featured product:
Cadence Incisive Palladium Dynamic Power Analysis
Voice over Wi-Fi Implementation with Single Stream
802.11n
low-power wireless: Near-Field Communications portable power: Auto-Zero Op Amps
NEW Section product focus: Low-Power Memory
September 2008
www.portabledesign.com
An RTC Group Publication
CEO Interview: Necip Sayiner, Silicon Labs
contents 14 device vendors in the future
analysts’ pages 12
Voice-over-Wi-Fi Implementation 16 with Single Stream 802.11n
Narasimhan Venkatesh, Peddi Indukuri and Subba Reddy, Redpine Signals
STA1
A-MPDU (TA1)
STA2
STA1 UTT
RIFS or SIFS
SIFS
A-MPDU (RA2)
19 psmp frame exchanges
MPBA (TA2) (TID1, TID2)
cover feature
AP
Data (TID1) Data (TID2) MTBA (TA1) alUStime or SIFS
NEW SECTION product focus 32 product focus 32 products for designers 38 design idea 42 product feature 43
SIFS
Initial PSMP sequence Downlink Phase Uplink Phase Data (TID1) Data (TID2)
NEW SECTION analyize this 10 analyze this 10
A-MPDU (TA2)
editorial letter 5 industry news 6
AMPDU (RA1)
departments
PSMP MorePSMP=1
STA2 UTT
1000
wireless communications
Near Field Communications 22 Steve Rackley
portable power Auto-Zero Op Amps: 28 Inherent Benefits in Portable Signal-Conditioning Applications Kevin Tretter, Microchip
PHY layer data rate (Mbps)
WUSB (Optional) WUSB (Mandatory)
100
25
IrDA VFIR
10
NFC range vs. 1
Bluetooth Class 1
data rate
B C
+5V 0.1 0.1
IrDA SIR 1
ceo interview
10 Range (m 1 kΩ
Necip Sayiner 44 Silicon Labs
F.S. = 10 mV
30 near field communications
September 2008
3
1 kΩ
team editorial team
Editorial Director Editor-in-Chief Managing Editor Copy Editor
Creative Director Art Director Graphic Designer Director of Web Development Web Developer
Warren Andrews, warrena@rtcgroup.com John Donovan, johnd@rtcgroup.com Marina Tringali, marinat@rtcgroup.com Rochelle Cohn
art and media team Jason Van Dorn, jasonv@rtcgroup.com Kirsten T. Wyatt, kirstenw@rtcgroup.com Christopher Saucier, chriss@rtcgroup.com Marke Hallowell, markeh@rtcgroup.com James Wagner, jamesw@rtcgroup.com
management team
Associate Publisher Product Marketing Manager Western Advertising Manager Western Advertising Manager Eastern Advertising Manager
Circulation
Chief Executive Officer Vice President Vice President of Finance Director of Corporate Marketing Director of Art and Media
Marina Tringali, marinat@rtcgroup.com Aaron Foellmi, aaronf@rtcgroup.com Stacy Gandre, stacyg@rtcgroup.com Lauren Trudeau, laurent@rtcgroup.com Nancy Vanderslice, nancyv@rtcgroup.com Shannon McNichols, shannonm@rtcgroup.com
executive management Untitled-1 1
4/13/08 3:57:29 PM
John Reardon, johnr@rtcgroup.com Cindy Hickson, cindyh@rtcgroup.com Cindy Muir, cindym@rtcgroup.com Aaron Foellmi, aaronf@rtcgroup.com Jason Van Dorn, jasonv@rtcgroup.com
portable design advisory council
HOW WELL DO YOU KNOW THE INDUSTRY?
Ravi Ambatipudi, National Semiconductor Dave Heacock, Texas Instruments Kazuyoshi Yamada, NEC America
corporate office The RTC Group 905 Calle Amanecer, Suite 250 San Clemente, CA 92673 Phone 949.226.2000 Fax 949.226.2050 www.rtcgroup.com
WWW.EMBEDDEDCOMMUNITY.COM
BPA Worldwide believes that consumer media owners, advertisers and agencies should expect more from audits. That’s why we build on the foundation of our worldwide reputation for accuracy and integrity, adding even greater value with reporting transparency and depth, as well as auditing timeliness.
For reprints contact: Marina Tringali, marinat@rtcgroup.com. Published by the RTC Group. Copyright 2008, the RTC Group. Printed in the United States. All rights reserved. All related graphics are trademarks of the RTC Group. All other brand and product names are the property of their holders. Periodicals postage at San Clemente, CA 92673 and at additional mailing offices. Postmaster: send changes of address to: Portable Design, 905 Calle Amanecer, Suite 250, San Clemente, CA 92673. Portable Design(ISSN 1086-1300) is published monthly by RTC Group 905 Calle Amanecer, Suite 250, San Clemente, CA 92673. Telephone 949226-2000; 949-226-2050; Web Address www.rtcgroup.com. Ride along enclosed. embeddedcommad_14v.indd 1 4 PORTABLE DESIGN
11/13/06 5:55:59 PM
editorial letter
A
As I write this, Texas is just starting to dig out from under the wreckage left by Hurricane Ike. Thanks to mandatory evacuations and lessons learned post-Katrina, there was very little loss of life, though a tremendous amount of property damage. And unlike Katrina, Ike never became a communications emergency, which is what usually happens when transmission lines and cell towers go down and landlines are cut. First responders from a wide range of agencies were able to synchronize their actions through a complex of wireless networks. The State of Texas’ emergency agencies had carefully planned for an event like Ike, setting up numerous interlocking networks on HF, VHF and UHF frequencies. Then they ran simulated emergency tests. Locally those ran out of the state’s Combined Transportation, Emergency and Communications Center (CTECC), a 75,000 sq. ft. facility in Austin. President Bush visited CTECC as Ike approached, congratulating them on having their act together. They certainly did. CTECC is a partnership involving the City of Austin (911 dispatch, police, fire, EMS), Travis County (911 dispatch, sheriff, constable), the Capital Metropolitan Transportation Authority (the local bus system), and the Texas Department of Transportation (TxDOT) for the freeway. This bunkered facility has a range of transmitters and an antenna farm that few can equal. It’s also a strong argument for bringing local and regional wireless communications under one roof. APCO 25 The real eyes and ears of a facility like CTECC and the Houston emergency operations center (EOC) are the mobile units in the police cars, fire trucks, ambulances, Red Cross vans and National Guard helicopters. In the past, few of these units could talk to anyone other than their base stations; certainly not to different types of units; and often not even to similar units (police, fire) from out of the area. So after a disastrous series of fires in Southern California, during which police and fire from different jurisdictions—even neighboring towns—were unable to communicate with each other, the Association of Public-Safety Communications Officials International, Inc. (APCO International) started Project 25, an industry-wide effort to set voluntary standards of uniform digital two-way radio technology for public safety organizations. APCO 25 provides a roadmap for conversion of existing analog systems to digital systems
using compatible modulation formats. The advantages it brings include improved spectrum efficiency; competition among multiple vendors through an Open Systems Architecture; and more effective, efficient, and reliable intra-agency and inter-agency communications. APCO 25 falls short of an SDR any-frequency, any-protocol radio, but it’s a big step forward.
Wireless Where It Matters john donovan, editor-in-chief
A basic requirement for new digital radio equipment is backward compatibility with standard analog FM radios. This supports an orderly—thus not terribly fast—migration into mixed analog and digital systems. Full compatibility is still a ways down the road, though not nearly as far as it used to be. SDR will be here some day; APCO 25 is here now. APCO 25 conversion is well under way, thanks to grants from the Department of Homeland Security, who won’t finance any equipment that isn’t APCO 25 compatible. Let’s hear it for the hams Having written a paean to amateur radio in this space a couple of months ago, permit me a quick tip of the hat to my fellow hams, who played an important communications backup role as Ike passed through Texas. During Katrina, hams were often the only communications link on the ground. Fortunately during Ike that wasn’t the case, though the Amateur Radio Emergency Services (ARES) throughout the Delta and West Gulf Divisions were on full alert, running HF, VHF and UHF nets using every available analog and digital mode; sharing SKYWARN and NOAA alerts; manning stations at local EOCs and hospitals; and, in the immediate storm area, passing message traffic into the hams’ National Traffic System (NTS) and providing backup communications for first responders. This is wireless where it matters. Nice work, people.
September 2008
5
news
The Androids Among Us
The buzz is we’ll have Androids in time for Christmas. The FCC has approved the HTC Dream, the first handset based on Google’s Android mobile platform. According to an AFP re-
port, T-Mobile will launch the device, probably in time for Christmas sales. The Washington Post reports that the HTC Dream will feature a touchscreen, Wi-Fi, a BlackBerry-style “jogball,” a Safari Web browser and Google applications such as Gmail, Maps and YouTube. nd Meanwhile, the Android Software Development Kit (SDK) is still in beta test, though er exploration ether your goal Google recently released version 0.9, which it speak directly claims is “now pretty stable and we don’t expect ical page, the ght resource. major changes.” Still, this is a pretty unproven technology, platform on which companies are expected to es and products make multi-million dollar bets. Still, the lure of ed avoiding paying licensing fees to Microsoft or Symbian for their mobile OSs has induced a lot of big players to sign on with Google’s Open Handset Alliance. Android may take a while to get its legs under it, but it’s certainly a well-funded startup companies providing solutions now and already creating huge waves in exploration into products, technologies and companies. Whether your goal is to research theone latestthat’s datasheet from a company, mp to a company's technical page, the goal of Get Connected is to put you in touchthe withhandset the right resource. level of it’s finalized. marketWhichever even before gy, Get Connected will help you connect with the companies and products you areThis searching willfor. be an interesting ride. onnected --John Donovan, Editor-in-Chief Google, Inc., Mountain View, CA. (650) 253-0000. [www.openhandsetalliance.com].
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
6
Spansion Unveils Plans for Highest Performance, Smallest NAND
NOR flash market leader Spansion Inc. has unveiled plans for its upcoming MirrorBit ORNAND2 product family, designed to support 25
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
percent faster write performance and up to twice the read performance at a significantly smaller die size than today’s floating-gate NAND. With the MirrorBit ORNAND2 family, Spansion plans to further diversify its product portfolio to address a broader portion of the high-value wireless and embedded markets, with initial single chip densities ranging from 1 Gbit to 4 Gbit at both 1.8 and 3.0 volts using a Single Level Cell (SLC) architecture. The MirrorBit ORNAND2 product family leverages the company’s proprietary MirrorBit charge-trapping technology and features a SONOS-like cell connected in a true NAND array. The technology features 25 percent fewer process steps than the first-generation MirrorBit ORNAND and MirrorBit NOR solutions, enabling significantly lower cost. Spansion’s MirrorBit ORNAND2 products
based on an SLC architecture, 3.0V operating voltage and an industrial temperature range, are intended for embedded applications such as plug-in the wall, car multimedia, networking infrastructure and industrial control systems. Spansion expects to integrate “managed NAND” controller and interface functionalities with MirrorBit ORNAND2 memory on a single chip, for the creation of a full product family of differentiated solutions. The integration of a managed NAND interface will help simplify system level design to complement industrystandard chipsets and microcontrollers. In the handset segment, Spansion plans to expand its product portfolio with the addition of MirrorBit ORNAND2 solutions. Customers will be able to utilize the high read performance
of Spansion MirrorBit NOR and MirrorBit Eclipse products for fast code execution, and add the fast write performance and high capacity of MirrorBit ORNAND2 products for data storage. MirrorBit ORNAND2 offers handset platforms designed using a “Store and Download” (SND) architecture low-latency demand paging for faster application loading. Spansion Inc., Sunnyvale, CA. (408) 962-2500. [www.spansion.com].
