January 2013

High Frequency Wireless Power

Mark Adams
Senior Vice President of CUI, Inc.

Wireless Charging Station

Pushing
the

A look at CUI's Novum Advanced Power Modules

Envelope

All the forces in the world are not as powerful as an idea whose time has come.
Victor Hugo, 1800

Power Developer contains new ideas that come every month.

Power Developer Editors, 2013

POWERDEVELOPER

Read Power Developer, the monthly newsletter for Engineers:

h t t p : / / w w w. e m b e d d e d d e v e l o p e r. c o m / n e w s _ l e t t e r /

Power Developer

CONTENTS

4 10 20 28 34

eGaN FETs for High Frequency Wireless Power Transfer

TECH COLUMN

Making Wires and Cords a Thing of the Past

FEAUTRED ARTICLE

Mark Adams - Senior Vice President of the Business Unit at CUI, Inc.

COVER INTERVIEW

Compensation Methods in Voltage Regulators

TECH ARTICLE

Pushing Things to the Limit? Asymmetric IGBT Modules for Smooth Swithching

TECH COLUMN

Power Developer

TECH COLUMN

CEO of Efficient Power Conversion (EPC)

Alex Lidow

Power Developer
The previous columns in this series discussed the advantages of eGaN FETs in high frequency hard-switching, resonant, and soft-switching designs. In this installment, highly resonant, loosely coupled, 6.78 MHz ISM band wireless power transfer will be presented that show how eGaN FETs are enabling this technology. The aim of this column is to show efficient wireless energy transfer using current eGaN FETs and the LM5113 eGaN FET half-bridge driver, and present examples of a voltage mode class D and class E approach.

To read the previous installment, click the image below:

Wireless Energy Transfer Systems

Wireless energy transfer applications are gaining popularity for mobile device charging solutions that demand a low profile and high robustness to changes in operating conditions such as coil spacing, coil alignment, and load power demand. The superior characteristics of eGaN FETs, such as low input and output capacitance, low parasitic inductances, and small size have made them ideal for use in these systems. In this article two classic switch mode based RF amplifiers will be compared; (1) the voltage mode class D [1, 2, 3], and (2) a single-ended class E [4, 5]. Excluded from this comparison will be advance control techniques as all circuits will be operating at a fixed frequency (6.78 MHz) and duty cycle (50%).

Figure 1: Source coil loop, device coil and rectifier circuit (left) with simplified representation (right).

TECH COLUMN
eGaN FETs have previously been demonstrated in a 15 W, 6.78 MHz classic voltage mode class D wireless energy transfer system [1, 2] that had a peak efficiency over 70%, and provided 4% higher total efficiency than a comparable MOSFET version. It was shown in [1, 2] that the system would benefit from using a smaller device as the dynamic switching losses were high in comparison to the conduction losses. This means that a small device with lower output capacitance (COSS) and higher RDS(on) would be beneficial to the efficiency performance on the system. With the introduction of the 3rd generation series of eGaN FETs [6], a small device, the EPC8004 [7], with excellent high frequency characteristics has become available. This new part will be compared in the same source amplifier as the original design. The class E amplifier has also received a lot of attention as a suitable candidate for wireless energy transfer due to its high theoretical efficiency and ability to absorb the COSS of the switching device into the matching network circuit. Practically however, the losses due to RDS(on) and the inductors and coils (ESR) will reduce the efficiency of this approach. All the amplifiers will be compared against each other using the same coil set and load. Only the changes to the source coil necessary to establish proper operation for the specific amplifier will be made. The device coil and rectifier in each case is the same circuit, which enables a direct comparison based on its merits. The load, rectifier, coil set with deviceside matching is shown in Figure 1 (left). This circuit is simplified to a single impedance parameter (Zload) for design purposes. This allows all the designs to be compared simply by the differences required for operation.

Figure 2: Voltage Mode Class D wireless transfer system schematic (left) with ideal waveforms (right).

Figure 3: Class E wireless transfer system schematic (left) with ideal waveforms (right).

Voltage Mode Class D Operation

The voltage mode class D, with the schematic shown in Figure 2 (left), operates the load at resonance by resonating Cs with the reactive component of Zload to establish a sinusoidal current in the load and to overcome the leakage inductance of the coil set [8]. Each half cycle a switch will carry the half sinusoidal current as shown in Figure 2 (right). This will establish a zero current switching (ZCS) event for the switches. The ZCS event will eliminate turn off commutation losses but will suffer from

Figure 4: Efficiency results for the class D and class E wireless energy transfer systems.

Power Developer
turn on losses, including commutation and COSS losses. To maximize performance, operation of the converter should be altered to operate slightly above resonance, using Lm and Cm, to achieve zero voltage switching (ZVS), eliminating the more penalizing turn on losses.

Class E Operation
The class E, with the schematic shown in Figure 3 (left) also operates the load at resonance by resonating Cs with the reactive component of Zload to establish a sinusoidal current in the load and to overcome the leakage inductance of the coil set [8]. However, with only one switching device that is used in conjunction with a matching network (Extra inductance (Le) and extra shunt capacitance (C she)) to establish a resonating voltage, this converters device experiences a zero voltage switching (ZVS) event as shown in Figure 3 (right). A shunt capacitor (Cshe) across the device is used as part of the resonant circuit which is used to absorb COSS into the resonant circuit. The inductor LRFck is used as a current source for the circuit.

Figure 5: Measured switch-node voltage waveform of the class D system showing the difference in voltage transition due to COSS between the EPC2014 and EPC8004 devices.

