(on) and increasing current capabilities. Smaller SiC die reduce CTE

stresses, allowing direct die-to-Cu soldering. Grouped smaller SiC die can further reduce the thermal resistance with wider heat-cones than single large die. Smaller SiC die improve yields and minimize higher costs created by today’s SiC higher defect densities and smaller wafers. Compact DS-SMD packages enable simpler and smaller Inverter and Motor Controller architectures, by eliminating long internal leads and redundant internal electrical & thermal interfaces, and can better accommodate heatsink (and cold-plate) isolation and structure. Three compact (29mm x 17.5mm x 0.75mm) DS-SMD HB packages can create an Inverter for a 200kW EV Motor Controller.

Unique DS-SMD Module ArchitectureThe DS-SMD HB technology is basically a double-sided power QFN package built with a one piece leadframe. The bottom-side power die are soldered into cavities and those die are exposed at the bottom-side of the package for direct soldering to the positive and negative

bus bars, while leaving the top-side of the leadframe available to solder other components like bump-chip gate-drivers and chip capacitors die This double-sided assembly approach works for smaller HEV Inverters, with integrate gate-drivers and passive components on the top-side. Even higher power EV inverters are possible by using the top surface to attach an additional top-side heatsink/cold-plate, creating a single-sided assembly with double-side heatsinks and high current external pads. The type of unique DS-SMD architecture is shown in a HB example using the proprietary µMaxPak technology.

A cross-section of that half-bridge (HB) package with

double-sided cooling is shown in Figure 1a. The HB packages are soldered directly onto the positive and negative bus bars, and an output lead is soldered directly onto the top of each HB package. The package is only 0.75mm (0.030”) thick, and the top and bottom pads provide all electrical and thermal connections. Since the power SiC die are soldered into the leadframe bottom-side cavities and to the copper bus bar and output leads, there are no HB package leads or wire bonds, eliminating associated inductance and resistance. This can reduce current-loop inductance to 0.1-0.2nH, and eliminates redundant lead/terminal interfaces. The power density and inductance are

further reduced by co-packaging the hi-side and low side switches in the HB package. The internal layout of the 650V/650A HB is shown in the Figure 1B X-ray View of the HB package. It demonstrates internal details and interconnects. The Figure 1B HB example uses two 8mm x 7mm SiC MOSFETs & two 4mm x 7mm SiC Diodes for both high-side and low-side switches. The high-side die drains sit directly on the positive bus and the low-side die source sit directly on the negative bus. The output terminal sits directly onto the common leadframe pads for both hi-side sources and low-side drains. The gate & sense (g/s) pad locations are shown in the Figure 1b X-ray View, but vertical structures are not shown. These will be dependent on the type of associated structure and components. Current sense leads are not shown, but when required, they may be implemented just like the g/s leads. All power package pads are soldered directly to the bus or output leads, eliminating

Figure 1a: Cross- section

Figure 2a: Top-Side pads Figure 2b: Bottom-Side pads

Figure 1b: XRay top-view HB


PACKAGING

higher thermal resistance press contacts. The positive and negative bus provides both electrical connections, and a thermal path to the bottom-side heatsink. The output pad shown in Figure 2a provides the high-current electrical connection to one phase of the motor, and the thermal paths to the top-side heat-sink. The bottom-side positive and negative pads shown in Figure 2b provide the DC input voltage from the bus bars, and the thermal path to the bottom-side heatsink. The heavy copper bus bars and output lead provides excellent heat spreading and heat transfer from the DS-SMD packages to the heatsinks, but since they are electrical connections they must be electrically isolated from the

heatsinks. The heavy copper leads also provide higher heat capacity at the die/packages, minimizes junction temperature (Tj) spikes, and reducing power die dynamic thermal resistance (Rjs) to the heatsinks/cold-plates. This maintains more uniform average junction temperature with 50% HB duty cycle at lower motor RPM. The isolation to heatsinks must be provided to the Inverter with DBC substrates, like aluminum nitride (AlN) or silicon nitride (Si3N4) substrates. They are placed between the power leads and the heatsinks, and ideally soldered. Since the DS-SMD package are small with SiC devices, the ceramic isolators can be much smaller than in conventional power Si IGBT modules. This minimizes breakage from CTE mismatched and mechanical stresses. Breakage of the ceramic isolators can be caused by the substrate warpage and excessively large area substrates. Smaller substrates can improve reliability and reduce ceramic costs, and thinner high-strength ceramic can further lower thermal

resistance, improving the performance, especially Rds(on) and reliability. For example, if an AlN isolating ceramic thickness is reduced from 0.025” to 0.015”, and the AlN is 50% of the total thermal resistance (Rjs),

the junction-to-case thermal resistance (Rjs) is reduced by 20%. This is very significant for high power EV Inverters, where we fight for an extra 1% or 2% improvement.

Other Motor Controller Advantages Enabled by DS-SMD HB and InverterThe Motor Controller structure is outside of the scope of this article, but it is important to briefly discuss how to use the DS-SMD HBs to best advantage in the Motor Controller. The exact structure will vary with power level, gate-drivers, protection circuitry, and other specific components, circuitry, functions and specifications. A basic Inverter interconnect structure is shown in Figure 3a and Figure 3b, and focuses on optimizing the electrical and thermal connections, and the locations of the DBC isolators described above. This optimization is essential to capitalizing on DS-SMD package driven performance, reliability and cost advantages listed below:• Smaller SiC die and smaller

DS-SMD packages make EV Motor Controller much smaller and lighter

• SiC MOSFETs and diodes can increase Motor Controller efficiency from 97 to between 99.0% and 99.5% of those built with Si IGBT modules

• With total power dissipation being reduced by a factor of greater than 3, there are proportional reductions in heatsink size and costs, and in the off-line cooling system’s size and cost

• One cannot overlook the significance of secondary performance and reliability advantages of lower operating temperatures (Tj)

• Smaller Inverters further reduce the system current-loop inductance with shorter bus-bars and interconnects, in addition to inductance reductions within the HB DS-SMD HB packages

• Reductions of Inverter’s mechanical, electrical and thermal interfaces further reduce the complete Motor Controller size and weight, improving reliability and reducing overall system costs

• Small and closer conductors and components must be managed for high-voltage spacing, which usually means potting, coating, under-fill, etc. These must meet safety agency rules and regulations, which can be better implemented inside the Motor Controller enclosure. It is important to use

Figure 3a: 3P - Inverter x-ray view

Figure 3b: Inverter HB cross-section

materials and spacing that comply with UL/EN Pollution Degree 1 regulations. These minimum HV spacings and materials may be new to PCB assembly, but have been used successfully for many years inside high power IGBT modules.

• Advantages of using thinner and smaller ceramic isolators have been described earlier, but such ceramic isolators must also be mechanically support to not create excessive pressure and twisting in the larger Motor Controller system assemblies

In Summary• High density efficient power

SiC enables ultra-thin SMD modules at EV power levels

• Ultra-thin DS-SMD packages allow SiC devices to operate at their full performance and efficiency potential

• DS-SMD SiC modules are much smaller and lighter than conventional Si IGBT modules, and enable even larger reduction in the EV Motor Controller system size and weight

• Ultra-thin DS-SMD packages are built using conventional QFN technologies. This technology offers easy customization, short time-to-market, and reliable low cost packaging.