Accellera Board Approves New Version of Analog, MixedSignal Standard
The Accellera Board of Directors and Technical Committee members—systems, semiconductor and design tool companies— have approved a new version of its VerilogAnalog Mixed-Signal (AMS) standard, Verilog-AMS 2.3, as an Accellera standard for analog and mixed-signal design and simulation. The new Verilog-AMS standard unifies the Verilog-AMS 2.2 specification with the IEEE Std. 1364-2005 or Verilog hardware description language (HDL) standard. Verilog-AMS 2.3 enables users to develop standard and tightly integrated Verilog-AMS modules and allows EDA software tool developers to implement EDA tools without ambiguities in the language interpretation. Verilog-AMS 2.3 encompasses analog and mixed-signal extensions to IEEE Std. 1364, which is widely used today for digital circuit design and verification. The previous Accellera Verilog-AMS standard, VerilogAMS 2.2, was approved in 2005. Apart from IEEE-1364 integration, Verilog-AMS 2.3 introduces new analog and mixed-signal features to support and enable improved top-down AMS design and verification methodologies. These include enhancements to table_model, support for
multiple analog blocks, and resolution of language conflicts with the SystemVerilog IEEE Std. P1800, such as changing the digital domain name to “ddiscrete” from “logic” as logic is a keyword in SystemVerilog, and making the usage of array literals consistent. The next phase of Accellera’s AMS technical activities will include integration of the AMS standard with the SystemVerilog language, IEEE Std. P1800, and extensions to the AMS language for mixed-signal assertions and behavioral modeling support.
“By integrating our front-end modules with Ember’s ZigBee silicon, we are offering our customers a high-performance plug-and-play solution that is truly unmatched worldwide.” Skyworks’ SKY65336 is a high-efficiency 8 x 8 mm multi-chip module (MCM) FEM for ZigBee and other 2.4 GHz industrial, scientific and medical (ISM) band applications. The SKY65337 is a FEM that is footprint compatible with the SKY65336, with the LNA removed for applications that require lower cost. Ember’s EM250 system-on-chip (SoC) integrates an IEEE 802.15.4 radio, a 16bit microprocessor, flash, random access memory (RAM) and peripherals delivering excellent performance at a low system cost. Ember’s EM260 network co-processor combines a 2.4 gigahertz (GHz) IEEE 802.15.4-compliant radio transceiver with a flash-based microprocessor running the EmberZNet PRO ZigBee stack.
Accellera Organization Inc., Napa, CA. (707) 251-9977. [www.accellera.org].
Skyworks and Ember Partner to Develop Front-End Modules for ZigBee Applications
Skyworks Solutions and Ember have announced that they are partnering to develop the industry’s first portfolio of ZigBee frontend modules (FEM) targeting applications such as smart meters in energy management, home area networks and industrial automation. “Skyworks’ partnership with Ember accelerates our efforts in the energy management industry by more than doubling our addressable market for front-end modules,” said Stan Swearingen, Skyworks’ vice president and general manager of Linear Products in a statement.
West Technology Research Solutions is forecasting the ZigBee/IEEE 802.15.4 market to grow at an annual rate of over 117 percent— from approximately 8.4 million units shipped in 2007 to as many as 516 million in 2012. Skyworks Solutions Inc., Woburn, MA. (781) 376-3000. [www.skyworksinc.com]. Ember Corporation, Boston, MA. (617) 951-0200. [www.ember.com].
September 2008
7
news
Ire at IR
On the surface it seemed to make sense. On August 15 Vishay made an unsolicited offer of $1.6 billion to acquire International Rectifier, a venerable company whose discrete powermanagement IC product lines would complement Vishay’s. IR had fallen on hard times, their stock had tanked, and Vishay apparently saw this as an acquisition opportunity. IR, in turn, just saw it as opportunistic and rejected the offer out of hand. As Richard J. Dahl, IR’s Chairman of the Board, put it in his response, “Your proposal is inadequate, opportunistic and not in the best interests of International Rectifier and its shareholders.”
nd
er exploration ether your goal speak directly ical page, the ght resource. technology, es and products
ed
Not being one to give up easily, on September 9 Vishay increased its offer slightly to $1.7 companies providing solutions now billion and announced its intention to nominate exploration into products, technologies and companies. Whether your goal is to research the latest datasheet from a company, mp to a company's technical page, the goal of Get Connected is to put you in touchthree with the right resource. level of candidates to Whichever IR’s board. This is reminisgy, Get Connected will help you connect with the companies and products you arecent searching for. Icahn’s move on Yahoo’s board afof Carl onnected ter they repeatedly rejected Microsoft’s buyout offer. Jerry Yang still has his job but no longer his job security. Whether Oleg Khaykin, IR’s CEO, winds up in the same position remains to be seen. Vishay rightly sees an IR acquisition as complementing its product lines, though whether their cultures are complementary is unclear. What is clear is that Vishay can’t afford to leverage itself too much further in order to top up its offer, though if it did get the support of both Get Connected with companies mentioned in this article. boards, it could raid IR’s cash kitty of $700 www.portabledesign.com/getconnected million to complete the financing.
End of Article
8
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
An old line company with strong market share, depressed stock and a big cash stash is a tempting takeover target. IR is a proud company and wants to work its way out of the doldrums. Vishay is facing a market with shrinking liquidity, so raising the cash won’t be easy. Stay tuned to see how this shakes out. --John Donovan, Editor-in-Chief International Rectifier Corporation, El Segundo, CA. (310) 726-8000. [www.irf.com]. Vishay Intertechnology, Inc., Malvern, PA (402) 563-6866. [www.vishay.com].
Digital Signal Controllers Analog Serial EEPROMs
Microchip Technology’s new UNI/O serial EEPROM uses only ONE connection to the host microcontroller. This compares to two or three pins for I�C™ and three to six pins for Microwire or SPI buses. This new, flexible bus offers advanced features like a status register and write protection on demand, along with all I/O, data and command functions through a single pin. Simplify designs & reduce system cost
t Free up pins on the MCU – Enhance your design by adding new features – Move to a smaller MCU = lower cost t Free up pins on your connector – Smaller connector = lower cost t Compact: Tiny packages and no pullup resistors
Optimized for embedded applications t t t t t
Software Write Protection ¼, ½, or full array Flexible data rate 10 - 100 kHz set by host 1-million E/W cycles, 200-year data retention Low standby current - 1 μA Real EEPROM, up to 125˚C
Device 11AA010 11LC010 11AA020 11LC020 11AA040 11LC040 11AA080 11LC080 11AA160 11LC160
Density Operating (bits) Voltage 1K 1.8-5.5V 1K 2.5-5.5V 2K 1.8-5.5V 2K 2.5-5.5V 4K 1.8-5.5V 4K 2.5-5.5V 8K 1.8-5.5V 8K 2.5-5.5V 16K 1.8-5.5V 16K 2.5-5.5V
www.microchip.com/UNIO
GET STARTED TODAY!
Purchase the MPLAB® Starter Kit for Serial Memory Products (DV243003), which supports the UNI/O family of serial EEPROM and all other Microchip serial EEPROMs for $79.98 at www.microchipDIRECT.com.
Microchip Direct... 2nd line
The Microchip name and logo, the Microchip logo, dsPIC, MPLAB and PIC are registered trademarks of Microchip Technology Incorporated in the U.S.A. and other countries. UNI/O is a trademark of Microchip Technology Incorporated in the U.S.A. and other countries. All other trademarks mentioned herein are property of their respective companies. © 2008, Microchip Technology Incorporated. All Rights Reserved.
Microcontrollers
NEW Single-wire EEPROM Family In a Tiny 3-lead Package!
analyze this
L
Like many other Laptop Power Nomads, I’ve migrated from a luggable laptop that weighed north of 8 pounds (they never list the weight of the power module) to a 4-pound notebook that has circled the globe with me. I’m now looking for something newer and lighter, much lighter, and one that has battery life that will allow me to continue working as I fly across the ocean. All along, I’ve been accompanied by a succession of cell phones, from an early twochannel GSM voice-centric one to my current quad-channel HSDPA unit with a 2 megapixel camera and 2 gigabytes of storage for pictures, video and address book. More than once, I needed a phone number of someone not listed in my cell phone. To crank up my notebook
Mobile Internet Devices on the Horizon
will strauss, forward concepts
to access my more extensive database took an agonizing (and sometimes embarrassing) several minutes waiting for Windows to come alive. In short, I need a communications device that can handle a number of PC functions. But, hope is on the way in two forms. First out of the gate is a gaggle of Ultra-Mobile PCs (UMPCs). UMPCs were engendered by Intel’s new Atom chipset and based on the legacy X86 instruction set architecture, while dissipating a fraction of the power of its Pentium-legacy brethren, thereby eliminating the need for fan cooling. These devices are relatively inexpensive mini-notebooks that 10
PORTABLE DESIGN
have smaller screens and smaller keyboards. Some run Microsoft Vista, while others run Windows XP or a variant of Linux. There are a number of UMPC form factors, from conventional clamshells to tablets to accordionfold devices. The other form that’s getting increasing attention are so-called Mobile Internet Devices (MIDs). Unlike UMPCs, there’s no thought of hard disk drives, since they are considered relatively inexpensive pocketable communications-centric devices with very long battery life that have full wireless Internet access with 4-inch or larger screens and VGA or better resolution. In short, it’s a Smartphone on steroids. Most early units will feature 3G cellular, Wi-Fi, Bluetooth and GPS. But doesn’t an Apple iPhone have these capabilities? Yes, but with a smaller, lower-resolution screen, a less inviting keyboard and lack of some desirable features, like TV-out or an optional Mobile TV capability. Because MIDs are expected to be fully operable for 24+ hours and with standby capabilities measured in days, the current Intel Atom chipset is inadequate from a power consumption standpoint. But, Intel promises that next year they will have a worthy contender with their Moorestown chip...which they claim will dissipate a tenth the power of its current Atom solution. It will be an SoC (system-on-chip) design scheduled for release in mid-2009. In contains an integrated CPU, graphics, video and memory controller on a single die and is expected to be 50% smaller in size than the Atom processor. Meanwhile, competing SoC chips based on ARM processor cores are now being fielded by companies like Qualcomm, Nvidia and Texas Instruments who claim their solutions already consume a tenth the power of the current Atom. Low-power MIDs based on these ARM-based solutions are not expected to hit the street until late this year or early in 2009, but they will have several months to eat into the emerging Atom-based market. We don’t envision MIDs as ever employing Microsoft’s Vista operating system because of its heavy drain on resources of memory and battery; however, we believe that Microsoft has a MID play with future versions of Win-
analyze this About the Author: Will Strauss is the president of Forward Concepts (www.fwdconcepts.com) and is considered an authority on digital signal processing and wireless markets.
MID & UMPC Comparison
Parameter
MID
UMPC
Target Market
Consumer-class always-on mobile lifestyle device.
Business-class mobile PC: typically shut down after each use.
Processor
SoC-based low-power processor core, such as the ARM Cortex A8.
X86 PC architecture, like Intel Atom or Via Nano
OS
Windows Mobile or Linux.
Full PC OS, typically Windows XP or Vista.
Application
Optimized for 3G wireless communications, multimedia and web surfing with proprietary applications.
Optimized for enterprise-grade, standard applications like Outlook, Word, Excel with focus on security. 3G is an option
Boot-up
Instantaneous (2-4 seconds)
PC-style extended boot cycle
Display
4-to 6-inch display with touch screen capability
5-to 10-inch display with touch screen capability
Storage
1–8 GB flash
4-to 8-GB flash (low-end), 20+GB HDD, 32-64 GB SSD (high-end).