Experimental Verification
Two class D systems were built and tested, one using the EPC2014 and the other the EPC8004 devices. A class E system was built and tested using the EPC2012 device. Figure 4 shows the efficiency results for two versions of the voltage mode class D systems and the class E system. The results show that the lower COSS has an efficiency advantage due to the shorter transition time despite the higher RDS(on) in the case of the class D approach. The class E system is significantly more efficient, due to having one switching transition occurring with simultaneous zero voltage and zero current and the other at a reduced current. Operation to a higher power output is possible due to the higher voltage rating of the device. Figure 6: Measured class E amplifier drain to source voltage and extra inductor current waveforms. In the case of the class D system, the voltage overshoot of the switch-node would limit the maximum output power.

TECH ARTICLES
Figure 5 shows the switch-node voltage waveform of the class D systems. It is clear that the lower COSS of the EPC8004 device allows for significant reduction in the voltage transition time. This ultimately results in higher class D efficiency. Figure 6 shows the drain to source voltage (VDS) waveform of Q1 for the class E amplifier together with the current in the extra inductor Le.

References
[1]  M. A. de Rooij, Low Power Wireless Energy Converters, Efficient Power Conversion white paper WP014, http://epc-co. com/epc/documents/papers/eGaN%20 FETs%20for%20Wireless%20Power%20Transfer%20Applications.pdf [2]  M. A. De Rooij and J. T. Strydom, eGaN FETs in Low Power Wireless Energy Converters, Electro-Chemical Society transactions on GaN Power Transistors and Converters, October 2012, Vol. 50, No. 3, pg. 377 388. [3]  S-A. El-Hamamsy, Design of High-Efficiency RF Class-D Power Amplifier, IEEE Transactions on Power Electronics, Vol. 9, No. 3, May 1994, pg. 297 308. [4]  F. H. Raab, Idealized operation of the class E tuned power amplifier, IEEE Transactions on Circuits and Systems, December 1977, Vol.24 , No. 12, pg. 725 - 735. [5]  W. Chen, et al., A 25.6 W 13.56 MHz Wireless Power Transfer System with a 94% Efficiency GaN Class-E Power Amplifier, IEEE MTT-S International Microwave Symposium Digest (MTT), June 2012, pg. 1 3. [6]  M. A. De Rooij and J. T. Strydom, Introducing a Family of eGaN FETs for MultiMegahertz Hard Switching Applications, Application Note AN015, September 2013, http://epc-co.com/epc/documents/ product-training/AN015%20eGaN%20 FETs%20for%20Multi-Megahertz%20Applications.pdf [7]  EPC8004 datasheet, http://epc-co.com/ epc/Products/eGaNFETs.aspx [8]  K Siddabattula, Wireless Power System Design Component And Magnetics Selection, Texas Instruments http://e2e.ti.com/ support/power_management/wireless_ power/m/mediagallery/526153.aspx. [9]  Z. Xu, H. Lv, Y. Zhang, Y. Zhang, Analysis and Design of a Class E Power Amplifier employing SiC MESFETs, IEEE International Conference of Electron Devices and SolidState Circuits, December 2009, pg 28 31.

Summary
In this column we presented highly resonant, loosely coupled, 6.78 MHz wireless energy transfer for both voltage mode class D and class E approaches. In each case it was demonstrated that eGaN FETs enable wireless energy transfer by providing efficient and compact solutions. In the case of class E, the efficiency exceeded 80% over a substantial portion of the output load range. It has been shown that eGaN FETs support both approaches and the selection of converter topology would further be dependent on application specific requirements such as sensitivity to load range and coil coupling variations. eGaN FET is a registered trademark of Efficient Power Conversion Corporation.

About the Author

Alex Lidow is co-founder and CEO of EPC, the leader in enhancement-mode galliumnitride-on-silicon (eGaN) FETs. Prior to EPC, he was CEO of International Rectifier. Coinventor of the HEXFET power MOSFET, he holds many patents in power semiconductor technology and has authored numerous publications. He is a graduate of Caltech with a Ph.D. from Stanford.

CLICK HERE
to comment on the article.

Power Developer

10

TECH ARTICLES

Wires and Cords a Thing of the Past

Just like in a TV Western, the Calvary is coming to the rescuebugles sounding the charge and all. In this case, its a highly innovative Israeli start-up company Humavox thats saving mankind from the morass of tangled wires, cords, and connections associated with your growing patchwork of connected devices. Theres more good news; you wont need an additional gizmo or gadget to slap onto your wireless device in order to charge it. This version of wireless charging is as simple as putting your mobile or wireless device in a cup-like receptacle and letting it do its thing. Humavox recently launched its Eterna hardware platform, which enables handheld devices to wirelessly charge. Instead, the technology works by using radio frequency signals to send power to those mobile devices in need of a charge. By adding the ThunderLink enabler added to the device, the user can begin wirelessly charging. With this revolutionarily simple method, Humavox will certainly make waves in an industry that is crowded with cables.

Making

11

Power Developer
Vaporizing Cables Since 2010 Humavox CEO Omri Lachman is the first to tell you hes not an engineer. He doesnt have an electrical engineering background, nor has he ever taken an engineering course. Instead, as a youngster, he was surrounded by an industrial environment. Early on, he found his niche in digital marketing with an entrepreneurial bent. Back then, around 2007, there was considerable talk about Qualcomm and Nokia among Lachmans circle of friends. His best friend and soon to become co-founder and CTO, Asaf Elssibony, was eager to explore the space and talked endlessly about wireless power. Then it hit them; It suddenly crossed our minds that wireless charging would be a magnificent idea, Lachman told us. Humavoxs initial idea was to develop wireless charging not specifically for mainstream devices like tablets and smartphones, but more toward a universal wireless power solution for virtually any mobile wireless device. In other words, wireless charging would be developed to make it easier, more convenient, he said. During these early stages, the developers focused on RF as the energy source for their wireless charging scheme. This enabled Humavox to overcome user-usability and engineering-related constrains that are associated with other wireless charging forms that required too many user interactions and proportional engineering efforts to integrate. They set their sights on developers and manufacturers as their target market. Today, their sales pitch centers around how manufacturers can quickly and costeffectively integrate wireless charging into their own products and branding.