Semiconductor Packaging Solutions

Reliable. Available. Now.

www.tracopower.com

Ultra-compact, fully regulated and isolated 1 and 3 Watt DC/DC converters.

TRN 1(SM) und TRN 3(SM) Series• Package occupies only

1 cm2 on the PCB• Available in SIP package or

for SMD mount• Wide input voltage ranges up

to 75 VDC includes the low voltage range 4.5–13.2 VDC

• Operating temperature range –40 °C to +90 °C

• I/O-isolation 1600 VDC, continuous short circuit protection

USA_EU_en_TRNSeries_PSD_68x273.indd 1 05.10.17 20:55

2524

POWER SUPPLIES



Lighting - from stadiums to vegetables

By: Patrick Le Fèvre, Powerbox

Could lighting help feed 10 billion people in 2050?

The lighting market seg-ment is very diversified, but Solid State Lighting (SSL), based on LEDs,

has made the Edison bulb effec-tively obsolete and is now infring-ing on areas where fluorescent lighting excels. The possibilities offered by SSL bring advantages in industrial applications, such as for roads and parking lots, stadiums and stages (Figure 1), as well as for urban farming, horticulture, water purification, and medical lighting and light therapy.

When electrons meet photons, it requires power designers to work very closely with LED manufactur-ers. A good example of this inter-section is “GaN (Gallium Nitride) lighting”, which uses GaN transis-tors in the power stage, and GaN-on-Silicon in the LED element. It reflects the industrial maturity of GaN in both the power and light-ing industries. For power design-ers, it is interesting to follow both technologies and there could be huge benefits in that association.

SSL lighting currently dominates some segments of the lighting market where the cost of replacing a light bulb is inordinately expen-

sive and could cost the end user more than the light itself. Lights on tall poles where a lift truck is required to even reach the lighting fixture is an exam-ple of this. Having to stop or reroute traffic on a bridge or in a tunnel is another example. These types of ap-plications benefit from having very long lasting SSL Additionally, SSL is much more efficient than the typical high pressure light that it is replacing, so the power consumption for providing the same level of light is notably reduced, often resulting in a very good ROI for the end user as well as contributing to reduce energy consumption.

Houston's NRG Stadium (USA) became one of the first profes-sional venues to use energy-efficient LED lights in 2015. The field was illuminated exclusively by 65.000 LED lights, consuming 337 kilowatts when at full power, which

is about 60 percent less energy than the previous conventional stadium-lighting system. Future technologies will save even more energy, with 75% saving the target. The energy saved is remarkable, but it is only the beginning of what is possible when combining SSL technology and efficient power management. Larger infrastructure providers and cities are renovating their lighting systems in favour of digitally controlled SSL, and com-bined with renewable energy, we are approaching the mythic zero

emissions lighting circle (from generation to utilisation).

Lighting for food!Right now, there are around 7.6 bil-lion people in the world and every year, the world’s population is expanding by 83 million people. By 2050, there will be a global popula-tion of just short of 10 billion. Pro-viding food to 10 billion people will require agriculture to develop very efficient production processes, while preserving the environment by reducing hazardous chemicals and optimizing water utilisation. In a recent communication, the World Bank highlighted the situ-ation and laid out the expected future of food supply around the world. “The world needs to pro-duce at least 50 percent more food to feed 10 billion people by 2050. But climate change could cut crop yields by more than 25 percent. The land, biodiversity, oceans, forests, and other forms of natu-ral capital are being depleted at unprecedented rates. Unless we change how we grow our food and manage our natural capital, food security – especially for the world’s poorest – will be at risk.”Considering all the parameters and requirement to produce food with the highest respect for the environment, in 1999, Dr. Dickson Despommier with his students developed the idea of modern indoor farming, revitalizing the term coined in 1915 by the Ameri-can geologist Gilbert Ellis Bailey: “Vertical farming.” There has been a lot of articles about industrial buildings being converted into

vertical farms, and there are an amazing number of technology in-novations contributing to optimize the energy delivered to the plants for optimal growth. From space utilization, there is 100 times more food produced per square meter compared with regular agriculture. Water utilization is reduced by 90% and hazardous chemicals are eliminated. Indoor farming is very attractive, though to be really efficient it requires a very efficient lighting system (Figure 2).

Not all vegetables can be grown with limited soil and nutrition, but for the ones that can be grown using this farming method, the

results are impressive and getting even more so when using mod-ern computer-controlled lighting technologies. For power designers this is a very interesting challenge to explore, combining advanced power electronics and modern agriculture with software.Since its introduction, indoor farming engineers have conducted research to validate the spectrum and energy required by different plants to grow efficiently. From wide spectrum fluorescent or halogen lamps to more narrow spectrum lighting, the conven-tional lighting industry innovated a lot but those technologies are not flexible nor efficient enough to

Figure 1: Solid State Lighting saving more than 60% energy compared to conventional lighting in Houston

Figure 2: Solid State Lighting growing vegetables in indoor farming

Figure 3: The Photosynthetic Photon Flux Density required to grow plantstypically starts at 450 nm (blue light) and goes through 730 nm (far red)

26

POWER SUPPLIES


respond to the demand.

Following experiments in Japan from 2005-2008, agronomical researchers investigated differ-ent lighting methods to adjust the spectrum and energy delivered to specific plants. Researchers concluded that the specific light spectrum to grow plants and vegetables typically starts at 450 nm (blue light) and goes through to 730 nm (far red) (Figure 3). The Photosynthetic Photon Flux Den-sity (PPFD) required ranges from 50 micromoles (µmol) for mush-rooms up to 2,000 µmol for plants like tomatoes and some flowers that thrive in full summer light (Figure 4). Agricultural experts tell us that for optimal results different

plants types may require different light spectra as well as differing light balance and intensities be-tween the seedling and harvest-ing stage. This often results in a requirement for the artificial light to have a number of different spec-tra channels that are individually adjustable for intensity.

Urban farms are increasingly moving to modern SSL lighting, especially as the amount of light energy per watt of power steadily increases. This increased efficacy also lowers the cooling costs as the produce yield is negatively impacted by too high air or soil temperatures. LED lighting allows the grower to use lights that only consume energy in the spectra

that the plants require, generally red and blue, thus saving energy over delivering full spectrum lighting, where the majority of the light is not used by the plants. This goes back to the elementary school question, “Why are plants green?” (because most plants don’t absorb (use) green light and thus it is reflected back to your eyes making the plant look green).

Today multiple LED lamps are commonly used to energy efficient-ly grow vegetables though more advancement can be achieved by integrating intelligent power sources in LED modules. One re-search area is to create micro-LED panels with growth index monitor-ing, which are able to modulate the light locally (1/2 square meter area). That will require a very ef-ficient distributed power solution that is able to adjust all the param-eters of the vegetable’s growth.