Keyboard
Soft or hard keyboard
Predominantly built-in hard keyboard
Battery Life
24+ hours under normal use
6-8 hours under normal use
Price
$300-$500 (low-end), $750+ (highend)
$500-$650 (low-end), $2000+ (highend)
Source: Forward Concept
MOBILE INTERNET DEVICE MARKET 40
35
SHIPMENTS (M units)
dows Mobile. Of course, Linux will be popular in MIDs for its lower processing overhead and tighter OEM control. Intel claims that only Windows-based X86 platforms offer true Internet access, but availability of the Firefox browser and Adobe Flash for Linux mutes that claim. As an always-on communications-centric device, we believe that 3G cellular is also required for MIDs, although some point out that Wi-Fi fulfills the wireless Internet access need. But we don’t see Wi-Fi alone meeting the always-on, always-connected need. In that regard, Qualcomm seems to have the cellular upper hand, with both (ARM Cortex A8based) Snapdragon core and 3G & GPS modem capability with their QSD8250 (HSPA) and QSD8650 (HSPA & 1xEV-DO) products. Intel, Nvidia and even TI will likely have to acquire their 3G modems from another party (like Qualcomm, Infineon, or Icera, perhaps through a modem module vendor like Option, Novatel Wireless or Sierra Wireless). With a market size predicted by Forward Concepts to be almost 40 million MID units in 2012, there’s strong incentive to enter this market. Intel claims to be first with OEM MID products on the street, but, collectively, Qualcomm, Nvidia and TI claim to be working on over three dozen OEM product designs with their SoC MID solutions, so it will soon be exciting times in the mobile Internet marketplace.
30
25
20
15
10
5
0
2008
2009
2010
2011
2012
Source: Forward Concept
September 2008
11
nd
er exploration ether your goal speak directly ical page, the ght resource. technology, es and products
ed
analysts’ pages Top 20 Semiconductor Supplier Ranking
IC Insights’ recent research uncovered a big shakeup in the 1H08 top 20 semiconductor supplier ranking (Figure 1). There are eight U.S. companies in the top 20 (including three fabless semiconductor suppliers), six Japanese, three European, two South Korean and one Taiwanese company (IC foundry supplier TSMC) in the ranking. As shown, it required at least $2.1 billion in 1H08 sales to make the top 20 ranking. Although the top four ranked companies remained the same, there were a number of “movers and shakers” up and down the remainder of the 1H08 ranking as compared to their full-year 2007 positions. 1 H08 Top 20 Semiconductor Sales Leaders ($M) 1H08 Rank
1 2 3 4
2007 Rank
1 2 3 4
Company
Intel Samsung TI Toshiba
5
6
TSMC*
6 7 8 9
5 8 7 9
ST** Renesas Hynix Sony
10
14
Qualcomm***
11 12 13 14
12 15 13 10
Infineon NEC Micron NXP
15
11
AMD
16 17 18 19
16 17 21 19
Freescale Fujitsu Panasonic Nvidia***
20
23
Broadcom***
-
-
Total Top 20
*Foundry
Headquarters
U.S. South Korea U.S. Japan
1Q07 Tot
2Q07 Tot
1H07 Tot
1Q08 Tot
2Q08 Tot
1H08 Tot
Semi
Semi
Semi
Semi
Semi
Semi
8,072 4,697 3,115 3,249
7,916 4,552 3,257 2,510
15,988 9,249 6,372 5,759
8,841 5,569 3,191 3,055
8,655 5,618 3,175 2,789
17,496 11,187 6,366 5,844
1H08/1H07 % Change 9% 21% 0% 1%
Taiwan
1,922
2,258
4,180
2,767
2,894
5,661
35%
Europe Japan South Korea Japan
1,953 1,949 2,569 1,716
2,086 1,984 1,998 1,573
4,039 3,933 4,567 3,289
2,179 2,142 1,675 1,946
2,391 2,195 1,824 1,484
4,570 4,337 3,499 3,430
13% 10% -23% 4%
U.S.
1,259
1,367
2,626
1,620
1,762
3,382
29%
Europe Japan U.S. Europe
1,282 1,346 1,370 1,390
1,361 1,366 1,328 1,461
2,643 2,712 2,698 2,851
1,566 1,475 1,391 1,443
1,580 1,523 1,529 1,445
3,146 2,998 2,920 2,888
19% 11% 8% 1%
U.S.
1,233
1,378
2,611
1,505
1,349
2,854
9%
U.S. Japan Japan U.S.
1,295 1,047 870 847
1,310 1,061 905 896
2,605 2,108 1,775 1,743
1,304 1,226 1,058 1,158
1,389 1,108 1,170 985
2,693 2,334 2,228 2,143
3% 11% 26% 23%
U.S.
**Not incl. flash in 1H07 & 1H08
897
892
1,789
986
1,153
2,139
20%
42,078
42,078 41,459
83,537
46,097
46,018
92,115
10%
***Fabless
companies providing solutions now
Climbers exploration into products, technologies and companies. Whether your goal is to research the latest datasheet from a company, mp to a company's technical page, the goal of Get Connected is to put you in touch with resource. of • Cthe ellright phone IC Whichever supplierlevel Qualcomm used a gy, Get Connected will help you connect with the companies and products you are searching 29%for. year-over-year 1H08 growth rate to
onnected
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
12
jump four spots and rank as the 10th largest semiconductor supplier in 1H08. • The third largest fabless supplier, Broadcom, jumped three positions and is now positioned as the 20th largest semiconductor supplier in the world. • Japanese consumer electronics powerhouse Panasonic (formerly Matsushita) and NEC were boosted by the strong Japanese yen versus the U.S. dollar. Each jumped three places and moved into the 18th and 12th positions, respectively, in the top 20 ranking.
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
• Moving up one place each in 1H08 were TSMC, Renesas and Infineon.
“Descenders”
• DRAM-supplier Qimonda’s nightmare continued in 1H08 as the company dropped 12 positions from being ranked 18th overall in 2007 to 30th in 1H08. The company endured a 1H08/1H07 sales decline of 47% (which would have been worse if not for the strength of the euro) and suffered a loss of more than $2.2 billion in its first three quarters of fiscal 2008! • Despite a relatively flat 1H08/1H07 sales result, NXP fell four spots to 14th from 10th. • AMD posted a decent 9% increase in 1H08/1H07 sales, yet slid four positions from 11th to 15th. • Eighth-ranked memory supplier Hynix fell only one spot in the ranking although its 1H08/1H07 sales tumbled 23%, the only 1H08/1H07 sales decline of any top 20 semiconductor supplier! • ST fell one position in the ranking and now occupies the 6th spot. It should be noted that the company’s 1H08 and 1H07 figures do not include FMG (flash memory group) sales, which are now part of Numonyx’s sales. As shown, the company registered a solid 13% year-over-year growth rate in 1H08 and has a relatively positive outlook for its second half as well. • TI did not drop in the ranking but the company registered the second-worst growth rate of any top 20 semiconductor supplier in 1H08 (0%). TI has been a significant supplier to the cellular handset market segment for many years. However, in IC Insights’ opinion, it would be wise not to extrapolate the company’s weak 1H08 performance to the overall cell phone marketplace. Moreover, the excellent 1H08 performance of Qualcomm and ST, also major IC suppliers to the cell phone industry, appears to indicate that TI’s poor 1H08 results are more a reflection of a TI-specific “problem” than an overall cell phone market issue. IC Insights, Inc., Scottsdale, AZ. (480) 348-1133. [www.icinsights.com].
Luxury Brands Aim for MultiBillion Dollar Revenues from the Mobile Handset Market
Dior recently introduced its own line of inhouse designed mobile phones. It took three years to develop the clamshell, which will be available in two versions: a $5,000 basic version and a $26,000 luxury version, which features a 2.6” touch screen with an integrated media player and is decorated with 640 stones of 3,251 carats of Swarovski crystals imprinted on a crocodile skin case.
differentiate their intangible assets: premium value and brand power. Since a handset comprises a mass of technological attributes, its market value is basically determined by all of the components and labor, and its technological performance. This is not likely to change. But consumers are also very aware of brand image and premium value. This gives manufacturers additional leverage in securing customer loyalty, and increasingly forms part of vendors’ long-term strategies for growth. ABI Research, Oyster Bay, NY. (516) 624-2500. [www.abiresearch.com].
Sensors Take the Growth Baton in the Global MEMS Market Other prestige brands are also adding cellular handsets to their product lines. As personal items used frequently in public, mobile handsets can take on a powerful symbolism, representing the user’s social status and personality. In partnership with handset manufacturers and vendors, luxury brands are starting to carve out a distinct niche within the cellular handset industry. ABI Research forecasts that revenues from luxury branded handsets will exceed $11 billion next year, increasing to more than $43 billion in 2013. Research director Kevin Burden comments, “For luxury goods producers, mobile phones are a logical addition to their basic product portfolios of jewelry, watches, and other fashion accessories. From the perspective of handset manufacturers, a luxury mobile phone does not simply mean a new handset model, it represents a meaningful strategic approach to increased brand equity.” Given the heated price competition in handset markets, vendors have struggled to deal with low margins. Cost-cutting alone is to some extent seen as of questionable effectiveness as a way around this impasse. Given the competitive environment in the handset market, manufacturers are looking for ways to
Driven by new demand from consumer electronics and wireless applications, the global market for Microelectromechanical Systems (MEMS) will expand to $8.8 billion in 2012, up from $6.1 billion in 2006, iSuppli Corp. predicts. “The markets for mainstay MEMS actuator products—like inkjet heads and Digital Light Processor (DLP) chips from Texas Instruments Inc.—finally have passed the baton to MEMS sensors to drive the next growth wave in the market,” said Jérémie Bouchaud, director and principal analyst, MEMS, for iSuppli. “The
new wave is partly founded in the rapid rise of consumer electronics applications such as motion sensors for gaming, laptops and digital still cameras. Mobile handsets will also be a strong area, with MEMS sensor revenue in this area to rise at a 22.9 percent Compound Annual Growth Rate (CAGR) to reach $925 million in 2012.” The figure presents iSuppli’s forecast of global MEMS revenue by device. The worldwide market for accelerometers, gyroscopes, microphones, pressure sensors, Bulk Acoustic Wave (BAW) filters, flow sensors, micro-fluidic chips, microbolometers, thermopiles and oscillators is growing at an 11 percent rate annually, and overtook the actuator market for the first time ever in 2007. Four main segments collectively will account for slightly more than 60 percent of total MEMS market revenue in 2012: consumer electronics, mobile handsets, automotive and industrial process control. Aside from the consumer and wireless applications, market pull is being exerted by the automotive sector, an established area set to receive new impetus as a result of mandates for safety and new emissions standards. Demand also will be driven by a diverse range of applications in industrial processing and control. The one bright spot for actuators is in Radio Frequency (RF) MEMS switches used in mobile handsets and test equipment. This market will grow at an annual rate of more than 100 percent from 2006 to 2012, and will account
September 2008
13
analysts’ pages
for $261 million in actuator revenue in 2012, up from just $6 million in 2007. “The consumer electronics and mobile communications fields are much more dynamic than the previous mainstay markets for MEMS, i.e. inkjet heads, diverse industrial applications and automotive uses,” Bouchaud said. “Existnd ing companies have a great opportunity to ride this wave and new players have a chance to ader exploration ether your goal dress a relatively open market. However, deep speak directly R&D pockets are essential to compete in this ical page, the ght resource. area, and the companies that will succeed will technology, be those that bet on and invest in building dedies and products cated mass-production facilities.” ed Pioneering companies with brand new MEMS products tend to enjoy a lengthy monopoly in the market, such as Texas Instruments with its DLP chip or Knowles with MEMS microphones. Meanwhile, in fast-moving markets like concompanies providing solutions now sumer electronics anda automotive sensors—the exploration into products, technologies and companies. Whether your goal is to research the latest datasheet from company, mp to a company's technical page, the goal of Get Connected is to put you in touchlatter with the right resource. Whichever level of driven by mandates—economies of scale gy, Get Connected will help you connect with the companies and products you arewill searching for. norm, with sensor companies atbe the onnected tempting to address both sectors. iSuppli also expects the pace of mergers and acquisitions to accelerate in 2008 and 2009, concentrating MEMS market share among fewer players.
End of Article
iSuppli Corporation, El Segundo, CA. (310) 524-4000 [www.isuppli.com].
Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
14
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
No “Comfort Zone” for Mobile Device Vendors in the Future
Mobile devices are the biggest-selling consumer electronics products in the world, with more than one billion shipping every year. However, the market is in a state of rapid flux. “Three or four years from now, no mobile device vendor—no matter what their market position today—will be in a ‘comfort zone’,” says ABI Research vice president and research director Stuart Carlaw. Developed and developing markets for these devices are being shaped by divergent forces. Developed markets in particular are typically highly saturated, highly competitive and highly segmented, with strong product innovation. Carlaw adds, “The advent of wider mobile broadband access, the drive to maximize data revenue, the desire to push smartphone operating systems down into mid-tier handsets, and rapid innovations in user interfaces will all make the mobile devices of 2010 radically unlike those of today.” As the technologies evolve and market trends emerge, we will see the user interface advances pioneered by Apple’s iPhone continue, with wider use of accelerometers and the addition of haptic feedback to touchscreens. Mobile Internet Devices (MIDs) will become an important market segment, and vendors will increasingly look to diversify their product and service offerings. “The mobile device market is expanding, not consolidating,” notes Carlaw, “and shows increasing micro-segmentation. Disruptive influences abound.” ABI Research, Oyster Bay, NY. (516) 624-2500. [www.abiresearch.com]
After Sluggish Debut, WUSB Prospects Looking Up
Wireless USB (WUSB) is currently hampered by the high price of the underlying UWB silicon, reports In-Stat. This is expected to limit its appeal until prices fall substantially, the high-tech market research firm says. Targeted at PCs, PC peripherals, consumer electronics (CE) and mobile phones, the first devices with WUSB technology shipped in 2007, including notebook PCs and hub and dongle solutions. “The beginnings of a new wireless ecosystem have launched, and should lead to increased shipments in the years ahead,” says Brian O’Rourke, In-Stat analyst. “Ultimately, UWB and WUSB will succeed because they solve problems that no other technology can; the ability to transmit large amounts of data from one device to another with relative power efficiency.”
Recent research by In-Stat found the following: • Just under 100,000 WUSB devices shipped in 2007, a total that will surpass 190 million in 2012. • Competition will come from other WiMedia-based standards, as well as technologies such as Wi-Fi. • Notebook PCs will lead the adoption of WUSB. In-Stat, Scottsdale, AZ. (480) 483-4440. [www.in-stat.com].
Bluetooth’s Growth Beginning to Reach Maturity
Bluetooth seems to have reached a maturation period with growth rates that, while strong, do not match the growth of the recent past, reports In-Stat. The Bluetooth chip market should benefit as new standards, such as Low-energy and High-speed Bluetooth, are expected to be introduced into devices beginning in late 2009, and in volumes beginning in 2010, the high-tech market research firm says. “These new silicon products will spur additional growth in Bluetooth-enabled product shipments, and High-speed Bluetooth ASPs (average selling prices) will help to keep revenues strong as other Bluetooth ASPs continue to decline,” says Brian O’Rourke, In-Stat analyst. “Current market growth for Bluetooth devices is primarily attributed to sales growth in mobile phones, which is the top market for Bluetooth silicon shipments.” Recent research by In-Stat found the following: • Growth of Bluetooth-enabled device shipments will increase by 26% in 2008. • Silicon manufacturers are continuing to roll out integrated radio chips with Bluetooth, Wi-Fi, GPS and FM. • Sales of mono headsets, another high-volume application for Bluetooth, will continue to grow, particularly as regulations surrounding mobile phone usage while driving continue to be made standard in many states. In-Stat, Scottsdale, AZ. (480) 483-4440. [www.in-stat.com].
September 2008
15
cover feature VoWiFi with 802.11n
Voice-over-Wi-Fi Implementation with Single Stream 802.11n The 802.11n standard provides for increased throughput and greater range in VoWiFi devices. This article looks in detail at the benefits as well as the implementation issues for mobile devices. by N arasimhan Venkatesh, Chief Wireless Architect; Peddi Indukuri, Product Manager; and Subba Reddy, Engineering Manager, Redpine Signals
T
The growth of wireless networks based on the IEEE 802.11 Wireless LAN family of standards has been one of the most outstanding success stories of the technology industry in recent years. Apart from the standards themselves, the universal pervasion of WLANs has been assisted and accelerated by the availability of interoperability testing and certification by the Wi-Fi Alliance— so much so that the term “Wi-Fi” is widely used interchangeably with “WLAN.” The initial growth of WLAN was in its intended role of providing a wireless data networking capability as a replacement to a wired LAN connection. However, as its capabilities grew—with the standards being enhanced to offer higher data rates, better quality of service and special modes such as power-save—it quickly became an integral part of a large variety of electronic devices including phones, gaming devices, music players, sensors and other consumer devices. Among these, one of
16
PORTABLE DESIGN
the fastest growing applications has been the transport of voice over the wireless network— thanks in part to the popularity of several commercial Voice-over-IP (VoIP) services. The Wi-Fi Alliance, taking cognizance of the significant potential of Voice-over-Wi-Fi (VoWiFi), has released a certification program called Voice-Personal that helps ensure that the underlying requirements of VoWiFi in Wi-Fi devices are met, in a home or small office environment. Expanding their focus of certification beyond protocol adherence and interoperability, the WiFi CERTIFIED Voice-Personal program focuses on a specific application and is based on performance testing. The program was released in July 2008 and Redpine Signals’ Lite-Fi is among the first products to support the certification. In this article, we provide a background to VoWiFi performance by examining the factors that enable it to provide a satisfying user experience. We elaborate on some of these require-
Requirements of VoWiFi
Voice has traditionally been carried over fixed latency, connection-oriented, low error rate transport medium. These attributes have to be specially provided for in networks based on WLAN. From a user’s point of view, the experience of a voice call over WLAN would be similar to that over an alternative designed-forvoice network if certain requirements are met. These include the following: Latency: A two-way, interactive communication like voice requires the medium to introduce limited packet latency. The connection should permit a tempo of speech to that in a face-toface conversation. The generally accepted limit on latency in a VoWiFi network is 50 ms. Jitter: This is the variation of the time of arrival of packets. Although handled by a jitter buffer at the receiving end, the network is nevertheless required to curb this. Packet drops: The protocols employed to carry voice packets do not provide for a mechanism to re-transmit packets lost in transmission. Indeed, even if they did, it would not be effective since a retransmitted packet would almost certainly have latency requirements violated. The WLAN protocol, of course, does provide for packet retries at the MAC level and this mechanism helps take care of the occasional packet errors that occur even in good channel conditions. VoWiFi devices are expected to limit packet loss to a few percent in good to average channel conditions. Power savings: A wireless device would be actually communicating only part of the time— wireless phone, for instance, would be “in use” typically only an hour or two a day. However, the devices would be expected to be “on” all the time—ready to receive a call if one was to come in, and ready to respond to the user to make a call instantly. The original 802.11 specifications do provide for a client device to be in a “sleep” mode and wake up occasionally to check for pending packets, but this power-save specification is not helpful for voice calls where packets arrive every 20 to 30 ms. The 802.11e standard and the Wi-Fi certification of WMM-PS provide for viable power-save states while maintaining
cover feature
ments and describe how they are implemented in VoWiFi devices.
other requirements specific to voice calls. This article describes this in detail. Range of operation: Voice requires a bandwidth that is only a small fraction of the data rate possible in a Wi-Fi network, but it requires it reli-
figure 1 DTIM Interval DTIM
TIM
TIM
DTIM
Buffered frame for other station
Buffered Frame Time
Power-save station activity PS-POLL
Legacy Power Save
ably in a location-independent manner. The wireless client implementation at the physical layer is therefore required to handle the multipath and interference scenarios that occur in corners and other remote locations of offices and homes. Roaming: A VoWiFi user in a large office could easily wander beyond the range of his Access Point while attending a call. Although enterprises are equipped with sufficient access points so as to provide Wi-Fi connectivity throughout the premises, the switch-over from one AP to the other— roaming—must be done quickly enough to keep the voice connection unbroken to the user. The mechanisms involved in roaming are described in greater detail in the sections that follow. The QoS mechanisms provided in the standards help meet these performance requirements. The WMM certification from the Wi-Fi Alliance verifies for the right implementation of these mechanisms and ensures interoperability with other certified client and access point products. The 802.11n standard provides for SEPTEMBER 2008
17
cover feature
increased throughputs as well as greater range. Range improvement through techniques such as STBC and beamforming, however, do benefit VoWiFi devices. More importantly, the uniform use of 802.11n even in handheld clients is of great benefit to an enterprise that has installed 11n equipment since it ensures that the throughput advantages of 802.11n are preserved. In the following sections, we elaborate on the important considerations of low-power operation and enterprise-wide roaming in VoWiFi client devices.
Power Save
VoWiFi phones may be stand-alone phones, or may be integrated with other telephony devices like mobile handsets and cordless handsets. In all cases, the devices are battery operated and sensitive to power consumption. There are two functional modes in these devices from the power consumption point of view. The first is during the state when a phone is kept “on” and is waiting for a call. The power consumed in this mode will determine the standby time of the phone. The second mode is during an active call—this would determine the talk time of the phone. In this section we will go through the various power saving techniques the 802.11 standard offers and their applicability to VoWiFi.
Legacy Power Save Mode
The 802.11 standard allows stations to go into power save and wake-up periodically to listen to
figure 2 Ack
MPDU
MPDU with EOSP flag set
AP Activity Trigger Frame
Station Activity Service Period
STA Power Profile
the access point’s beacons. The AP buffers the packets of a station if the latter is in power-save mode and indicates the availability of pending packets in the beacon frame. The station, when it wakes up, receives and processes the beacon to check for pending packets and goes back to sleep, that is, power save mode, if there are no pending packets buffered for it. If there are pending packets as indicated in the beacon, the station would continue to stay in the active state and would send a PS-POLL frame to retrieve each of the buffered packets. It goes into sleep state when the “more data” flag is cleared in the received data or management frames. The access point also buffers the broadcast packets for the stations in power save state and delivers them after the DTIM beacon. Stations, therefore, have to wakeup for DTIM beacons to receive broadcast packets. Figure 1 shows the wakeup profiles of two stations in power save. VoWiFi stations during standby can use this mode. The station, however, will not be able to stay in this mode once the call starts as packets could suffer a latency of, say, 100 ms, for a beacon interval of 100 ms, which is not acceptable for VoWiFi.
APSD
APSD, or Automatic Power Save Delivery, was made available in 802.11e. It defines U-APSD, or un-scheduled APSD, and S-APSD, or scheduled APSD. APSD provides a more efficient way of retrieving buffered packets from an AP. With S-APSD, all buffered packets are delivered to a station at a pre-intimated time with respect to the TSF. The station would have to wake up before this and must be ready to receive these packets. In U-APSD, the station would have to send a trigger frame and the AP would deliver the corresponding delivery-enabled frames. It can use any pending data packets as trigger frames or it can send a null frame if there are no pending packets. The station would then continue receiving the packets until it receives a frame with EOSP flag set. Figure 2 shows the frame exchanges in a U-APSD service period. The Wi-Fi WMM-PS certification includes U-APSD operation. VoWiFi stations can use APSD mode during a call. The wakeup period can be configured based on the codec rate. State-of-theart WLAN modules, including Redpine’s Lite-Fi, consume less than 20mW during a VoWiFi call through the use of these power-save techniques.
PSMP Frame exchanges during a U-APSD activity
18
PORTABLE DESIGN
U-APSD being contention-based is only suitable when there are only a few VoWiFi clients
A-MPDU (TA1)
STA2
STA1 UTT
cover feature RIFS or SIFS
PSMP MorePSMP=0
MTBA (RA2)
A-MPDU (RA1)
Last PSMP sequence
etc...