For designers and manufacturers, we created wireless charging that can be actively accessed and rapidly integrated. Its something that does not require a lot of engineering effort.

That messaging may not be sufficient because Humavox is not alone in this exploding market, which some say is still in its early stages. The latest market research forecast for the wireless charging market is at $7.161 billion by 2017. This involves various technologies including inductive, magnetic resonance, RF, microwave, and optical beam for power charging. The Technology Lachman said it best: The basis for our technology development is actually driven not so much from a marketing perspective, but considerably more from the reasoning of a users experience like dropping a coin in a box.

12

TECH ARTICLES

Omri Lachman CEO of Humavox

13

Power Developer
For consumers, the Humavox team wanted to make charging an effortless and intuitive action something that fits all types of devices, but also all types of users from children to seniors. For designers and manufacturers, Lachman stated, we created charging that can be actively accessed and rapidly integrated. Its something that doesnt require a lot of engineering effort. There are three main pieces to the wireless charging arrangement. First, theres the Eterna technology, which Humavox calls a flexible and versatile hardware platform. According to Lachman, Eterna is a, quick and easy pathway for wireless device manufacturers to adopt and integrate this technology without making formative changes in existing architecture, design, and components. Eterna sold as intellectual property to device manufacturers so they can integrate the wireless charging element in whatever method works best for them. This eliminates the need for consumers to buy external hardware to enable wireless charging. Humavox also provides its ThunderLink wireless charging receiver. The ThunderLink is said to be a flexible way to integrate the wireless charging enablement into any electronic device. There is an assortment of integration choices and form factors available to a device manufacturer, and this integration is performed without any extra or unnecessary engineering design work. The third item is called the NEST Power Station, which is a design-free RF resonator. The NEST Power Station is the first wireless power generator under ETERNA chargers category and was conceived after long observation and user behavior analysis. We realized theres a certain user usability

Eterna is a quick and easy pathway for wireless device manufacturers to adopt and integrate this technology without making formative changes in existing architecture, design, and components.

14

TECH ARTICLES

NEST Charging Station

pattern a habit, Lachman told us. Consumers customarily place their precious electronics, be it for healthcare, wellness, or entertainment purposes, in all sorts of enclosures, from protection, sterilization or just to keep it away from others. Humavox wanted to mimic that usability habit. NEST allows brands and developers to unify their product storing with wireless charging enablement. Since theyre already designing box and cases, they might as well get it enhanced for user convenience. Basically, its a receptacle where you place the wireless product so it can be re-charged. Humavox said there are endless options when it comes to shapes and sizes to choose from, or it can make the right one according to a manufacturers specifications. Plus, there are no placement or orientation requirements with the NEST power station. After we achieved wireless charging, Lachman said, we ended up creating energy hotspots inside the NEST that allowed us a high level of wireless power transfer efficiency. Also, NEST does multiple-device

charging in a smart way to help manage energy. RF actually works great if resonated and applied in a particular area of confinement. Too Many Bees around the Hive Lachman argues that his RF wireless charging product has a competitive edge over those other technologies. His reasoning is it is more ergonomic and easier for a person to simply drop his or her wireless device into a cup or bowl rather than precisely placing it on a pad. Meanwhile, Humavox is working diligently to safeguard its IP. Lachman said, We are actively filing patents in areas serving as an umbrella that guards our unique RF as energy developments. Simultaneously, the company strategy is to search out and target meaningful partners that parallel Humavoxs market direction. Once we find manufacturers we can successfully work with, Lachman told us, we can then start telling our story and approach more mainstream applications.

15

Power Developer
However, the challenges they face deal with technology deployment and market strategy, and they go hand in hand, he said. Manufacturing is not a challenge. Our platform allows device manufacturers to experience how we design our power and charging solution the way they want us to from controlling cost to controlling industrial designthe whole user experience, Lachman said. But as far as the go-to-market plan, Humavox can ill afford to be overly deliberate in its marketing and sales since theyre battling a growing list of competitors and a variety of technologies. For example, a competitor, with a similar technology, PowerbyProxi of Auckland, New Zealand, already has industry giant Samsung supporting it with a $4 million investment. All those market entrants and growing list of disparate technologies arent helping market growth according to some market researchers. The market research firm, iSuppli, said one way to spur market adoption is to adopt a common standard that ensures interoperability among solutions being developed. The industry will remain fragmented until the industry defines a standard, it said. Lachman isnt convinced about adopting a standard or about industry alliances being formed. Were trying to avoid coming up with another standard right now. There are too many already and theyre overwhelming the industry. The focus should be on achieving a unified solution that works exceptionally well for engineering groups and then make that the standard. As far as wireless charging alliances, there is the Wireless Power Consortium

After we achieved wireless charging, we ended up creating energy hotspots inside the NEST that allowed us a high level of wireless power transfer efficiency"

(WPC) Steering Committee and the Alliance for Wireless Power or A4WP, which is a consortium that Qualcomm and Samsung founded. To date, WPCs claim to fame is Qi, which is the widely adopted wireless charging specification. Multiple alliances are creating too many bees around the hive, Lachman said. As a marketer, the first question I ask is: Is this technology scalable enough? We keep challenging ourselves by saying we have this wonderful solution that fits like a glove for all sorts of healthcare and wellness wireless device applications. But will it fit other devices? Thats why we maintain close engagements with major companies from different market categories from battery makers to wearable electronics firms, Internet of all things applications, all the way to medical applications. We see the solution as scalable. Creating that intuitive type of wireless charging makes a lot of sense to people.