Despite a lot of articles and pa-pers presented at conferences, indoor farming is still at its infant stage, though the demand on the agriculture to produce more with less environmental impact is an important factor in its develop-ment. The combination of the latest technologies in SSL, power management, and a software con-trolled environment will improve productivity and tools for modern farmers to be in position to grow food for the 10 billion people in 2050.

Powerboxhttp://www.prbx.com

Figure 4: Light energy required ranges from 50 micromoles (µmol) for mushrooms up to 2.000 µmol for light intensive plants

Special Report:Electric & Hybrid Vehicles

Inside:

Maximize Run Time in Automotive Battery Stacks Even as Cells Age...

Automakers Shift to 48v Mild Hybrid Systems...

Advances in Substrate Technology Enable New Power Modules for Automotive Application...

Power Film Capacitor Design for EV and HEV Applications...

28

33

37

40

2928

SPECIAL REPORT : ELECTRIC & HYBRID VEHICLES



Maximize Run Time in Automotive Battery Stacks Even as Cells AgeBy: Samuel Nork. GM, Battery Chargers, ASSPs & PMICs & Tony Armstrong, Director of Product Marketing, Power Products, Analog Devices Inc.

All Series-Connected Cells Need to Be Balanced

Large battery stacks consist-ing of series-connected, high energy density, high peak power Lithium poly-

mer or Lithium-Iron Phosphate (LiFePO4) cells are commonplace in applications ranging from all-electric (EV or BEV) and hybrid gas/electric vehicles (HEVs and plug-in hybrid electric vehicles or PHEVs) to energy storage systems (ESS). The electric vehicle market in particular is forecast to create tremendous demand for large ar-rays of series/parallel connected battery cells. The 2016 global PHEV sales were 775,000 units [Source: EVvolumes.com], with a forecast of 1,130,000 units for 2017. De-spite the growing demand for high capacity cells, battery prices have remained quite high and represent the highest priced component in an EV or PHEV with prices typically in the $10,000 range for batteries capable of a few 100s of kilometers of driving range. The high cost may be mitigated by the use of low cost/refurbished cells, but such cells will also have a greater capac-ity mismatch, which in turn reduc-es the usable run time, or drivable distance on a single charge. Even the higher cost, higher quality cells

will age and mismatch with repeat-ed use. Increasing stack capacity with mismatched cells can be done in two ways: either by start-ing with bigger batteries, which is not very cost effective, or by using active balancing, a new technique to recover battery capacity in the pack which is quickly gaining momentum.

All Series-Connected Cells Need to Be BalancedThe cells in a battery stack are balanced when every cell in the stack possesses the same state of charge (SoC). SoC refers to the current remaining capacity of an individual cell relative to its maxi-mum capacity as the cell charges and discharges. For example, a 10A-hr cell with 5A-hrs of remain-ing capacity has a 50% SoC. All battery cells must be kept within a SoC range to avoid damage or lifetime degradation. The allow-able SoC min and max levels vary from application to application. In applications where battery run time is of primary importance, all cells may operate between a min SoC of 20% and a max of 100% (or a fully charged state). Applica-tions that demand the longest

battery lifetime may constrain the SoC range from 30% min to 70% max. These are typical SoC limits found in electric vehicles and grid storage systems, which utilize very large and expensive batteries with an extremely high replacement cost. The primary role of the bat-tery management system (BMS) is to carefully monitor all cells in the stack and ensure that none of the cells are charged or discharged be-yond the min and max SoC limits of the application.

With a series/parallel array of cells, it is generally safe to assume the cells connected in parallel will auto-balance with respect to each other. That is, over time, the state of charge will automatically equal-ize between parallel connected cells as long as a conducting path exists between the cell terminals. It is also safe to assume that the state of charge for cells connected in series will tend to diverge over time due to a number of factors. Gradual SoC changes may oc-cur due to temperature gradients throughout the pack or differences in impedance, self-discharge rates or loading cell to cell. Although the battery pack charging and

discharging currents tend to dwarf these cell-to-cell variations, the accumulated mismatch will grow unabated unless the cells are peri-odically balanced. Compensating for gradual changes in SoC from cell to cell is the most basic reason for balancing series connected batteries. Typically, a passive or dissipative balancing scheme is adequate to re-balance SoC in a stack of cells with closely matched capacities.

As illustrated in Figure 1a, passive balancing is simple and inexpen-sive. However, passive balanc-ing is also very slow, generates

unwanted heat inside the battery pack and balances by reduc-ing the remaining capacity in all cells to match the lowest SoC cell in the stack.

Passive balancing also lacks the ability to effectively address SoC errors due to another common

occurrence, capacity mismatch. All cells lose capacity as they age and they tend to do so at different rates for reasons similar to those listed above. Since the stack cur-rent flows into and out of all series cells equally, the usable capacity of the stack is determined by the low-est capacity cell in the stack. Only active balancing methods such as those shown in Figures 1b and 1c can redistribute charge throughout the stack and compensate for lost capacity due to mismatch from cell to cell.

Cell-to-Cell Mismatch Can Dra-matically Reduce Run Time Cell-to-cell mismatch in either ca-pacity or SoC may severely reduce the usable battery stack capac-ity unless the cells are balanced. Maximizing stack capacity requires that the cells are balanced both during stack charging as well as stack discharging.

In the example shown in Figure 2, a 10-cell series stack comprised of (nominal) 100A-hr cells with a

Figure 1a(top) 1b (bottom left) 1c (bottom right): Typical Cell Balancing Topologies

Figure 2: Stack Capacity Loss Example Due to Cell-to-Cell Mismatch

3130




+/- 10% capacity error from the minimum capacity cell to the maxi-mum is charged and discharged until predetermined SoC limits are reached. If SoC levels are con-strained to between 30% and 70% and no balancing is performed, the usable stack capacity is reduced by 25% after a complete charge/discharge cycle relative to the theo-retical usable capacity of the cells. Passive balancing could theoreti-cally equalize each cell’s SoC dur-ing the stack charging phase, but could do nothing to prevent cell 10 from reaching its 30% SoC level before the others during discharge. Even with passive balancing during stack charging, significant capacity is lost (not usable) during stack discharge. Only an active balanc-ing solution can achieve capacity recovery by redistributing charge from high SoC cells to low SoC cells during stack discharging.

Figure 3 illustrates how the use of ideal active balancing enables 100% recovery of the lost capac-ity due to cell-to-cell mismatch. During steady state use when the stack is discharging from its 70% SoC “fully” recharged state, stored charge must in effect be taken from cell 1 (the highest capacity cell) and transferred to cell 10 (the lowest capacity cell) – otherwise cell 10 reaches its 30% minimum SoC point before the rest of the cells, and the stack discharg-ing must stop to prevent further lifetime degradation. Similarly, charge must be removed from cell 10 and redistributed to cell 1 during the charging phase – other-

wise cell 10 reaches its 70% upper SoC limit first and the charging cycle must stop. At some point over the operating life of a bat-tery stack, variations in cell aging will inevitably create cell-to-cell capacity mismatch. Only an active balancing solution can achieve capacity recovery by redistributing charge from high SoC cells to low SoC cells as needed. Achieving maximum battery stack capacity over the life of the battery stack re-

quires an active balancing solution to efficiently charge and discharge individual cells to maintain SoC balance throughout the stack.