MTBA (TA1)
STA1
SIFS
Data (TID2)
PSMP MorePSMP=1
Data (TID1) Data (TID2)
AMPDU (RA1)
SIFS
A-MPDU (RA2) Data (TID1) Data (TID2) MTBA (TA1) alUStime or SIFS
AP
SIFS
MTBA
SIFS
Initial PSMP sequence Downlink Phase Uplink Phase
MPBA (TA2) (TID1, TID2)
Roaming or handoff is a key requirement of VoWiFi and influences the design of networks supporting VoWiFi in a large way. A mobile STA needs to roam from one AP to another when it moves out of its BSS range, or when the current operating channel conditions get deteriorated. While switching from one AP to the other, at the worst case, an application session may be terminated—requiring it to be re-started after the handover is completed; or in other cases may either experience a temporary outage or at the best case a delay in transfer of packets for a short period of time. Voice applications would not tolerate an outage or session termination, and even a delay in delivery of queued up packets must be minimized—ideally to about 50 ms or less if the mobile user is not to notice a brief degradation in connection quality. In order to maintain application continuity without compromising the key aspects of the connection such as security and power save mode, a VoWiFi-capable mobile station would require using an intelligent roaming mechanism. In order to understand the challenges of roaming, it is important to understand the process involved in 802.11 roaming. A typical scenario is illustrated in Figure 4. As shown in Figure 4, a mobile user starts
figure 3
A-MPDU (TA2)
Roaming
from point A and connects to AP-1 as that is the only AP within his range. He then moves towards point C, via point B. The signal quality of AP-1 at point B is poor relative to that of AP2. At this point, the Wi-Fi-enabled mobile device should detect AP-2, and seamlessly switch from AP-1 to AP-2 without causing disruption to the user’s voice call. This switching involves following tasks, which are to be performed by the mobile STA: • Discover available Access Points in the vicinity • Disconnect from its current associated AP • Establish a connection with the new AP Generally, in normal traffic conditions, the
PSMP MorePSMP=1
in the network. When the number of VoWiFi clients starts increasing, packet collisions rise and adversely affect the QoS requirements of VoWiFi. Also, a station may have to wake up and wait for its turn to be serviced by the AP, consuming power during the period. Power Save Multi Poll, or PSMP, made available in the draft 8-2.11n standard, addresses these issues. In this mechanism, the AP sends a PSMP frame in which it communicates the uplink and downlink time slots for each of the stations. Figure 3 shows the frame exchanges in PSMP. MTBA, or Multi-TID Block Ack, is also a technique introduced in 802.11n to be used in conjunction with PSMP. With MTBA, a single block-ack frame can include block-acks of different TIDs. A station would have to wake up for PSMP and, with the knowledge of its slot times, would be able to go back to sleep until its uplink or downlink slot time arrives. Usually, though, there would be a fixed cost to going to sleep and coming out of it, and it may not be worthwhile going into a sleep mode for a short duration. The primary advantage of PSMP is seen, therefore, only when there are several VoWiFi clients active at the same time.
STA2 UTT
PSMP Frame Exchanges
802.11 disconnecting process can be completed in less than 1 ms. But discovering APs in the vicinity and establishing a connection with the new AP in, say, an enterprise security mode, can take longer—even more than one second. This latency is, of course, not tolerated by voice applications, which are highly sensitive to packet delays. In practice, clients minimize the AP-discovery and connection establishment delay through a combination of the standardsbased and proprietary techniques that generally involve completing most of the tasks that are part of the roaming process in advance. The IEEE standard defines two methods for discovering Access Points, one through passive scan and the other through active scan. It does not, however, address how a mobile STA can discover APs once an active connection exists between the STA and an AP. As shown in Figure 5, clients use proprietary background scan methods to discover the available networks on SEPTEMBER 2008
19
cover feature
different channels, while maintaining connectivity with the current AP. Step 1: The client indicates to the current AP that it is going into power save mode by sending a null frame with PS bit set. Step 2: It moves to a different channel. Step 3: It sends a broadcast probe request that would be read by all APs within range in that channel. Step 4: The client collects the BSS list based on one or more probe responses or beacons received from APs in that channel. Step 5: The client reverts back to its previous operating channel. Step 6: It informs the AP that it is coming out of PS mode by sending a null frame with PS bit off, and resumes data transfer.
figure 4
nd Router
Internet
er exploration ether your goal speak directly ical page, the ght resource. technology, es and products
AP-2 B
C
BSS-2
ed A AP-1
companies providing solutions now
exploration into products, technologies and companies. Whether your goal isBSS-1 to research the latest datasheet from a company, mp to a company's technical page, the goal of Get Connected is to put you in touch with the right resource. Whichever level of AP-1 signal level at Point-B: Poor gy, Get Connected will help you connect with the companies and products you are searching for.
AP-2 signal level at Point-B: Good
onnected
Roaming scenario while moving from point A to C
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
20
During a VoIP call, the client would have only about 20 ms between packets to carry out this background scan. It may, therefore, have to repeat this in case all desired channels are not scanned during one opportunity. Having built a list of APs, the client is ready to roam when it’s time to. The decision to roam is the responsibility of the client. Designers of client devices use their own proprietary mecha-
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
nisms in making this decision. Factors such as Received Signal Strength Indicator (RSSI), signal-to-noise ration, frequency of packet retries are among those that the client devices analyze in deciding to roam. Sometimes a fresh background scan may take place upon one of these measures reaching a critical value. The next important task during roaming is connection establishment with the new AP. Depending on the security settings of the AP, this may involve the 802.1x/Extensible Authentication Protocol (EAP) mechanism. The 802.1x/ EAP is a time-consuming process based on the EAP type that is employed. The simplest 802.1x authentication, which uses Lightweight EAP (LEAP) method, can take a minimum of 100 ms to a maximum of 1.2s, depending on network conditions. The VoWiFi capable mobile STA should minimize this time to less than around 50 ms for meeting the constraints of the voice application. In the open system mode, there is no need for 802.1x/EAP authentication. As a result, in this mode, the client can typically complete the roaming process in less than 20 ms, through its active-join mechanism. In WPA/WPA2 Pre-Shared Key (PSK) security mode, the reassociation process requires an additional 4-way handshake, but not 802.1x/ EAP. This handshake takes additional time, but typically would be done within the constraints of VoWiFi. The advanced WPA/WPA2 enterprise security mode involves the most time-consuming 802.1x/EAP authentication together with the four-way handshake mechanism, during reassociation. As mentioned earlier, this process may take more than 1s, which is beyond the targeted delay for VoWiFi. In order to meet this requirement, the IEEE standard proposes a pre-authentication technique, wherein the client completes the 802.1x/EAP process with the new AP through the distribution system before it decides to roam to this AP. During handoff, therefore, the client would only need to use four-way handshake to obtain the PTK from the new AP. As a result, the client would only take as much time to roam as in the PSK case. Future roaming methods would use provisions in the forthcoming 802.11k and 11r standards. 11k defines quantifiable measurements
cover feature on current BSS and neighboring BSSs, so that a mobile STA can make an informed decision to roam from one BSS to the other; and 11r helps a mobile STA by providing a standardized approach to fast BSS transition to minimize roaming time. And there are also proprietary mechanisms being followed today to hasten the roaming process in enterprise security modes.
ing all product and customer deliverables for Redpine’s WLAN related products. Mr. Indukuri holds an M.B.A from Lancaster University Management School, U.K, a top-30 global B-school, and B.Tech in Computer Science &
figure 5
Certification
Apart from Voice-Personal, there are several certification programs from the Wi-Fi Alliance that are relevant for VoWiFi devices. Certified products have a distinct edge in their ability to provide a uniform user experience in diverse scenarios. The quality of service requirements of VoWiFi are provided by WMM, which covers traffic classes and the priorities afforded to each. WMM-PS helps conserve battery life while a voice or multimedia application is in progress. WPS—Wi-Fi Protected Setup facilitates easy setup of security using a Personal Identification Number (PIN) or a button located on the Wi-Fi device. Voice-Personal, as mentioned earlier, helps ensure the delivery of good quality voice over a Wi-Fi network in a home environment, while the forthcoming VoiceEnterprise certification would do the same in an enterprise environment, adding the requirements of roaming and enterprise security to the test suite to bring about a complete VoWiFi experience in offices and public locations. Redpine Signals, Inc., San Jose, CA. (408) 748-3385. [www.redpinesignals.com].
Author Bios
Narasimhan Venkatesh is chief wireless architect at Redpine Signals and has over 20 years experience in communications engineering with expertise in wireless systems design, telecommunications and optical networking. Mr. Venkatesh’s responsibilities include leading the development of wireless algorithms and hardware at Redpine’s development center in Hyderabad. Mr. Venkatesh holds a Masters Degree in Electrical Engineering from the Indian Institute of Technology, Madras, India. Peddi Indukuri is a product manager at Redpine Signals. He is responsible for manag-
Step 1 Step 2
Step 3
Step 4 Step 5
Step 6
Current AP
Lite-Fi
New AP(s)
Background scan mechanism
Engineering from S.V. University, India. Subba Reddy is an engineering manager at Redpine Signals. He leads the engineering team involved in the design and development of ultra low power WLAN related products. Mr. Reddy is an M.Tech. in Electronics Design and Technology from Indian Institute of Science, Bangalore, India.
SEPTEMBER 2008
21
wireless communications near field communications
Near Field Communications With a range measured in centimeters, NFC is the ultimate in short-range wireless for portable devices, with a wide range of uses. This article explains how it works.
by Steve Rackley
N
Near field communication (NFC) is an ultra short range wireless communication technology that uses magnetic field induction to enable connectivity between devices when they are in physical contact or within a range of a few centimeters. NFC has emerged as a technology for interconnecting consumer electronic devices from the convergence of contactless identification (e.g. RFID) and networking technologies, and aims at simple peer-to-peer networking through automatic connection and configuration. The key difference between NFC and standard RF wireless communication is the way in which the RF signal is propagated between the transmitter and receiver. Standard RF communications, such as Wi-Fi, is described as “far-field� since the communication range is large compared to the size of the antenna. Near field communication relies on direct magnetic or electrostatic coupling between components
22
PORTABLE DESIGN
within the communicating devices rather than free space propagation of radio waves. Because of the very short range, NFC devices can communicate using extremely low electric or magnetic field strengths, well below regulatory noise emission thresholds, so that there are no limitations on frequency band usage due to licensing restrictions. NFC technology is a joint development of Philips and Sony, and is based on the ECMA 340 standard. The technology is being promoted by the NFC Forum, whose sponsor members also include MasterCard, Motorola, Nokia and Visa International. The ECMA 340 standard was adopted by the ECMA General Assembly in December 2004, and defines NFC communication modes using inductive coupled devices operating at a center frequency of 13.56 MHz. The definition is also known as the near field communication interface and protocol (NFCIP-1). Similar
The Newest
Embedded Technologies New Products from:
�����������������
RN-41 Bluetooth® Module www.mouser.com/ rovingnetworks/a
MatchPort™ b/g Embedded Device Server www.mouser.com/lantronix/a
BISM II Bluetooth® Modules www.mouser.com/ezurio/a
The ONLY New Catalog Every 90 Days
Experience Mouser’s time-to-market advantage with no minimums and same-day shipping of the newest products from more than 366 leading suppliers.
ConnectCore™ 9C ARM9-based Module www.mouser.com/digi/a
The Newest Products For Your Newest Designs
www.mouser.com Over A Million Products Online
(800) 346-6873
to the more familiar IEEE standards, ECMA 340 specifies the modulation and data coding schemes, data rates and frame format for NFC device interfaces. A simple link layer protocol addresses link initialization and collision avoidance, and a transport protocol covers protocol activation, data exchange and deactivation.
NFC PHY Layer
ECMA 340 specifies a magnetic induction interface operating at 13.56 MHz and with data rates of 106, 212 and 424 Kbits/s, which is compatible with Philips’ MIFARE and Sony’s
FeliCa contactless smart card interfaces. Rather than measuring transmitter power and receiver detection levels in dBm as is the case for far-field RF communication, the strength (H) of the magnetic field used in NFC is measured in amps/meter (A/m). ECMA 340 specifies the field values as shown in Table 1. The ECMA 340 standard defines two communication modes—active and passive. In the active mode, communication is started by an RF field generated by the initiating device (the Initiator), and the target device (the Target) also generates a modulated RF field to respond to
figure 1 Data bits (Responder 1)
1
0
1
0
0
table 1 1
1
0
0
1
Field level
Field strength
Description
Hthreshold
0.1875 A/m
Minimum field detection level
Combined response Manchester coded
Hmin
1.5 A/m rms
Minimum unmodulated field strength
Detected bits (Initiator)
Hmax
7.5 A/rn rms
Maximum unmodulated field strength
Random retransmit delay Data bits (Responder 2)
1
0
0
1
0
1
0
1
0
Random retransmit delay
1
0
1/0
First two bits
1 Initiator transmits
0
Next two bits
1/0 Resend
NFC Collision Detection with Multiple Responding Targets
table 2 Bit rate
Modulation method
Bit coding method
106 kbps
ASK (100% modulation)
Pulse position coding (Modified Miller) — pulse transmitted at the center of a hit period for each 1-bit, or at the start of a hit period for an opening 0-bit or a repeated 0-hit.