16

TECH ARTICLES

PURE DIGITAL POWER

ZL8800 Combines Digital Control Loop Technology and PMBus Interface for the Industrys Most Advanced Digital Power Solution Not only can you now track and control every aspect of your power supply without additional monitors or sequencers, the ZL8800s ChargeMode control loop technology delivers best-in-class transient response and eliminates the need for complex compensation. This all adds up to a pure digital solution that lowers system cost, reduces design time and increases available board space.

INTERSIL.COM/ZL8800
Download Datasheet Read App Note Order Demo Board Watch Videos

Stop Coding and Start Designing with Intersils PowerNavigatorTM

Download free at intersil.com/powernavigator

New Digital Power Knowledge Center

Watch and learn at intersil.com/DigiPowerTraining

2013 Intersil Americas LLC. All rights reserved.

17

MEDICAL AC-DC POWER SUPPLIES

experience... quality... cost effective
PMP85SF Series PM60 Series PMP220 Series

85 Watt Desktop 2x MOPP Class II Input Single Output EN60601-1 3rd Edition Certified

60 Watt PCB/Cover 2x MOPP Class I / Class II Input Single and Multi Outputs EN60601-1 3rd Edition Certified

220 Watt Desktop 2x MOPP Class I / Class II Input Single Output EN60601-1 3rd Edition Certified

Protek Power delivers high quality 15W-650W medical grade AC-DC power conversion solutions that are certified to international EN60601-1 standards. Discuss your application for our standard products or modified designs with us.

www.protekpowerna.com
43 Broad Street Hudson, MA 01749 Tel: 978.567.9615 sales@protekpowerna.com

Visit us at www.protekpowerna.com

Online Circuit Simulator

PartSim includes a full SPICE simulation engine, web-based schematic capture tool, and a graphical waveform viewer.

Some Features include:

Simulate in a standard Web Browser AC/DC and Transient Simulations Schematic Editor WaveForm Viewer Easily Share Simulations

Try-it Now!
www.partsim.com

Join Today

eeweb.com/register

Power Developer

Pushing the
Mark Adams Senior Vice President of CUI, Inc.

Envelope

CUI is a technology company headquartered in Tualatin, Oregon. Over the past two decades, the company has become a leading provider of power products. To stay ahead of the curve, CUI continually integrates the latest technologies and design techniques into their ac-dc and dc-dc power modules. The companys Novum Advanced Power product line was developed to address the growing complexities in todays advanced designs. We spoke with Mark Adams, Senior Vice President of the Business unit at CUI, about the companys Novum Advanced Power module, the advantage of digital power over analog, and the companys unique approach to innovation.

20

COVER INTERVIEW

21

Power Developer
What are some of CUIs core products and target markets?
There are two product areas within CUI Power and Components. The Power product group is comprised of a broad portfolio of ac-dc power supplies and dcdc converters that range from 0.25 watts all the way up to 2400 watts. Our Components product group supports a broad range of board-level electromechanical products in the interconnect, audio, and motion control spaces. Within our Power group there is a separate business unitwhich I head-upcalled Novum Advanced Power. This business unit focuses specifically on advanced dcdc board mount power for intermediate bus architectures, where isolated dc-dc bricks convert 48 volts down to 12 volts, and then 12 volts down to the voltages that are required on the board via non-isolated point of load modules. Novum Advanced Power is primarily focused on the higher-power, high-complexity design requirements of todays customers.

What kind of design resources do you have for engineers looking to implement these power modules into a product?
We have a complete engineering team at our Tualatin, Oregon HQ, as well as engineering resources in San Diego and China. Additionally, we have an extensive design application support group available to assist customers. What we are really trying to do, however, is create a simple solution for a very complex problem by injecting a level of intelligence into our product portfolio so a customer doesnt need to call on a team of FAEs. We want to make the design process as easy as possible for the engineer, giving them the tools necessary via our Novum ACE GUI to follow a simple set of design rules and optimize to their specific power requirements. We are really focused on designing self-supportive products, so a lot of the work thats done here internally can be leveraged by our customer base to make their job easier.

22

COVER INTERVIEW
What sort of overall trends do you see in the power industry?
The ability to cut and paste discrete designs from previous generations is no longer feasible in many instances as the processing density in todays ICs have significantly increased complexity in the power architecture, so we definitely see more customers moving towards modules. We also see customers asking for more functionality in the power moduleit is not just about power conversion anymore. Once the customer has a product up and running, they want to know whats going on in their system. Once you have that, you can take it to the next level of doing predictive failures and start getting very sophisticated. We also work to integrate our own IP into the design to further boost performance. For example, we have our Solus Power Topology, which is a SEPIC-fed buck topology that allows us to greatly increase efficiency, power density, and transient response when compared to traditional topologies typically used in todays designs. So not only do we take the best of what others do and package that into a product, we also take our own IP and add that into the mix, creating products that lead the market in the performance areas crucial for our customers nextgeneration designs.

How does CUI approach innovation?

As a power supply company, we see ourselves as an integrator of technologies. Its our job find the best technologies and marry them in a power supply, balancing the performance and cost for our customer.

Our Solus Power Topology is a SEPICfed buck topology that allows us to greatly increase efficiency, power density, and transient response when compared to traditional topologies typically used in todays designs.