High Efficiency Bi-Directional Balancing Provides Highest Capacity RecoveryThe LTC3300-2 (see Figure 4) is a new product designed specifi-cally to address the need for high performance active balancing. The LTC3300-2 is a high efficiency, bi-

directional active balance control IC that is a key piece of a high perfor-mance BMS sys-tem. Each IC can simultaneously balance up to 6 Li-Ion or LiFePO4 cells connected in series.

SoC balance is achieved by redistributing charge between a selected cell and a sub-stack of up to 12 or more adjacent cells. The balancing decisions and balancing algorithms must be handled by a

Figure 3: Capacity Recovery Due to Ideal Active Balancing

Figure 5: LTC3300-2 Power Stage Performance

Figure 4: LTC3300-2 High Efficiency Bi-Directional Multicell Active Balancer

separate monitoring device and system processor that controls the LTC3300-2. Charge is redistributed

from a selected cell to a group of 12 or more neighboring cells in or-der to discharge the cell. Similarly,

International Exhibition and Conferencefor Power Electronics, Intelligent Motion,Renewable Energy and Energy ManagementNuremberg, 5 –7 June 2018

»Power Electronics is your Passion? If so, PCIM Europe is your platform - exhibit in 2018! pcim-europe.com/exhibitors

Information: +49 711 [email protected] # pcimeurope

32



charge is transferred to a selected cell from a group of 12 or more neighbor cells in order to charge the cell. All balancers may operate simultaneously, in either direction, to minimize stack balancing time. The LTC3300-2 has an SPI bus compatible serial port. Devices can be connected in parallel, using digital isolators. Multiple devices are uniquely identified by a part ad-dress determined by the A0 to A4 pins. On the LTC3300-2, four pins comprise the serial interface: CSBI, SCKI, SDI and SDO. The SDO and SDI pins may be tie together, if desired, to form a single bidi-rectional port. Five address pins (A0 to A4) set the part address. All serial communication related pins are voltage mode with voltage levels referenced to the VREG and V– supplies.

Each balancer in the LTC3300-2 uses a non-isolated boundary mode synchronous flyback power stage to achieve high efficiency charging and discharging of each individual cell. Each of the six bal-ancers requires its own transform-er. The primary side of each trans-former is connected across the cell to be balanced, and the secondary side is connected across 12 or more adjacent cells, including the cell to be balanced. The number of cells on the secondary side is limited only by the breakdown volt-age of the external components. Cell charge and discharge cur-rents are programmed by external sense resistors to values as high as 10+ amps with corresponding scaling of the external switches

and transformers. High efficiency is achieved through synchronous operation and the proper choice of components. Individual balancers are enabled via the BMS system processor and they will remain enabled until the BMS commands balancing to stop or a fault condi-tion is detected.

Balancer Efficiency Matters!One of the biggest enemies faced by a battery pack is heat. High am-bient temperatures rapidly degrade battery lifetime and performance. Unfortunately, in high current bat-tery systems, the balancing cur-rents must also be high in order to extend run times or to achieve fast charging of the pack. Poor balanc-er efficiency results in unwanted heat inside the battery system, and must be addressed by reducing the number of balancers that can run at a given time or through expen-sive thermal mitigation methods. As shown in Figure 5, the LTC3300-2 achieves >90% efficiency in both the charging and discharging di-rections, which allows the balance current to be more than doubled relative to an 80% efficient solu-tion with equal balancer power dissipation. Furthermore, higher balancer efficiency produces more effective charge redistribution, which in turn produces more ef-fective capacity recovery and faster charging.

ConclusionWhile new applications such electric and PHEVs that are grow-ing rapidly, consumer expectations for long operating life and reliable

operation remain unchanged. Automobiles, whether they be battery or gasoline powered, are expected to last for over 5 years without any perceptible degrada-tion in performance. In the case of an EV or PHEV, performance equates to drivable range under battery power. EV and PHEV sup-pliers must provide not only high battery performance, but also a multi-year warranty, guarantee-ing a minimum range to remain competitive. As the number and age of electric vehicles continues to grow, irregular cell aging within the battery pack is emerging as a chronic problem and primary source of run time reduction. The operating time of a series-connected battery is always limited by the lowest capacity cell in the stack. It only takes one weak cell to compromise the whole battery. For the vehicle suppliers, replac-ing or refurbishing a battery under warranty due to insufficient range is a very expensive proposition. Preventing such a costly event can be accomplished by using larger, more expensive batteries for each and every cell, or by adopting a high performance active balancer such as the LTC3300-2 to compensate for cell-to-cell capacity mismatch due to non-uniform aging of the cells. With the LTC3300-2, a severely mismatched stack of cells has nearly the same run time as a perfectly matched stack of cells with the same average cell capacity.

Analog Deviceswww.analog.com

Automakers Shift to 48V Mild Hybrid Systems

By: Ed Kohler, Intersil Corporation, a Renesas company

Regulatory CO2 emissions and fuel economy requirements are driving the adoption of 48V mild hybrid systems.

Worldwide auto-makers have begun releasing vehicles with a

new 48-volt (48V) power subsys-tem, also known as board net. The first announcements came from European OEMs, like Audi, who announced a 48V system with its SQ7 TDI model. The Audi 48V sys-tem powers an electric compres-sor that eliminates turbo lag and an electromechanical active roll stabilization system that improves ride and handling. The 48V initia-tive has now gone global. Chinese OEMs including Geely and FAW Group announced they will launch 48V mild hybrids in the coming years, and Italian-American OEM Fiat Chrysler has made a similar announcement. Korean OEMs Hyundai and SsangYong have also demonstrated 48V mild hybrids.

Legislation Drives Mild Hybrid Rollouts48V subsystems do introduce an incremental vehicle cost, but the widespread acceptance shows that automakers acknowledge it as a means to achieve lower CO2 emis-sions, improved fuel economy, and vehicle drivability enhancements.

Regulatory CO2 emissions and fuel economy requirements are driving the adoption of 48V mild hybrid systems. The penalties are severe for OEMs that do not meet these legislated targets. Figure 1 demonstrates the CO2 emission limits enacted worldwide. In the U.S., Europe, China, and Korea, automakers will need to achieve a 22-36% CO2 emissions reduc-tion in their fleets over the next 10 years to comply with the current legislation.

Figure 2 shows the fuel economy targets. In order to meet these

targets, automakers must improve fuel economy by 24-49% over the same 10-year time period. While the regulations are getting tighter, there is a further push worldwide to make the testing cycles more representative of real-world driving conditions. As test methodologies improve, it is expected that auto-makers will find it even more diffi-cult to meet the regulation targets.