212/424 kbps
ASK (8-30% modulation)
Manchester coding — transition at the center of each hit period; low to high for a 0-bit, high to low for a 1-bit. Reverse polarity (i.e. high to low for a 0-bit, low to high for a 1-bit) is also allowed.
ECMA 340 Active Mode Modulation and Bit Coding Methods
24
PORTABLE DESIGN
ECMA 340 NFC Magnetic Field Strength Specification
the Initiator’s command. Modulation and bit coding methods used in active mode are shown in Table 2. In the passive mode (Table 3), the Initiator starts the communication using an RF field but the Target responds by load modulation rather than by generating an RF field in response. Load modulation entails modulating the load in the target device that the initiating RF field is applied to. This generates sidebands on the original carrier frequency (13.56 MHz) that are detected by the Initiator.
Protocol Stack
Since NFC is not attempting to provide the full range of network features captured in the OSI model, the protocol stack is very limited and consists of a single simple transport pro-
curring around a single data transfer. Initiator and Target negotiate a communication speed, starting with the lowest (106 Kbits/s), in a parameter selection step during transport protocol initiation.
table 3 Bit rate
Modulation method Load modulation
fc/16 = 847.5 kHz
212/424 kbps
Load modulation
-
Carrier modulated using Manchester coding. Reverse polarity allowed.
ECMA 340 Passive Mode Modulation and Bit Coding Methods
figure 2
an elegant process to
1000
resolve collisions that
volves an elegant process to resolve collisions that will occur when several targets respond at the same time, particularly when targets are responding in passive mode (Figure 1). Collision detection at the bit level is made possible by the use of Manchester coding, since a collision is detected when a full bit period occurs without a transition being sensed. This can only occur when a 1-bit transmitted by one target collides with a 0-bit transmitted by another target. Bits received before the collision can be recovered and the targets are requested to resend data starting with the unrecovered bit. A random delay used by responding targets ensures that this process does not get stuck in a repeating loop. The data link between devices is transaction based, with initiation and termination oc-
WUSB (Optional) PHY layer data rate (Mbps)
same time.
Bit coding method
106 kbps
device IDs involves
targets respond at the
Subcarrier frequency
Subcarrier modulated using Manchester coding. Reverse polarity not allowed.
The discovery of target
will occur when several
wireless communications
tocol, which defines activation, data exchange and deactivation on an NFC link. The vestiges of a Data Link layer are also evident in the form of media access control based on CSMA/CA. An Initiator checks for an existing RF field before commencing communication, and similarly a Target device in active mode checks for an existing RF field before responding. A single initiating device can interact with multiple targets, each of which generates a random 40-bit ID at the start of the device selection process. The discovery of target device IDs in-
WUSB (Mandatory)
100
IrDA VFIR
10
1
Bluetooth Class 1
NFC
0.1 0.1
Bluetooth Class 2
IrDA SIR
Bluetooth Class 3
Zigbee 1
10 Range (meters)
100
1000
PAN Technologies; Range vs. Data Rate
NFC in Practice
Four basic NFC usage models are currently envisaged, as shown in Table 4. Apart from these usage models in which the NFC connection is used to transfer end-user data, NFC can also be used to securely initiate another connection between two NFC-enabled devices. For example, NFC-enabled Bluetooth SEPTEMBER 2008
25
nd
wireless communications
or Wi-Fi devices may use NFC to initiate and configure the longer range link. Security is assured by the close proximity requirement for NFC operation. Once the Bluetooth or Wi-Fi link is configured, the devices can be separated for longer range communication.
table 4 Usage model
Description
Example
Touch and go
The user brings the device storing a ticket or access code close to the reader for applications such as event or transport ticketing and access control, or for simple data capture, such as picking up an Internet URL for further information from a smart label on a poster or other advertising.
You see a poster advertising an event such as a concert you want to attend. Bring your PDA or mobile phone near the poster to download event information from a smart chip in the poster.
Touch and confirm
Transactions such as mobile payment where the user is required to enter a password or other confirmation to authorize the interaction.
Event tickets could be purchased online or from an electronic box office and stored on your handheld device.
Touch and connect
Two NFC-enabled devices can be linked to enable peer- to-peer data transfers, such as exchanging photos or synchronizing contact information.
If you take pictures with your mobile phone's built-in camera, you can later touch an NFC-enabled computer or TV to display the images, or touch and transfer them to a friend's mobile phone.
NFC-enabled devices may offer a range of possible functions, including other highspeed connectivity options. The simple NFC connection will allow the user to explore a device's capabilities and access other available services or functionality.
By simply touching two devices together it will be possible to transfer large ides between the devices, for example. using NFC to identify and configure a separate high-speed wireless connection.
er exploration ether your goal speak directly ical page, the ght resource. technology, Touch and es and products
explore
ed
NFC Usage Models
companies providing solutions now
exploration into products, technologies and companies. Whether your goal is to research the latest datasheet from a company, mp to a company's technical page, the goal of Get Connected is to put you in touchCurrent with the rightand resource. Whichever Developments level of Future gy, Get Connected will help you connect with the companies and products you areof searching for. NFC
onnected
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
26
The first examples of NFC in use have been trials followed by commercial deployment for transport ticketing and payment on a local bus network in Hanau, Germany and on Taipei’s Mass Transit Rail system in Taiwan. These trials have been based on the Nokia NFC shell, which clips onto a Nokia 3220 mobile phone. Future data rates up to 1.7 Mbits/s are currently planned, approaching the 3 Mbits/s of Bluetooth 2.0, and market research points to 50% of mobile phones being NFC enabled by 2010.
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
Summary
The simple PAN landscape, dominated by IrDA and Bluetooth, is becoming increasingly diverse, as shown in Figure 2, with many new technologies being developed that will offer the user a wide range of choices in terms of data throughput, range, power consumption and battery life. Printed with permission from Newnes, a division of Elsevier. Copyright 2007. “Wireless Networking Technology, From Principles to Successful Implementation” by Steve Rackley. For more information about this title and other similar books, please visit www.elsevierdirect.com.
portable power auto-zero op amps
Auto-Zero Op Amps: Inherent Benefits in Portable SignalConditioning Applications This self-correcting architecture provides excellent power-supply rejection and common-mode rejection, plus low bias current and noise to portable designs. by Kevin Tretter, Microchip
A
At first glance, the term “auto-zero” operational amplifier (op amp) may appear to be something new, but in reality this architectural concept has been around for decades. This article will explore the history behind auto-zero op amps and provide a high-level overview of the architecture. Additionally, the article will explore the inherent benefits of this architecture for signalconditioning applications. Finally, an example application will be analyzed to further compare the auto-zero architecture to that of traditional op amps.
A Brief History
Chopper amplifiers have been around for decades, dating back close to 60 years. The chopper amplifier was invented to address the need for an ultra-low-offset, low-drift op amp—something that was superior to the bipolar op amps available at the time. In the original chopper amplifier, the amplifier’s input and 28
PORTABLE DESIGN
output are switched (or chopped), causing the input signal to be modulated, corrected for offset error and then unmodulated at the output. This technique allowed for low offset voltage and low drift, but also had limitations. Since the input to the amplifier is being sampled, the input-signal frequency had to be limited to less than half of the chopping frequency in order to prevent aliasing. In addition to the bandwidth limitation, the act of chopping causes significant glitches to appear, requiring filtering on the output to smooth out the resulting ripples. The next generation of self-correcting amplifiers improved on the chopper amplifier by creating a chopper-stabilized op amp. This architecture uses two amplifiers—a “main” amplifier and a “null” amplifier, as shown in Figure 1. The null amplifier corrects its own offset error by shorting the inputs and applying a correction factor to its own null pin, after which it monitors and corrects the offset
Low Drift Over Temperature and Time All amplifiers, regardless of process technology and architecture, have an offset voltage that changes over temperature and time. Most op amps specify this offset drift over tempera-
figure 1 Main +
VIN+
VOUT -
VIN-
Null Input Switches
Advantages of the Auto-Zero Architecture
Null Output Switches
+
As described above, the auto-zero architecture continually self-corrects for the offsetvoltage error of the amplifier. This results in several distinct advantages over traditional op amps. Low Offset Voltage The nulling amplifier continually cancels its own offset voltage, and then applies a correction factor to the main amplifier. The frequency of this correction varies depending upon the actual design, but typically occurs thousands of times per second. For example, the MCP6V01 auto-zero amplifier from Microchip Technology corrects the main amplifier every 100 μs, or 10,000 times each second. This continual correction allows for ultra-low offset voltages that are much lower than traditional op amps. Additionally, the process of correcting the offset voltage also corrects other DC specifications, such as power-supply rejection and commonmode rejection. Therefore, auto-zero amplifiers are able to achieve superior rejection to that of traditional amplifiers.
portable power
of the main amplifier. This architecture has a big advantage over the older chopper amplifiers, as the main amplifier is always connected to the input and output of the IC. Thus, the bandwidth of the main amplifier determines the input-signal bandwidth. Therefore, the input bandwidth is no longer dependent upon the chopping frequency. Charge injection from the switching action is still an issue, which can cause transients and can couple with the input signal, causing intermodulation distortion. The auto-zero architecture is similar in concept to that of a chopper-stabilized amplifier in that there is a nulling amplifier and a main amplifier. However, significant improvements have been made over the years to minimize noise, charge injection and other performance issues associated with chopper-stabilized op amps. Various manufacturers use different terms to define this architecture, such as “auto-zero,” “autocorrelating zeroing” and “zerodrift.” Regardless of the terminology, the basic underlying architecture is the same.
Null
Simplified Chopper-Stabilized Functional Diagram
ture in terms of volts per degree Celsius. This drift can vary substantially from amplifier to amplifier, but for a traditional amplifier is typically on the order of several micro-volts to tens of micro-volts per degree Celsius. This offset drift can be very problematic in high-precision applications; unlike initial offset errors, this drift cannot be accounted for with a one-time system calibration. In addition to drifting over temperature, an amplifier’s offset voltage tends to change over time, as well. For traditional op amps, this drift over time (sometimes called aging) typically isn’t specified in the datasheet, but it can create significant errors over the life of the device. SEPTEMBER 2008
29
portable power
The auto-zero architecture inherently minimizes both the drift over temperature and time by continually self-correcting the offset voltage. In this way, an auto-zero amplifier can achieve significantly better drift performance over traditional op amps. For example, the MCP6V01 op amp mentioned previously has a maximum temperature drift of only 50 nV/°C. Eliminates 1/f Noise 1/f noise, or flicker noise, is a low-frequency phenomenon caused by irregularities in the conduction path and noise due to the bias currents within the transistors. At higher frequen-
figure 2 +5V
1 kΩ
nd
er exploration ether your goal speak directly ical page, the ght resource. technology, es and products
500 kΩ
F.S. = 10 mV
VOUT +
MCP606/MCP6V01
ed 1 kΩ
500 kΩ
companies providing solutions now
exploration into products, technologies and companies.Simplified Whether yourApplication goal is to research the Diagram latest datasheet from a company, Circuit mp to a company's technical page, the goal of Get Connected is to put you in touch with the right resource. Whichever level of gy, Get Connected will help you connect with the companies and products you are searching for.
onnected
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
30
cies, 1/f noise is negligible as the white noise from other sources begins to dominate. This low-frequency noise can be very problematic if the input signal is near DC, such as the outputs from strain gauges, pressure sensors, thermocouples, etc. In an auto-zero based amplifier, the 1/f noise is removed as part of the offset-correction process. This noise source appears at the input and is relatively slow moving, hence it appears to be a part of the amplifier’s offset and gets compensated accordingly.