23

Power Developer
What types of applications does the Novum Advanced Power line target?
Our primary target is the datacom/telecom market. Complex designs in this space need high reliability and leading-edge performance. Down time is bad when you are dealing with equipment supporting billions of web searches or executing split-second stock transactions. These are applications where fractions of seconds of down time can cost millions of dollars, so reliability and intelligence are crucial. This market is the most applicable for this product line; however its starting to spread as processing power becomes important in other markets.

What kind of feedback are you getting from your customers about this product line?
Ive been involved in digital power for almost nine years now. Its been a transitional period for the industry, slowly moving from analog power architectures to a digital power structure. As todays ICs become more advanced, digital power is actually becoming a requirement in many applications. This is creating a challenge for our customers, as there just arent a lot of new analog power engineers coming out of college to address these challenges. What ends up happening is that digital engineers employed by the networking companies get paid for the IP they put into the DSP or ASIC, but they are also tasked with trying to figure out how to power these devices. The latest chipsets are pushing 75 to 100 amps thats not an easy task. Furthermore, these chips are operating at voltages around 0.9 volts with extremely tight voltage tolerances and transient requirements, that, if you go outside of a certain band gap, searches may be lost. How do you address that without bringing in a dump truck of output capacitance? It can take us three iterations of designs to get it perfect, and that is in a controlled environment. You can image how challenging this can be for an engineer who doesnt specialize in complex power design. We continually hear from customers that power is the necessary evil. At CUI, we have the ability to put 90+ amps on a module thats already tested and proven, allowing the customer to plug it into their design and connect to our simple-to-use Novum ACE GUI. They can then make the required adaptations on-the-fly via software, allowing the engineer to devote more time and energy to other aspects of their design.

24

COVER INTERVIEW

At CUI, we have the ability to put 90+ amps on a module thats already tested and proven, allowing the customer to plug it into their design and connect to our simpleto-use Novum ACE GUI.

25

Power Developer

 We have products that we will release this quarter that will be the highest density and highest output available in  the marketplace.
In terms of moving your product line forward, what are the biggest challenges you are facing?
Our biggest challenge is making sure that we understand what the next generation of chips will require as the pace of innovation continues to accelerate in the networking space. We strive to address the most challenging sockets in each application. Many times, its the core of those boards or a network switch that has transient requirements and performance metrics that are extremely difficult to meet with a standard type of design. If we can address these power rails, then we should be able to address everything else with our product portfolio thanks to the digital nature of our products.

What excites you the most about the future of power supply technology?
What excites me is that power is not going away. Without power, nothing else works. It will always be there and it will always be a problem that needs to be solved for the customer. The problems are endless, and will continue to grow as core voltages drop and current requirements rise in advanced ICs. Now that, thanks to digital, power has a brain, you can really start to get sophisticated. The challenge there is how to take a sophisticated scheme and simplify it so that its not a major task for our customers to implement. There are a lot of things we can do, not only from a density and power conversion point of view, but from the intelligence aspect. With added intelligence, you can impact performance, impact reliability, do predictive algorithms, and make the solution much more efficient. There are countless things you can do with digital that you cant do with an analog scheme.

26

Poweringyour your next design. Powering next design.

Let us be the your power expert. We understand that you dont have time to master every aspect of electronic design. time to have As the master every aspect of electronic a leading manufacturer of power supplies we are here design. to collaborate with you to ensure next project is a success. As a leading manufacturer of your power supplies we are here to collaborate with you to ensure your next project is a success.
Let us be your power expert. We understand that you dont

Novum Advanced Power

Ac-Dc Power Supplies

Dc-Dc Converters

Novum Advanced Power

Ac-Dc Power Supplies

Dc-Dc Converters
www.cui.com/PowerExpert

www.cui.com/PowerExpert

Power Developer

V OLTAGE REGULATORS
Compensation Methods In

An evolution from analog to digital

Bruce Rose, Technical Marketing Manager, CUI Inc
Feedback signals are used in voltage regulator circuits in order to produce a controlled output voltage. When properly implemented, feedback will improve the performance of the circuit. A major contributor to proper implementation of a feedback circuit is the compensation network. This article will give an overview of some methods used to implement compensation in voltage regulators including techniques for automatic digital compensation.

28

TECH ARTICLES

29

Power Developer
The Automotive Analogy
An analogy to assist in understanding compensation in voltage regulators lies in the suspension system of an automobile. Operators of cars desire different styles of ride depending upon the use of the car. Riders in limousines would like to enjoy a smooth ride and not notice any external disturbances. At the other extreme, racecar drivers would like their cars to respond quickly to the external forces of starting, stopping and turning. While both of these styles of cars may experience similar disturbances, different reactions of the cars to the disturbances are desired. As a result, the suspension systems of the cars are tuned to react in the desired manner to specific disturbances. A properly tuned suspension system will give the car its desired ride qualities. Tuning the suspension of a car is similar to adjusting the feedback compensation circuit of a voltage regulator. contains switching FETs (field effect transistors), a circuit to control the switching of the FETs and an output filter that includes inductance and capacitance. The control stage provides the signals to the power stage, such that the switching regulator produces the desired output voltage waveform. The control stage consists of an attenuator, an error amplifier, a gain circuit and a compensation circuit. The switching regulator can either be built with discrete components soldered directly to the host circuit board or obtained from manufacturers that offer voltage regulator point of load (POL) modules with the components placed on a daughter circuit board which is then connected to the host circuit board. Some advantages of POL modules are that much of the voltage regulator circuit design has been done by the module vendor, and the modules can occupy less space on the host board than would a discrete solution. In most analog switching regulators, internal nodes are brought external to the circuit so that the user can select the circuit compensation components. This external compensation feature allows the user to optimize the performance of the switching regulator for their application. Optimizing the voltage regulator transient response involves measuring or modeling the circuit and then calculating the values of the compensation components. The circuit is then modeled or measured with the compensation components installed. This process is often repeated many times until the desired result is achieved. Optimizing the compensation of a digital switching regulator is accomplished in a similar manner; however changes are made using firmware rather than physical components. Proper implementation of the compensation network within an analog-switching regulator requires engineers with special tools, skills and experience. If an analog switching regulator is measured during the compensation design phase, then the circuit board needs to be resoldered many times. If the circuit is modeled and not measured, there is still the need to