In light of the increasingly strin-gent regulations and higher penal-ties for non-compliance, vehicle manufacturers are highly moti-vated to develop technologies that

Figure 1: CO2 legislation




SPECIAL REPORT : ELECTRIC & HYBRID VEHICLES POWER SYSTEMS DESIGN 2017DECEMBER

tors include the addition of the 48V starter/generator, which significantly improves the start-stop system over the standard 12V implementation. The 48V starter reduces engine start time to just a couple hundred milliseconds from as much as 1 to 2 seconds in the traditional 12V system. This allows the engine to stop in cases where the need for quick acceleration or immediate restart would have required the 12V start-stop system to leave the engine running. The added capacity of the 48V bat-tery also allows the engine to be stopped for longer durations while still maintaining functions such as climate control and infotainment. In addition, the 48V starter can provide a boost or torque assist that reduces load on the combus-tion engine to improve efficiency

under acceleration. The 12-15kW starter/generator is able to capture more of the kinetic energy of the vehicle during braking than is possible with a typical 2kW alternator and the Li-ion battery is better able to quickly store the recovered energy, significantly improving the effectiveness of regenerative braking.

Engine Downsizing Improves DrivabilityAutomakers also realize another significant improvement, without affecting vehicle drivability, by downsizing the combustion engine and adding a 48V supercharger to provide additional power boost for better acceleration. Standard 12V turbo-charged vehicles also make use of engine downsizing to improve efficiency, but the turbo-

lag associated with a mechani-cally driven system and the lack of available boost at low engine RPMs results in a less than satis-factory driving experience. Using a 48V supercharge provides near instantaneous boost at any engine RPM, eliminating the negative impact on drivability. And there are additional benefits from weight reduction when 12V systems are converted to 48V. Typically, the wir-ing harness supplying power can achieve a 4x reduction in current required to supply the same power, and motors running off 48V are significantly lighter than with the traditional 12V system. There is a small negative impact due to the added weight of the new 48V sys-tems (where they do not replace an existing 12V device); however, this impact does not diminish the overall improvements.

While the 48V system will reduce CO2 emissions and improve fuel economy, automakers are also banking that consumers will value the improvements in drivability it offers. Audi has demonstrated that when us-ing a 48V electric compressor, they are able to increase off-the-line acceleration. The electric compressor spins up to 70,000 RPM in 250ms, providing torque-boost almost instantaneously, even at low engine rpms. With a 48V starter-generator, the engine starts in less time and with less vibration than with a convention-al 12V starter. A start-stop system using 48V also addresses the perception that these systems

Figure 4: A 3.5kW multiphase 12V/48V system

will help them reduce CO2 emis-sions and improve fuel economy. Not surprisingly, they are rapidly adopting 48V mild hybrid systems because they reduce emissions by 10-15% in small vehicles and 15-20% in larger models, and at a relatively low incremental cost.

Inside the 48V SystemThe 48V system typically consists of several core elements as shown in Figure 3. There is a 48V battery, a 48V starter/generator, and a 12-48V DC-to-DC converter. Typically, a 48V Lithium-ion (Li-ion) battery provides the energy storage for this board net. Li-ion is the battery technology of choice because of its high-energy stored-to-weight ratio (energy density) and its ability to accept energy quickly, enabling effi-cient regenerative braking systems. The 48V starter/generator replaces both the 12V alternator and 12V starter. It starts the engine and then acts as a dynamo, translating

the combustion engine’s rotational energy into electrical power. Dur-ing braking periods, the vehicle’s kinetic energy turns the dynamo and typically produces 12-15kW of power, while slowing the vehicle. Depending on the way the starter/generator is integrated into the powertrain, it may also be capable of providing a boost or torque as-sist during periods of acceleration. Finally, a 12V/48V bidirectional

DC/DC converter is used to link the existing 12V board net to the new 48V board net. The 12V/48V DC/DC typically supplies power up to 3.5kW from the 48V system to the 12V net to power the exist-ing vehicle electronics. In some situations, the 12V board net must supply the 48V electronics through the 12V/48V DC/DC as well, there-fore requiring bidirectional power transfer. Beyond these three com-mon elements of the 48V sub-system, other components that can be added include an electric compressor, electromechanical roll stabilization, electric screen clear, and various motors to support functions such as water pumps and cooling fans.

Reduce CO2 Emissions, Im-prove Fuel EconomyAs previously stated, one of the principal benefits that automak-ers can realize with the addition of the 48V board net is a reduction in CO2 emissions and improved fuel economy. Contributing fac-

Figure 2: Fuel economy legislation targets

Figure 3: Key components of the 48V board net



are irritating and don’t allow the driver to make necessary split-second decisions, such as entering into high-speed traffic.

Bidirectional DC/DC Controller Enables 12V-48V SystemsThe only additional component re-quired for a 12V/48V power-supply system is a relatively high power, up to 3.5kW, bidirectional DC/DC converter (as shown in Figure 4) to provide a bridge between the 48V and 12V systems. Previously DC/DC converters were designed for operation in only one direction and were limited to a few hundred watts or less which posed a major challenge. Fortunately, Intersil’s new controller is designed to provide bidirectional dc-to-dc conversion between the 12 and 48V board nets and is scalable for supplying power from under 1kW to over 3.5kW.

The ISL78226 is a 6-phase 12V/48V bidirectional synchro-nous PWM controller developed to solve the challenges associated with implementing a high power bidirectional bridge between the 12V and 48V power systems. To support this new requirement, automotive tier 1s initially had to build a system from existing de-vices intended for other applica-tions. One way was to use a DSP to implement digital power algo-rithms that could be used to drive a multiphase half-bridge power conversion stage. This approach was flexible, but required in-depth knowledge of control theory and implied extensive, long-term

support of the firmware. Another approach was to combine two unidirectional controllers with multiplexers and a microcontroller to activate the appropriate paths. This approach required redundant components that resulted in a large and complex final solution.

The ISL78226 is a single analog controller that performs bidi-rectional DC/DC power conver-sion, eliminating the downsides of alternative methods. It can control up to 6-phases and au-tomatically change the number of active phases to adapt to the load demanded by the output, thereby maximizing the DC/DC conversion efficiency. A single IC typically supports power conver-sion up to 3.75kW, but multiple ISL78226 ICs can be combined to manage more phases for higher power systems. When used in parallel to make higher power conversions, the ICs coordinate phase-to-phase current balancing and fault response. The control-ler regulates the output voltage (either 12V in buck mode or 48V in boost mode) and it includes a simple track interface to allow the output voltage to be ad-justed via a digital PWM signal or analog reference. And while it is regulating the output voltage, the device actively balances the average current flowing in each power stage phase. This helps equalize losses and prevents hot spots from developing in one of the phases that might otherwise occur when components such as inductors are mismatched.

Additionally, the ISL78226 regu-lates the maximum average cur-rent to limit the power transferred in either direction. This protects the DC/DC converter as well as the vehicle’s entire power net. It also allows for constant current charging in applications where either the 12V battery or the 48V battery is connected directly to the output. The controller includes a PMBus digital interface that provides system control and diag-nostics, enabling functional safety goals to be achieved with suffi-cient fault coverage. It also allows the system to take preventative action when conditions exceed warning thresholds, and in the event of a serious fault or failure, the controller can be configured to allow the vehicle to limp-home.