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
Low Bias Current Bias current is the amount of current flow into the inputs of the amplifier to bias the input transistors. The magnitude of this current can vary from µA down to pA, and is strongly dependent upon the architecture of the amplifier-input circuitry. This parameter becomes extremely important when connecting a highimpedance sensor to the input of an amplifier. As the bias current flows through this high impedance, a voltage drop occurs across the impedance, resulting in a voltage error. For these applications, a low bias current is required. Virtually all auto-zero amplifiers on the market today implement a CMOS input stage, which results in very low bias currents. However, the charge injection from the internal switching can result in slightly higher bias currents than that of a more traditional, CMOSinput op amp. Quiescent Current For battery-powered applications, quiescent current is a critical parameter. Because of the nulling amplifiers and other circuitry required to support the self-correcting auto-zero architecture, auto-zero amplifiers typically consume more quiescent current for a given bandwidth and slew rate, relative to traditional amplifiers. However, significant improvements have been made to increase the efficiency of this architecture. Some op amps, such as Microchip Technology’s MCP6V03, offer a Chip Select or shutdown pin in order to minimize quiescent current when the device is not active.
Application Example: Portable Pocket Weigh Scale
The previous section identified several parameters in which the auto-zero architecture helps to increase amplifier performance. This section will explore an example application using a strain gauge, which highlights some of the advantages of an auto-zero amplifier. Portable weigh scales are popular devices for weighing small items such as precious metals, jewelry and medications. These devices are battery powered and typically require accuracy down to a tenth of a gram, if not better. Therefore, this application requires high-precision, low-power signal con-
portable power
ditioning for the strain gauges used to measure the weight. A strain gauge uses electrical resistance in order to quantify the amount of strain caused by an external force. There are several different types of strain gauges, the most common of which is a metallic strain gauge. This type of strain gauge is composed of a wire or small piece of metal foil. When a force is applied, the strain on the gauge is altered (either positively or negatively), resulting in a change in the strain gauge’s electrical resistance. This change in resistance can then be measured and the magnitude of the applied force quantified. Typically, one or more strain gauges are arranged in a Wheatstone-bridge configuration, due to the excellent sensitivity that this circuit offers. The change in the resistance value is small, so the overall voltage output of such a Wheatstonebridge circuit is small. For this example, we will assume a 10 mV full-scale output. Figure 2 is a simplified circuit that will be analyzed for this application. Please note that this circuitry is not intended to be a complete representation, but is simplified to show the benefits of the auto-zero architecture. For example, the outputs of the Wheatstone-bridge circuit should be buffered to provide a highimpedance input, which is not shown in the circuit diagram. In this circuit, the amplifier is configured for a differential gain of 500, so a full-scale output from the Wheatstone bridge will ideally produce a 5V output from the amplifier. Due to the high amount of gain required in this application, the offset voltage of the amplifier becomes critical. Any voltage offset due to the amplifier will be multiplied by the gain. For example, the MCP606 is a CMOS op amp that implements non-volatile memory to trim the input offset voltage, resulting in a maximum offset of 250 μV (at room temperature). In this application, the maximum offset error of the MCP606 can result in 125 mV of error at the output of the amplifier, or 2.5% of the full-scale range. Let’s compare this to the MCP6V01 auto-zero amplifier, which has a maximum offset of only 2 μV (at room temperature). This offset will result in a maximum error of 1 mV at the output of the amplifier, which is only 0.02% of the full-scale output.
Another advantage of the auto-zero architecture is its low drift over time and temperature. For this example, let’s assume that the portable weigh scale is specified from 0°C to 50°C. The temperature drift of the MCP606 is specified to be 1.8 μV/°C. The error due to drift across this temperature range could be as much as 90 μV, which again would be multiplied by the gain of the circuit, causing an additional 45 mV of error at the output of the amplifier. The MCP6V01, on the other hand, specifies a maximum drift of only 50 nV/°C. Hence, the drift error for this application is only 1.25 mV at the output of the amplifier circuitry, which is over 30 times better than the performance with the MCP606 amplifier. As stated earlier, 1/f noise can be a limiting factor for low-frequency applications, such as the weigh-scale example used here. The MCP606 op amp exhibits the typical 1/f noise spectrum, with a corner frequency around 200 Hz. The 1/f noise begins to dominate at this point, resulting in a voltage-noise density well above 200 nV/√Hz, below 1 Hz. The MCP6V01 op amp, due to its self-correcting auto-zeroing architecture, does not exhibit this 1/f noise, which remains constant at low frequency. For weigh-scale applications, the output of the load cell is a very slow-moving signal, so 1/f noise can be a critical factor.
Conclusion
Today’s auto-zero architecture can date its roots back to the early days of chopper amplifiers, but has improved significantly since that time. The old chopper amplifiers have many shortcomings that made system-level design quite troublesome. The new auto-zero architecture is much more user friendly and provides substantially better performance. As shown in the application example, the autozero architecture can offer much better performance over a traditional op amp in highprecision applications.
About the Author
Kevin Tretter is product marketing manager, Analog & Interface Products Division, at Microchip Technology Inc. He can be reached at kevin.tretter@microchip.com. SEPTEMBER 2008
31
product focus low-power memories Low-Cost LPDDR Non-Volatile Memory Numonyx has introduced its Velocity LP NV-RAM product family—the industry’s fastest low-power, double-data-rate (LPDDR) non-volatile memory—giving mobile phones and other consumer electronics makers better memory performance at lower costs than current solutions. The new devices deliver two to three times faster read bandwidth performance improvement over traditional NOR flash memory, while providing a cost saving alternative to platforms using high DRAM content. The Numonyx Velocity LP NVRAM family also offers a seamless architectural path to the company’s upcoming phase change memory (PCM) products. Digital images, videos, games and other applications continue to put significant memory demands on wireless phones. To meet that demand, most mobile phone makers utilize combinations of non-volatile memory like NOR and NAND with costly RAM memory technologies. Each of these memory types require different hardware and software interfaces and often force designers to make trade-offs between memory system performance and supporting more interfaces, which can increase system cost. Numonyx Velocity LP NV-RAM combines the attributes of the execution memories in the system onto one, widely accepted LPDDR interface—helping to simplify system architecture, increase speeds and reduce costs. Because this non-volatile memory is also executable, designers can enable digital content to be read at RAM speeds and lower the memory system cost by reducing the requirements for LP RAM in the system. Support for the new interface also helps system architects to scale their memory systems to support more phones, from low-end to high-end, with the same architecture. The Numonyx Velocity LP NV-RAM family is compliant with the JEDEC Low Power Double Date Rate (LPDDR) NVM Standard announced in November 2007. Upon ratification, Numonyx will continue to work with the industry to support future generations of interface standards with their products. This is the first step in helping to deliver faster non-volatile memory, reducing the need for added RAM for mobile platforms. Through innovative packaging techniques, Numonyx enhances the benefit of sharing RAM and Velocity LP NV-RAM in a single package-on-package solution. This allows designers to deliver a high-performance, low-cost memory solution on top of the processor. By having RAM and Velocity LP NV-RAM on the same bus, Numonyx can stack the memories on top of each other, saving physical space or “memory footprint” on the circuit board. Numonyx BV, Rolle, Switzerland. +41.21.822.3700. [www.numonyx.com].
World’s First 1.8V 8 Mbit Serial Flash The industry’s first 8 Mbit Serial Flash device capable of operating at 1.8 volts, the AT25DF081 is designed to store both program code and data in battery-powered, portable consumer products such as personal media players (PMPs) and smartphones. The AT25DF081 addresses the key concerns when choosing to use Serial Flash for portable designs. First and foremost, because the AT25DF081 is used as a boot and code shadowing Flash memory, read performance from the Serial Flash is critical in order to keep the applications’ power-up boot time as short as possible to appease the instant-on expectations of consumers. The 66 MHz operation of the AT25DF081 addresses this concern and can enable an application to shadow the entire contents of the memory in under a very short 130 milliseconds. Unlike devices that are screened to operate at lower voltages and subsequently de-rated, the AT25DF081 was designed specifically to operate at 1.8 volts, enabling the device to retain the same high levels of performance as its 3-volt counterpart. By having a complete system operate at a unified 1.8 volts, system designs can be fully optimized to provide the lowest power consumption and longest battery life possible. Precious board space, weight and system cost can also be significantly reduced by simplifying the embedded power-supply circuitry and eliminating the need for an additional LDO (low dropout regulator) and associated passive components that would otherwise be necessary for a 3-volt Serial Flash memory. In addition, since the AT25DF081 can operate at the same voltage as the ASIC or processor/controller, there is no need for level shifters on control and data lines, resulting in further savings. The AT25DF081 is available now for both samples and production volumes. The AT25DF081 is offered in three different package options comprising of an 8-lead, 150-mil wide “narrow-body” SOIC, an 8-contact, 5 mm x 6 mm x 0.6 mm Ultra Thin DFN (UDFN), and an extremely small, 0.4 mm thin 11ball dBGA (WLCSP). In 10,000 unit quantities, prices start at $0.66, $0.70 and $0.83 for the three respective package options. Atmel Corporation, San Jose, CA. (408) 441-0311. [www.atmel.com].
32
PORTABLE DESIGN
STREGNTHEN YOUR [ARM DESIGN EXPERTISE]
REGISTER NOW www.arm.com/developersconference
Drive your business [FORWARD] A full 3-days of classrooms, panels and theater presentations you can’t afford to miss if you’re currently designing or evaluating an ARM Processor-based application. Nowhere else can you learn about the latest ARM processors, its road map and strategies, and how to integrate these into your next design? [COMMUNICATE] with over 2,500 other engineers and industry leaders. [CONNECT] with companies and colleagues developing on the cutting-edge of technology [CREATE] with products specifically designed for your needs.
October 7–9, 2008
Santa Clara Convention Center Santa Clara, CA USA Exhibitors / Sponsors as of September 18th: Platinum
Gold
Silver
prolific
find yourself [HERE] [COMMUNICATE] [CONNECT] Speed your time to market by communicating and exchanging ideas on the latest techniques for designing with the ARM architecture. This three-day event includes: > Over 120 conference sessions providing a full technical design tutorial for embedded applications from hardware and software partners and ARM Licensees. > Wide range of exhibits and product demonstrations from the ARM Connected Community. > Hands-on workshops and theater presentations on the exhibition floor. > Discussions on latest industry trends, strategies and design methodology in forums and panel sessions.
[CREATE] Strengthen your ARM Design Expertise with Special Industry Sessions that will immerse you in current design methodologies and issues Keynote Presentations Tuesday, October 7 Warren East, CEO, ARM The Web Driven Society – Trends, Challenges and Opportunities Kevin Meyer,Vice President, Chartered Semiconductor Interdependence in an Era of Accelerating Technologies Wednesday, October 8 Bill Miller, Director of FIRST Robotics Competition, FIRST Innovation Is Dependent on Inspiring the Youth
exhibitors and sponsors Actel Corporation Alchip Technologies, Inc. American Arium ARM Arrow Electronics Atmel Cadence Design Systems Carbon Design Systems, Inc. CEVA, Inc. ChipEstimate.com CodeSourcery, Inc. Common Platform Coverity Coware, Inc. CriticalBlue DFI Committee Elliptic Semiconductor Evatronix SA EVE USA, Inc. Express Logic Faraday Technology Fraunhofer IIS Green Hills Software GSA Hitex HUONE, Inc. IAR Systems Keil Software Lauterbach
Luminary Micro, Inc. Macraigor Systems Magma Design Automation Marvell Semiconductor Mentor Graphics Corporation Micrium, Inc. Micro Digital, Inc. NXP Open Kernel Labs Parasoft Embedded pls Development Tools Prolific Inc. QSound Labs, Inc. Quadros Systems, Inc. Signum Systems Silistix STMicroelectronics Symbian Synopsys The Dini Group Texas Instruments Text100 Public Relations TRANGO Virtual Processors VaST Systems Technology Vector Software videantis GmbH VirtualLogix VisualOn Inc. Wipro Technologies
ARM presents... ARM Product Roadmaps and Future Technologies
networking reception
Industry Panel Sessions
Tuesday, October 7 5:00pm-7:00pm
Experts will discuss the latest trends, design challenges and future technologies to address power management, mobile computing, gaming and multimedia, security, hardware and software integration and multi-processor technology. ARM Theater
Join the celebration! Everyone is invited! Fifty years ago Nobel Prize laureate Jack St. Clair Kilbey invented the first Integrated Circuit. Join us for a cocktail reception and ARM Powered fashion show in his honor. Brought to you by:
Acquire a broad understanding of how things are built, from choosing the processor, applications and tool chain, and putting it all together. The trademark symbols, logos, service marks and copyright material displayed in this publication (collectively “Marks”) are the property of ARM or third parties. Except as specifically permitted herein no portion of the information in this publication or the Marks contained therein may be reproduced in any form without the prior written permission from the Marks’ owners. Users are not permitted to modify, distribute, publish, transmit or create derivative Marks from any material found in this publication for any public or commercial purpose. The RTC Group is also not responsible for the accuracy or correctness of the content of the material provided to RTC by third parties.