Switching Regulators
In order to achieve good power conversion efficiency, design engineers often employ switching regulators (Figure 1). Typical switching regulators consist of two primary functional blocks; a power stage and a control stage. The power stage conducts the current flow in the voltage regulator. It

Figure 1: Analog Switching Voltage Regulator

30

TECH ARTICLES

eventually solder together a physical circuit to measure the performance. The process of re-soldering the compensation components introduces a substantial level of risk to the design process. It is common that the wrong value of a compensation component is installed, another part of the circuit is accidently modified, or the circuit board is damaged during the design process. It is also possible that the circuits drawing power from the voltage regulator can become damaged due to improper compensation of the voltage regulator. When any of these events occur, time delays and expenses are incurred to recognize the problem, identify a solution to the problem, and implement the repair. The above mentioned risks, procedures and required resources exist whether a discrete design of an analog switching regulator is implemented or a POL module based upon an analog switching regulator is used. One can think of the design of the analog switching regulator as a similar process to purchasing a kit car and then selecting and installing all of the suspension components. It takes specific tools, knowledge and experience to properly tune the suspension of a car. Adjusting the suspension in the car poses risks, including damaging the car while driving it with a poorly tuned suspension, or breaking components while making suspension adjustments. In either case, time and resources will be required to repair the damage caused while tuning the suspension. As a result of the increasing number of digital systems being implemented in todays designs, voltage regulator vendors are now offering analog switching regulators with digital wrappers (Figure 2) . The voltage regulator portions of these circuits are very similar to the traditional analog switching regulators. The digital wrapper enables the system to employ software to implement CCM (configure, control and monitor) functions of the voltage regulator, in a limited manner. The ability to use CCM functions in a voltage regulator via software control is of benefit to the design team during the development phase and to the user of the final product.

Figure 2: Analog Switching Voltage Regulator with Digital Wrapper

One can think of the design of the analog switching regulator as a similar process to purchasing a kit car and then selecting and installing all of the suspension components.

31

Developer Analog Power switching regulators with digital wrappers

Vin

are being oered to design engineers for discrete Error amplier and designs and as POL modules. Some module vendors gain stage FET have chosen to include most of the compensation Vout FET Vref + L control components module. The module - internal to the user is then provided a single internal compensation FET C Compensation Network node and is required to select only one resistor and one capacitor to adjust the performance of the Attenuator module. The advantage of this process is that tuning Control Stage Power Stage the performance of the module is simpler than when the user must select all of the compensation components. A trade-o of this compensation technique is that the user is not able to select the Vin complete set of compensation network components. The ability to select all of the compensation components would enable greater optimization of the Error amplier and gain stage FET performance of the voltage regulator. The ability to Vout FET Vref select only+a single resistor and capacitor L is similar control User interface to selecting the shock absorbers for a car, but not bus FET C Compensation being allowed to tune any other component in the Network system. Limitedsuspension digital
congue, control and monitoring Attenuator

engineers (gure 3), providing superior perform to earlier topologies. Similar to analog switching regulators, digital regulators require a control circuit and a power stage. The power stage for a digital switching regulator is similar to that for an analog switching regulator. The control circu in a digital regulator is implemented with digita and mixed-signal circuits. An advantage of this topology is that extensive CCM functions can be implemented. The extensive set of CCM function a digital voltage regulator provides greater bene than the limited CCM functions present in an an switcher with a digital wrapper. Another advanta of digital switching regulators is that optimizing performance of the circuit can be accomplished more easily and automatically.

Control Stage Digital Voltage Power Stage Compensating Regulators

The technical evolution of voltage regulators started with analog switching topologies for increased eciency, transitioning to the addition of digital wrappers for limited CCM functions. Today digital switching voltage regulators are available to design
Vin

Vref

Digital and mixed signal congure, control and monitor

FET FET control FET Vout L C

The compensation function in a digital voltage regulator can be implemented as Proportional, Integral, Dierential (PID) taps, which are coe used in the digital control circuit to dene the response of the voltage regulator. An advantage of using rmware PID taps is that the designer can congure and control the performance of t voltage regulator with software. An innite num changes can be made to the response characte of the circuit without risk of damaging compone or the circuit board. In addition, the behavior of Analog switching regulators with digital systemare can be monitored and the performance wrappers being offered to design engineers for discrete designs and as POL modules. Some throug voltage regulator circuit can be re-tuned module vendors have chosen to include most the life of the product. This ability to easily mod of the compensation components internal to the performance of the voltage regulator is sim the module. The module user is then provided apush-button single internalsuspension compensation node and is is availab tuning which required to select only one resistor and one some cars. capacitor to adjust the performance of the

User interface bus

Control Stage

Power Stage

Bild Switching Voltage Regulator Figure 3:3 ADigital Digital Switching Voltage Regulator

page 4

module. The advantage of this process is that tuning the performance of the module is simpler than when the user must select all of the compensation components. A Some advanced digital regulator controllers trade-off of this compensation technique to automatically compensate th is oer that the theability user is not able to select the complete of compensation network regulator set for optimum performance by monitor components. The ability to select all of the wavefo the characteristics of the output voltage compensation components would enable One advantage of automatic compensation greater optimization of the performance of the voltage regulator. The ability select is that the circuit designer doesto not need any only a single resistor and capacitor is similar to selecting the shock absorbers for a car, but not being allowed to tune any other component in the suspension system.