ConclusionAutomotive OEMs are fully em-bracing the 48V power subsystem and have signaled their intentions with new car model developments underway. Several automakers have already publically announced 48V mild hybrid vehicles. The impetus for adding another power bus is to improve fuel economy and reduce CO2 emissions through mild hybridization, while at the same time augmenting the drivability of the vehicle. The 48V board net achieves all of the automakers goals at a cost point that has more potential to reach wide-scale adoption than attained by full hybridization.

Intersil Corporationhttps://www.intersil.com/

Advances in Substrate Technology Enable New Power Modules for Automotive Application

By: Matthew Tyler, ON Semiconductor

Direct bonded copper (DBC) substrate construction is expediting the development of more advanced power modules

As we are all very well aware, there are mounting concerns about the state of

the environment and the impact that carbon emissions are having on the global ecosystem. There is also rising uncertainty about the long term supply of oil, as greater demand from emerging economies means that reserves are being consumed at an accelerated rate (with 35 billion barrels now being shipped each year, according to the International Energy Agency). Automobile manufacturers therefore face considerable challenges ahead. They need to develop car models that are capable of delivering improved levels of fuel economy. Through this they can ensure they continue to comply with international legislation and also keep ahead of their competitors, offering the car-buying public vehicles that are cheaper to run and more eco-friendly.

One of the key methods for

addressing this increasingly critical issue, and thus abide by the stringent regulatory guidelines that have now been put in place, is to reduce the mechanical load on the traditional petroleum powered engine. Another method is to reduce the overall mass of the automobile such that the less energy (fuel) is required to provide the same driving experience. The migration from the use of conventional mechanical systems to ones that are predominantly electrically based will help with both approaches. Electrical coolant and oil pump in addition to electrical power steering are common examples of this shift in design methodology. In addition the desire to further optimize more traditional implementations of electric motors in car such as side mirror adjustment, seat positioning, climate control, window opening/closing is further driving innovation in electromechanical design.

Currently both, brushed DC and brushless DC (BLDC)

motors are being employed for automotive applications. However, the popularity of BLDC implementations is rapidly increasing because they possess numerous attributes that are highly appealing to automotive engineers. Among the most prominent of these are: • BLDC motors offer a much

higher degree of control than DC devices, with variable speed operation supported (which is much more energy efficient).

• BLDC motors generally

exhibit greater compactness than a DC motor of equivalent output power. This means that their implementation will lead to the saving of valuable space (something that is becoming increasingly limited in modern automotive design, due to the greater volumes of wire harnessing and suchlike now being included).

• As BLDC motors do not





for more sophisticated, higher performance power modules to address automotive motor control, ON Semiconductor has positioned itself at the forefront of DBC based power module design. The company has introduced the STK984-190E. This is 30 A/40 V rated MOSFET power module, which is highly optimized for 3-phase BLDC motor drive applications within an automotive setting. It comprises 6 MOSFET devices in a 3-phase bridge configuration. These are accompanied by an additional MOSFET which is

assigned the role of providing a reverse battery protection switch mechanism. The module comes in dual in-line package (DIP) that measures just 29.6mm x 18.2mm x 4.3mm - a far smaller form factor than any equivalent product on the market. As all the MOSFETs featured in the STK984-190-E are fully AECQ101 compliant, the module has the capacity to deal with the difficult conditions that are found in automotive application settings. It is highly optimized for incorporation into

Figure 1: Schematic showing ON Semiconductor’s STK984-190E in an automotive fan system for climate control

Figure 2: The STK984-190E power module & LV8907UW motor controller IC in a climate control fan implementation; a) bottom view; b) top view

12V automotive applications, including pumps, fans and windscreen wipers. A 40°C to 150°C operational temperature range is supported. The STK984-190E can be used in conjunction with ON Semiconductor’s LV8907UW sensor-less motor controller IC (which incorporates a LIN transceiver for in-vehicle networking purposes) to provide a complete reference design for automotive BLDC motor control.

Through employing a more integrated strategy it will be possible to increase the efficiency 3-phase BLDC motor drives while simultaneously allowing the complete system to be much more compact and lightweight in nature than was previously possible, as well as exhibiting greater reliability characteristics. The high power density levels achieved by modules, such as the one just detailed, means that they will require only half the board space and half the component count that conventional discrete solutions would call for - a level of downsizing certain to have major impact on automotive electronics development moving forwards. These highly advanced metal substrate modules will consequently allow engineers to adhere to the improvements in fuel economy and space utilization now being expected by the automobile industry.

On Semiconductorwww.onsemi.com

require either brushes or commutators, there are thus no sparking or noise problems to contend with.

• BLDC motors are not subject to wear and tear (something that blights DC motors). This means they can support substantially longer operational lifespans - thereby mitigating the costs associated with repair or replacement.

Based on this, it is clear that BLDC motors can prove highly advantageous when they are installed into automotive systems. They do, however, require more sophisticated driving electronics than their DC counterparts. In many cases, discrete solutions are employed for driving BLDC motors, with a multitude of off-the-shelf components covering the power, control, protection and thermal management activities involved. Typically it will require the specifying and installing of 13-16 discrete components. This discrete approach is starting to be considered outdated, however, as both commercial and technical pressures become ever more acute. It relies on the procurement, placement and testing of too many elements to make it realistically viable in the future. Also it simply is not space efficient or particularly cost effective.

The use of power modules with

highly integrated construction has providing the industry with a more attractive alternative. It has enabled reductions in component count and space utilization to be witnessed, as well as increasing reliability (as there are fewer items involved that could potentially fail). Also it has meant that the inconvenience of placing discrete devices can be circumvented. In addition, this approach requires less engineering resources to be allocated to the task, resulting in shortened design cycles and lowering the investment needed. The emergence of next generation power modules with even higher degrees of integration and more innovative packaging technology (where a thermally conductive metal substrate has been utilized) holds the key to further progression in BLDC motor control.

Direct bonded copper (DBC) substrate construction is expediting the development of more advanced power modules that are able to bring about a number of operational and logistical advantages over either discrete deployments or conventional power module solutions. In basic terms the DBC substrate comprises an insulating ceramic which offers superior thermal conductivity. This ceramic layer is flanked on either side by a highly conductive copper layer. The external copper plate is covered with a finishing layer, while the internal copper

layer serves as the basis for a printed circuit layout. The components are soldered directly onto the DBC. Interconnect bond wires allow connection between the dies and the printed circuit layout, or between different components inside the module. This dramatically reduces the physical footprint of the system which yields immediate savings in size and weight of the finished solution.DBC also enables the thermal path of the power system to be significantly shortened without having to compromise on the electrical isolation. Through its employment power modules can be introduced where bill of materials costs have been dramatically lowered, development times accelerated and thermal performance improved. This technology also supports implementation of much smaller layouts than are possible via the alternative approaches previously outlined. As a result markedly higher power densities can be derived. It also dispenses with the need for additional insulation between the power module and the heat sink, which gives engineering teams greater design flexibility. The superior thermal performance also means that a smaller heatsink can be specified for the motor drive system - once again saving weight, space and lowering the cost of the total system.

In response to this growing need



Power Film Capacitor Design for EV and HEV Applications

By: John Gallipeau, Technical Marketing Manager for Power Capacitors, AVX Corporation

EV/HEV market expected to continue steady global growth worldwide

Automakers around the world are actively an-ticipating the decline of internal combustion

engine (ICE) models and the rise of clean energy fleets comprised of electric, hybrid electric, and plug-in hybrid electric vehicles (EVs, HEVs, and PHEVs). Due to the steadily increasing popularity of clean ener-gy automotive technology over the past few years, design engineers have already established a solid powertrain foundation for these vehicles. However, this market is expected to continue enjoying steady global growth worldwide for many years to come, and one of the primary reasons behind this projected growth is the expectation of significant technology advance-ments that will provide these vehicles with higher power density and improved performance, likely in line with government mandates expected to directly impact the power electronic systems of these vehicles.

The clean-energy vehicles of the future are expected to be just as reliable as current ICE models, and to provide range and safety fea-tures similar or equivalent to those of modern-day fuel-based vehicles.

Currently, that’s not often the case. To fulfill these expectations, invert-er manufacturers and OEMs will have to develop small, high-power, lightweight, and high-performance inverter designs capable of safety propelling these vehicles in a simi-lar fashion to ICE vehicles.

The achievement of the smaller, more advanced wide bandgap (WBG) inverters that this newer market demands to compete with experienced ICE technology at competitive costs relies on ad-vanced passive and active power electronics. Components like the DC filtering capacitors used be-tween batteries or generators and inverters to convert the DC voltage into AC drive power will have to

shrink down to the smallest size possible while remaining safe, reli-able, and effective.

Power Film CapacitorsAdvanced power film capaci-tors with controlled self-healing technology are one of the power electronics solutions that future EV and HEV engineers can rely on to meet the stringent size, weight, performance, and zero-catastroph-ic-failure reliability criteria of this demanding market.

Power film capacitors capable of delivering reliable design solutions for EVs and HEVs must meet sev-eral specific parameters regarding metallized film materials, process-ing, and design.

Self-Healing CapabilitiesOne of the primary advantages of film capacitors is their ability to overcome internal defects, or to exhibit self-healing capabilities. The dielectric films used in DC filter capacitors are coated with a thin metallic layer. If a defect oc-curs, this metal layer evaporates, isolates the defect, and effectively heals the capacitor.

Initially, this metallic layer was ap-plied to the whole of the film. This approach to the self-healing pro-cess is effective in low-energy appli-cations, but proves insufficient in high-power applications with high reliability standards. In 1979, AVX developed the first film capacitors with a controlled self-healing pro-cess by applying the metallic layer to the film in small patterns, rather than all over. This resulted in safer and more controlled power film ca-pacitor performance, and allowed these products to be employed in higher-power applications.

Controlled self-healing capabilities are achieved through the segmen-

tation of the metallized film. This segmentation divides the total capacity into elementary cells — up to several millions per capacitor — that each act as a fuse. When defects occur in power film capaci-tors with controlled self-healing capabilities, the cells with the weak points are the only ones that are insulated by essentially blowing the fuse. As a result, the capacitance decreases as a function of the ratio between the elementary cells and the total surface of capacitor, but no short circuit or failure occurs.

Power film capacitors with con-trolled self-healing capabilities are the only film capacitors that can reliably prevent catastrophic failures (e.g., short circuits with energy in parallel) in high-energy applications with electrical fields of more than 200V/µm, like EVs and HEVs. As such, all metallization manufacturers have long utilized some variety of laser segmentation to achieve controlled self-healing power film capacitors.

The advantages of controlled self-healing technology include its:

• Ability to satisfy challenging mechanical and electrical specifications

• Long life expectancy • Proven field reliability with

zero catastrophic failures, even under severe usage

• Broad market and application suitability

The suppliers with the most expe-rience manufacturing power film capacitors with controlled self-healing technology can also offer competitively low-cost solutions.

Flat Bobbin DesignsThe metallized film within power film capacitors also affects perfor-mance, as at least two metallized films have to be wound or stacked together in opposing fashion in order to achieve capacitance value.

Metallized film is often wound into flat bobbin designs to create film capacitor products. These elemen-tary bricks effectively balance current and inductance, and are designed for operating tempera-tures spanning -55°C to 115°C. They are wound on high-productivity

Figure 1: Metallized polypropylene film segmented into a mosaic design (left) and an electrical schematic illustrating how the segmented film operates during the controlled self-healing process.

Figure 2: A power film capacitor bobbin wound and metal-sprayed before soldering (left) and multiple bobbins soldered into a bus bar configuration (right).



winding machines and, although their width and external diameters vary, their generally medium size prevents winding forces from pull-ing on and damaging the intrinsic dielectric property of the polymer film. Basic connection topology in-duces the association of bobbins, and the design of the connection system achieves added integration functions within the capacitor.

Advantages of flat bobbin designs include: a good filling factor, high winding productivity, modularity in three dimensions, no critical thermal expansion, the highest electrical field and specific energy, lower cost, serial resistance, and balanced inductance.

Bus Bar Designs Bus bar designs directly affect the self-inductance value of film ca-pacitors, and traditional designs typically have self-inductance values spanning from 5nH to 100nH, depending on their design and size. The flexibility offered by the elementary brick/flat bobbin design concept can achieve bus bar designs with inductance values below 6nH.

Inductance below 6nH is needed to limit over-voltage during switch-ing. One way to achieve this is us-ing two bars parallel to each other, as this cancels out the most of the internal inductance by optimizing the recovery surface. However, this approach must be considered with caution, as it can generate addi-tional bus bar tooling costs and manufacturing costs.

Optimizing Power Film Capaci-tor DesignsIn order to optimize power film ca-pacitor designs for EVs and HEVs, it is very important to obtain a mission profile in term of volt-age, temperature, and root mean square (RMS) current. These three parameters directly influence both the dimensions and the functional cost of the capacitors.

Several leading capacitor manu-facturers are working with materi-als suppliers to further improve the operating parameters of film capacitors for future EV and HEV applications. Of particular interest are new materials that improve the upper limits of operating temperature ranges and increase dielectric constant. For example, higher quality polypropylene films generally have fewer amorphous phases and can withstand operat-ing temperatures up to 105°C — and even higher temperatures for short periods of time.

Thermal conditions are fundamen-tal for capacitor design. The rela-tionship between RMS current and ambient temperature induces hot spot temperature in these capaci-tors, and hot spot temperatures and voltage are fundamental pa-rameters for achieving the guaran-teed long-lifetime and high-reliabil-ity performance that EV and HEV applications demand. The thermal resistance and thermal conduction of the capacitor combined with the expected ambient temperature and cooling conditions noted in the mission profile can be used to calculate the hot spot and improve upon the final design.