Auto Compensation

32

TECH ARTICLES
Compensating Digital Voltage Regulators
The technical evolution of voltage regulators started with analog switching topologies for increased efficiency, transitioning to the addition of digital wrappers for limited CCM functions. Today digital switching voltage regulators are available to design engineers (figure 3) , providing superior performance to earlier topologies. Similar to analog switching regulators, digital regulators require a control circuit and a power stage. The power stage for a digital switching regulator is similar to that for an analog switching regulator. The control circuit in a digital regulator is implemented with digital and mixed-signal circuits. An advantage of this topology is that extensive CCM functions can be implemented. The extensive set of CCM functions in a digital voltage regulator provides a much greater benefit than the limited CCM functions present in an analog switcher with a digital wrapper. Another advantage of digital switching regulators is that optimizing the performance of the circuit can be accomplished more easily and automatically. Image 3 A Digital Switching Voltage Regulator The compensation function in a digital voltage regulator can be implemented as proportional, integral, differential (PID) taps, which are coefficients used in the digital control circuit to define the response of the voltage regulator. An advantage of using firmware PID taps is that the designer can configure and control the performance of the voltage regulator with software. An infinite number of changes can be made to the response characteristics of the circuit without risk of damaging components or the circuit board. In addition, the behavior of the system can be monitored and the performance of the voltage regulator circuit can be re-tuned throughout the life of the product. This ability to easily modify the performance of the voltage regulator is similar to push-button suspension tuning which is available in some cars. voltage waveform. One advantage of automatic compensation is that the circuit designer does not need any special tools, knowledge or experience to optimize the performance of the voltage regulator. In a regulator with analog compensation components, the compensation must be set such that the output voltage characteristics are acceptable over changes due to initial component tolerances, aging, temperature, input voltage and many other factors. This means that the circuit is never operating at the optimum performance point. Digital voltage regulators with automatic compensation enable the voltage regulator to operate at peak performance regardless of changes in the system. Automatic compensation of digital voltage regulators can bethought of as having an expert mechanic always in the car to optimize the ride without any burden on the driver or passengers. CUI, Inc. is the only POL module manufacturer in the industry to offer multiple families of commercially available digital voltage regulator modules with automatic compensation. Proper compensation of voltage regulators enables users to realize optimum performance from their circuits. Tuning the performance of a circuit using traditional analog switching regulators involves a substantial level of risk. Vendors of some analog voltage regulatorbased POL modules offer products that simplify the task of compensation by limiting the choices available to the user. Conversely, digital voltage regulators enable firmware based CCM functions, which permit the voltage regulator to operate at optimum performance. All of these topologies traditionally require a design team with special tools, knowledge and experience in power supply design to create an acceptable solution. The multiple families of digital POL modules from CUIs Novum Advanced Power line incorporate automatic compensation, allowing system designers ease of use and superior performance in next generation applications.

Auto Compensation
Some advanced digital regulator controllers offer the ability to automatically compensate the regulator for optimum performance by monitoring the characteristics of the output

33

Power Developer

Perfection in P

Asymmetric IGBT Modul

In order to move to renewable sources of energy from wind and the sun, we nee powerful inverters. IGBTs are the key components here. They also play an importan role when it comes to power electronics, such as controls for electric motor where IGBTs (Isolated Gate Bipolar Transistors) have become indispensable.

IGBTs are also widely used to switch extremely high loads at several 1000V. Fo such applications, small DC/DC inverters with excellent insulation properties a a must, as they guarantee a long service life and make green energy available t the next generation.

34

TECH COLUMN

Power

les for Smooth Switching

Thomas Rechlin Senior FAE for Europe at RECOM Engineering Gmunden, Austria

ed nt rs,

or are to

35

Power Developer

The photovoltaic plant shown in Figure 1 is a typical example of an IGBT application, and worth examining in more detail. IGBTs are used here twice. Photovoltaic panels generate fluctuating voltages, depending on the time of day and the weather, making it necessary to establish an intermediate circuit. In order to obtain a stable intermediate circuit voltage of typically 800VDC to 1000VDC, IGBT boost converters are installed. Normally, they come with MPP (Maximum Power Point) tracking. MPP ensures that each PV module is operated at optimum efficiency, irrespective of the sunlight available. The IGBT is operated at a variable switching frequency, which might be as high as 300kHz. It is obvious that this puts a huge strain on the associated components.

As, for obvious reasons, the intermediate circuit power cannot be directly fed to the grid, a second IGBT module is needed, acting as an inverter. Two IGBT pairs are controlled through a bridge circuit with a phase opposite PWM signal. To obtain a frequency as close as possible to 50Hz sine, the control frequency must be around 10kHz to 20kHz. The downstream LC filter smoothes the output voltage so that it can be fed to the grid without any problems.