Further developments will result from a combination of advanced dielectric materials capable of achieving the higher tempera-ture performance and higher hot spot temperatures required by EV and HEV applications, reduced film thicknesses, and enhanced

metallization and segmentation processes, and reduced film thick-nesses. To improve the electrical performance of power film capaci-tors the quality of the film itself must be improved. For instance, if the target is to increase epsilon G², where Epsilon = dielectric constant and G = gradient of volt-age, the volume of the capacitor is inversely proportional to the square of gradient of voltage.

Future Developments

Figure 3: An example of mission profile for an HEV/EV automotive application.

Table 1: A basic overview of the different types of dielectric materials.Currently, there are no established standards or working conditions for the power film capacitors that are employed in the automotive industry. Instead, capacitor manu-facturers work with automotive designers to develop optimized, application-specific capacitors using data including: voltage, cur-rent, temperature profile, lifetime expectancy, reliability, and cooling method (e.g., heat sinks, water tube systems, or liquid-in-vapor).

However, since this market is expected to continue its global growth, supported by significant technology advancements pre-dicted to achieve range and safety performance parity between EVs and HEVs and traditional ICE vehicles, both industry and gov-ernment mandates are currently being developed to ensure the safety and performance of the power electronics components responsible for enabling the next-generation WBG inverters that provide these vehicles with higher power density, improved perfor-mance, and zero-catastrophic-failure reliability.

AVXwww.avx.com

www.taiyo-yuden.com

Telecommunication, Information, Consumer, Industry and Automotive Electronics

HIGHEND QUALITY ELECTRONIC COMPONENTS



POWER SYSTEMS DESIGNFINALthought

By: Ally Winning, European Editor, PSD

The Philosophical Battle for the Internet

The Internet has always been

a great enabler of technol-

ogy. In itself, it provides a

platform that has changed

the way we live our lives. Companies,

such as Amazon and Google have

become the largest corporations in the

world by taking advantage of the tech-

nology. The Internet has also revolu-

tionised business, it allows companies

to sell produce around the world, and

to collaborate as never before.

Many other technologies are also built

on the Internet infrastructure, including

some that have the potential to change

other facets of our lives by as much

as the Internet itself did originally. The

Internet of Things is a great example of

this type of technology.

The development of Internet-based

technologies will be incentivised by

cheap and easy access to the technol-

ogy. The more potential customers

available, the more incentive there is for

companies to develop solutions. For ex-

ample, when more people got access to

broadband internet, streaming services

became viable and Netflix arrived.

Currently there are three separate

battles being fought over access, which

could see many US citizens deprived

of the open Internet. Firstly, and most

often discussed is net neutrality. At the

moment, Internet providers are banned

from discriminating against traffic from

different websites, so one provider can’t

offer Netflix at a different speed from

Hulu. The FCC is looking to drop that

restriction and allow providers to favour

certain sites with higher speeds. A vote

on net neutrality could be held as early

as December.

The second issue is that the FCC has

proposed abandoning the functional

test implemented in 2014 that required

telecom providers to ensure that ser-

vices wouldn’t be degraded if copper

lines were turned off. With the legisla-

tion abandoned, it means that Telcos

would be able to turn off copper lines

without conducting an analysis of the

effects on consumers.

The third issue concerns a scheme

that was introduced to ensure that low-

income Americans had access to cheap

Internet. The Lifeline program subsi-

dised the cost of broadband services

and guaranteed that poorer Americans

would have access to the program. The

proposed FCC change cuts 70% of the

providers from the program, while also

restricting the available budget.

These three issues could see access

to fast Internet curtailed for many

Americans, and in the case of net

neutrality, access to any sites outwith a

purchased package become restricted

or too slow to be worth visiting. The

US seems to be turning its back on

the philosophy that made the Internet

what it is, that all sites are equal and

users choose the best one, not Inter-

net providers.

The US was the country that first saw

the potential of the Internet and has

owned the first wave of innovation.

Now, the second wave of innovation

is upon us, and opportunities are

opening up for services using the IoT,

such as telemedicine. Will there be

the same incentive for US companies

to deliver solutions when many of the

people that would benefit most are

effectively banned from the medium?

PSD

www.powersystemsdersign.com

Registration Now Openwww.apec-conf.org

LT8650S

OUT1

OUT2

Easy PCB LayoutEasy PCB Layout

6.2µA IQ6.2µA IQ

94.4% Efficient @ 2MHz94.4% Efficient @ 2MHz

2x4AOUT2x4AOUT

200kHz to 3MHz200kHz to 3MHz CISPR25 Class 5CISPR25 Class 5

Linear Technology ASM STANDARD PAGE TRIM SIZES : 8 x 10.5” Minimum 8.5 x 11.25” / 216mm x 286mm Maximum

Agency contact: Jon MiwaPhone: 926-642-3053Email: [email protected]

More Info Silent Switcher 2 Family

, LT, LTC, LTM, Linear Technology, the Linear logo,and µModule are registered trademarks of Analog Devices, Inc. All other trademarks are the property of their respective owners.

www.linear.com/product/LT8650S1-800-4-LINEAR

The LT8650S combines high efficiency at high switching frequencies and an ultralow EMI design to deliver a very compact dual output 4A (6A peak)/channel synchronous step-down solution. The LT8650S delivers 94.4% efficiency while switching at 2MHz stepping 12VIN down to 5VOUT, enabling a very small solution footprint while keeping switching noise out of critical frequency bands. Its Silent Switcher 2 architecture uses special design techniques and integrates input, boost and INTVCC bypass capacitors to ensure ultralow EMI/EMC emissions on any PCB board or layout easily meeting CISPR25, Class 5 limits. The LT8650S consumes only 6.2μA quiescent current and is offered in a 4mm x 6mm LQFN package.

BreakthroughPerformance

Silent Switcher®2, Unrivaled Ease-of-Use

LT8609S 3.0 to 423.0 to 653.4 to 423.0 to 423.0 to 423.4 to 653.0 to 183.0 to 18

2A/3A Peak5A/7A Peak5A/7A Peak

2x2A/3A Peak2x4A/6A Peak

8A10A2x9A

0.80.970.970.80.80.80.80.8

200kHz to 2.2MHz200kHz to 3MHz200kHz to 3MHz200kHz to 3MHz200kHz to 3MHz200kHz to 3MHz200kHz to 3MHz200kHz to 3MHz

2.52.5170

66.22.516020

3x3 LQFN-164x4 LQFN-244x4 LQFN-243x4 LQFN-204x6 LQFN-324x6 LQFN-324x4 LQFN-244x7 LQFN-36

Part Number Frequency IQ (μA) PackageVIN Range (V) IOUT (A) VOUT(MIN) (V)

LT8640S

LT8643S

LT8653S

LT8650S

LT8645S

LT8642S

LT8652S

Find your local sales office: www.linear.com/contact

7th generation igbts improve thermal …...profile, thermally efficient solution with extended 11.5...

Documents