Control of IGBTs - what to look out for

We now know how IGBTs work. What might not yet be clear is the purpose of the DC/DC converter, as this only becomes apparent after examining the control of the IGBT more closely. The IGBT is controlled

36

TECH COLUMN

by means of an IGBT driver. This component is integrated into the power circuit and floats at the respective voltage (see Figure 1). It must therefore be insulated from the control circuit. This is done by means of an optocoupler for the control signal, while the supply line is insulated by means of two reinforced insulated DC/DC converters. Why two converters? Two converters are needed due to the properties of the IGBT, which is essentially a hybrid between a MOSFET (metal oxide semiconductor field-effect transistor) and a bipolar transistor. It has been devised for the switching of high powers at minimum loss. This is achieved by ensuring that the gate capacity is charged as quickly as possible, which, however, results in extremely high current peaks (di/dt), posing massive stress on the components. The gate resistance RG permits maximum switching speeds at just about tolerable di/dt values. This is the situation during switching on. For switching off, the opposite applies. The gate must be depleted as quickly as possible. This is made possible by a negative control voltage

VG. In connection with symmetric supply (where +15V are required for reliable switching on of an IGBT), we are looking at -15V, which appears reasonable. The fast depletion of the gate, however, produces extremely high voltage peaks (dv/dt), which again tend to reduce the service life of components. The problem can be solved by reducing the control voltage upon switching off. Experience shows that a VG- of -9V is most suitable, as the gate is still reliably depleted while the dv/dt values remain acceptable. These relationships are summarized in Table 1 above.

Fig. 1: Function diagram of solar system with two IGBT stages. Both the boost converter and the inverter must be controlled through reinforced insulated DC/DC converters in order to ensure proper insulation from the high-voltage side.

37

Power Developer

Fig. 2: Right: general design of IGBT driver. Left: current and voltage curves during switching on/ off. The curves clearly show the positive effect of a lower switch-off voltage on the dv/dt load.

Figure 2 shows the principle behind an IGBT driver circuit. The current and voltage curves during switching on and off illustrate the problem faced with IGBT control. Engineers basically have two options: opting for the compact design with one converter and 15V, with all associated disadvantages when it comes to switching off; or choosing the more efficient solution with two converters (15V and -9V respectively), which is also more expensive.

this. Given the high switching speeds and the associated steep dv/dt switching edges, the calculated insulation voltage is not even near what is required. As these peaks occur for only a few s, the effect on the insulating capacity of the converter is not immediately detectable. However, little strokes fell great oaks, as the saying goes, and it therefore comes as no surprise that the insulation suffers in the long run so that the service life of the component is significantly reduced. For the level of insulation, which in essence describes the type of insulation, air and creepage distances are a major concern. Generally, the insulation of IGBT drivers is designed for 50Hz. However, much higher frequencies up to several hundred kHz are not unusual in IGBT applications. Such high frequencies can affect the electromagnetic properties of the transformer materials in a way that is entirely unforeseeable. In addition, there are often parasitic switching capacities, triggered by the steep switching edge. Standard insulation consisting of a layer of lacquer on the transformer wires is therefore simply insufficient. To improve safety, double insulation or what is called basis insulation, where the varnished wires are complemented with other barriers, is required.

Criteria for the Selection of the Right DC/DC Converters

To overcome the above problems, engineers have come up with IGBT converters. These feature asymmetric outputs of normally +15V / -9V. It is thus possible to supply the IGBT driver with the optimum voltages for operation without excessive loads. The second important criterion is insulation, characterised by the insulation voltage and the level or type of insulation. At first sight, the calculation of the insulation voltage seems to be quite easy. According to the well-known rule of thumb, the insulation voltage should be twice the intermediate circuit voltage. There is, however, more to

38

TECH ARTICLES
In conclusion, the insulation voltage should be significantly higher than the expected peak voltage. This can be achieved with basis insulation or reinforced insulation, which is the preferred option, which makes the IGBT converter much more reliable. Another obstacle that you might encounter are the sometimes contradictory specifications of the insulation in the data sheet of the various manufacturers. To guide you through this jungle, RECOM has developed a userfriendly tool: the Isolation Calculator (figure 4). It allows you to compare like with like and to find products that meet your requirements.

Modern IGBTs Covering Wide Range of Use

To meet the requirements of the industry, RECOM has developed seven new IGBT converter ranges, which were launched at this years PCIM trade fair in Nuremberg. All models feature asymmetric outputs (+15V / -9V) required for the control of IGBT drivers at input voltages of 5V, 12V or 24V. The series were designed to provide optimum insulation at the various voltage levels. Starting at 3kV (RH-xx1509D) up to 6.4kV (RxxP1509D), the converters provide matching insulation voltages for virtually all applications. The RECOM engineers also paid close attention to the space requirements for the IGBT converters. The new models are available with a compact SIP7 housing (RP-xx1509D), the universal DIP14 housing (RKZ-xx1509D), or as DIP24 miniature units (RV-xx1509D) where space is restricted. These 1W and 2W modules are certified according to EN60950-1 and do not contain any hazardous substances (RoHS2 and REACH compliant). All modules come with a threeyear manufacturer warranty, as is standard for RECOM products.

Fig. 3: Supply of IGBT driver through DC/DC converter with asymmetric output.

CLICK HERE
to comment on the article.
Fig. 4: This useful CD is available free of charge from RECOM. Alternatively, access the Insulation Calculator from the RECOM website www.recom-international.com.

39

M o v i n g To w a r d s
a

David Elien
VP of Marketing & Business Development, Cree, Inc.

FUTURE
Hugo van Nispen, COO of DNV KEMA

Clean Energy

LIGHT
How Cree reinvented the light bulb

Let There Be

Cutting Edge

Modeling

SPICE

+
MCU Wars
32-bit MCU Comparison

Cutting Edge Flatscreen Technologies New LED Filament Tower

View more EEWeb magazines Click Here

Power Developer
O ct o b er

201 3

Sierra Circuits:
A Complete PCB Resource
Ken Bahl
CEO of Sierra Circuits

Reality
Designing for
TQ-Groups Comprehensive Design Process

From Concept to
+

Wolfgang Heinz-Fischer Head of Marketing & PR, TQ-Group

Durability
+

Freescale and TI Embedded Modules

PLUS: The Ground Myth in Printed Circuits

+
PCB Resin Reactor

ARM Cortex Programming

Low-Power Design Techniques