[THE EVENT] location
sponsors
Santa Clara Convention Center
PLATINUM
5001 Great America Parkway Santa Clara, CA 95054 408-748-7000
GOLD
October 7–9, 2008
schedule of events
SILVER
Tuesday, October 7, 2008 9:00am - 4:45pm 12:00pm - 7:00pm 11:00am - 12:00pm 12:00pm - 2:00pm 5:00pm - 7:00pm
Conference Exhibition - Opening Day Keynote Address Lunch Break in the Exhibit Hall Networking Reception
SPECIAL
Wednesday, October 8, 2008 9:00am - 5:45pm 10:00am - 4:00pm 11:00am - 1:00pm 1:00pm - 3:00pm
Conference Exhibition - Closing Day Keynote Address Lunch Break in the Exhibit Hall
Thursday, October 9, 2008
9:00am - 4:45pm Conference 11:45am - 12:45pm Lunch Break in Second Floor Lobby No Exhibits
Aisle Sign Sponsor
Press Room Sponsor
Lunch Sponsor
Lanyard Sponsor
Notebook and Pen Sponsor
Conference Bag Sponsor
MEDIA
Registration Hours
8:00am - 5:00pm Tuesday, October 7 8:00am - 5:00pm Wednesday, October 8 8:00am - 3:00pm Thursday, October 9
packages and pricing As a conference attendee, you will have access to: > All keynote presentations & panel discussions > Technical sessions, tutorials, lectures and case studies > Hands-on-labs > Special events
ARM DevCon Full Conference (3-Day) Pass 1-Day Pass Exhibits & Analyst Presentations
Early Bird By Sept. 23
Regular After Sept. 23
On-Site
$495
$595
$695
$295
$395
$495
Complimentary
Complimentary
Complimentary
Register now: www.arm.com/developersconference
product focus
Low-Power ComboMemory Series Targets EntryLevel Handsets Silicon Storage Technology’s (SST) 34WA Series of 1.8V high-performance flash memory/RAM combination memory products supports functions and features new to SST’s ComboMemory family, including burst-mode operation for faster performance and address/data bus multiplexing for a lower pin count and smaller device footprint. The products in the 34WA Series use a very small semiconductor package, measuring just 6 mm x 8 mm x 1 mm, making them the smallest address/data bus multiplex combination memory devices available on the market today. The performance and space savings provided by the 34WA Series make these products ideal for the entry-level mobile handset market’s need for high-performing, cost-effective, small form factor mobile phones. “Entry-level mobile handsets for emerging regions continue to be significant areas of growth in the mobile phone market,” said Paul Lui, senior vice president, Memory and Special Products Business Unit, SST.
“Mobile handsets with basic features that once served as entry-level products are no longer sufficient for the growing demands of customers in these regions. In light of this, OEMs need to deliver mobile handsets at aggressive price points, yet still provide strong performance, small size and long battery life. The introduction of our new 34WA ComboMemory Series confirms SST’s commitment to supporting these market segments, and we believe the enhanced feature set, low power consumption and small footprint make our 34WA products a compelling choice for handset OEMs.” The 34WA Series supports burst-mode operation, which allows the data in the devices to be read at high speed, thereby increasing overall system performance. The devices also support address and data bus multiplexing to allow the sharing of I/O pins to reduce package pin count and overall device size. The 1.8V operating voltage of the 34WA Series devices leads to lower system power consumption. The 34WA Series is available now. Prices range from $1.30 for the 34WA1601, with 16 Mb flash, to $3.20 for the 34WA32Ax, with 32 Mb flash + 16 Mb PSRAM. Silicon Storage Technology, Inc. (SST) Sunnyvale, CA. (408) 735-9110. [www.sst.com].
36
PORTABLE DESIGN
PCI Express Mini-Card for UMPCs The PCI Express Mini Card format has become the format of choice for the emerging ultra-mobile PC (UMPC) category. With its small footprint and industry-standard design, the mPCIe format is an ideal fit for embedded solid-state storage because it allows for high capacities and high performance while ensuring compatibility in ultra-mobile applications. Leveraging STEC’s proprietary, high-performance MACH4 SSD controller, the Express Mini Card offers extremely high performance (up to 90 Mbytes/s) and high capacities up to 32 Gbytes. The result is a no-compromise solid-state storage solution providing ultramobile system designers with a true plug-n-play storage device, allowing for short design cycles and fast time-to-market. Flash memory has quickly become the product of choice for applications requiring high reliability and high tolerance to shock, vibration, humidity, altitude and temperature, especially important to mobile applications. Because the MACH4 PCI Express Mini Card uses NAND flash technology (e.g. no moving parts), it is more reliable and has much lower latency compared to a traditional hard disk drive, while consuming half the power (approximately 1 watt compared to 2 watts for a typical mobile hard disk drive). STEC’s MACH4 PCI Express Mini Card also adheres to the PCI-SIG Express Mini Card mechanical specification as well as industry compliance and regulatory standards including UL, FCC, RoHS and ATA-7. This combined with a proprietary state-of-the-art flash memory controller supporting key flash management features including advanced wear-leveling and error checking and correction as well as drive monitoring (SMART), provides OEMs with the flexibility to address customer-specific applications with the industry’s highest reliability and endurance. Features include: PCI Express Form Factor; capacities up to 32 Gbyte; built-in power-down data protection; support the use of MLC NAND; full RoHS 6/6 compliance; full data-path protection with built-in 8-bit ECC engine for highest reliability; built-in wear-leveling and block management for ultra-high endurance; commercial and industrial operating temperature ranges. STEC, Inc., Santa Ana, CA. (949) 476-1180. [www.stec-inc.com].
Intel’s Z-P140 PATA Solid-State Drive (SSD), one of the tiniest in the industry, is aimed at handheld mobile devices. Smaller than a penny and weighing less than a drop of water, these 2 Gbyte and 4 Gbyte ultra-small devices are fast, low power and rugged, with the right size, capacity and performance for mobile Internet devices, digital entertainment and embedded products. SSDs use flash memory to store operating systems and computing data, emulating hard drives. The Intel Z-P140 PATA SSD has an industry standard parallel ATA (PATA) interface and is optimized to enhance Intel-based computers, and will be an optional part of Intel’s Menlow platform for mobile Internet devices. The Intel Z-P140 is the smallest SSD in its class, making it attractive to designers and manufacturers of mobile and ultra-mobile devices. Comparatively, the Intel Z-P140 is 400 times smaller in volume than a 1.8-inch hard disk drive (HDD), and at .6 grams is 75 times lighter. It is also a much more durable alternative to HDDs. The 2 Gbyte and 4 Gbyte capacities are large enough to store mobile operating systems, applications and data such as music or photos. It is extendable to 16 Gbytes for added storage capacity. The Intel Z-P140 PATA SSD offers read speeds of 40 Mbytes/s and write speeds of 30 Mbytes/s. Critical to mobile applications, its active power usage is 300 mW, and only 1.1 mW in sleep mode, which helps to extend a device’s battery life. With a 2.5 million hours mean time between failures (MTBF) rate, this PATA-based chip scale package delivers reliable solid-state performance in an extremely tiny footprint. The Intel Z-P140 is available now. The 4 Gbyte version will follow the 2 Gbyte product. Intel Corporation, Santa Clara, CA. (408) 765-8080. [www.intel.com].
1.8V Flash Memory Family Supports ADP, Burst Mode, Simultaneous Read/Write Spansion’s MirrorBit WS flash memory family offers a line of products with high density, high reliability and performance-enhancing features making it the ideal solution for multimedia rich, wireless handsets. The product line features 1.8-volt, 16-bank, fast access with burst mode, and simultaneous read/write operation with product density scaling from 128 Mbits to 2 Gbytes with both conventional NOR and Spansion’s MirrorBit Eclipse architectures. The WS family includes WS-P 90nm NOR; WS-R 65nm NOR; and WS-S 65nm Eclipse. The WS product family supports burst speeds up to 133 MHz as well as page-mode interface, which can improve read transfer rates by up to 50 percent, compared to standard asynchronous flash products. The WSxxxN family consists of 256 and 128 Mbit, 1.8-volt-only, simultaneous Read/Write, Burst Mode flash memory devices, organized as 16 or 8 Mwords of 16 bits. These devices use a single VCC of 1.70-1.95V to read, program and erase the memory array. A 9.0-volt VHH on ACC may be used for faster program performance if desired. These devices can be programmed in standard EPROM programmers. The Spansion S29WS256/128 are MirrorBit flash products fabricated on 110nm process technology. These burst-mode flash devices are capable of performing simultaneous read and write operations with zero latency on two separate banks using separate data and address pins. These products can operate up to 80 MHz and use a single VCC of 1.7V to 1.95V, making them ideal for today’s demanding wireless applications requiring higher density, better performance and lowered power consumption. All MirrorBit WS memory solutions combined with pSRAM are offered in a single footprint for the easiest migration with no PCB changes. Similarly, all WS memory solutions with DRAM are offered in either a standard MCP or a standard PoP footprint. These industry-standard reference footprints are optimized for and qualified on most chipset platforms, and offer customer-centric solutions in a small form factor. Spansion’s Universal Footprint and advanced packaging technology enable system designers to select the right form factor and memory configuration for their unique platforms, thereby empowering phone manufacturers to quickly bring to market handsets that attract new users and grow market share. Spansion Inc., Sunnyvale, CA. (408) 962-2500. [www.spansion.com].
SEPTEMBER 2008
37
product focus
PATA Solid-State Drive Is Ultra-Small, Fast and Low Power
products for designers Miniaturized Jupiter 32 xLP GPS Modules Navman Wireless OEM Solutions has announced that its latest low-power, small form factor Jupiter 32 xLP GPS module is now available in production quantities. Featuring the new 65 nm SiRFstarIII GSC3e(f)/LPx chipset and providing up to 30% greater power savings than previous designs, the new Jupiter 32 xLP ultra-highsensitivity module is the ideal GPS solution for devices where size matters. Delivering high accuracy and extremely fast position fixes, the Jupiter 32 xLP device integrates SiRFstarIII GSW 3.2.4 firmware code and unique Navman software features including Write-to-Flash for easy configuration storage and retrieval; Ephemeris Push for rapid satellite acquisitions and starts; and User Selectable Profiles to maximize operation in even the smallest products. Firmware features incorporate rapid time-to-first-fix in urban environments, high navigation sensitivity and stability, and quick transient condition response and jamming mitigation. Compatible with its predecessor, the popular Jupiter 32 design, the Jupiter 32 xLP is an extremely small GPS receiver sporting a tiny 17.0 mm x 15.0 mm x 2.7 mm form factor and weighing a scant 2.0 grams. Key features of the Jupiter 32 xLP include fully sealed EMI edge shielding, an ultra-high-sensitivity GPS receiver with an optimized RF front-end, proprietary shaped filtering to reduce noise, multi-path mitigation for track re-centering and jammer elimination, and dynamic search control to adapt to severe environments. The Jupiter 32 xLP can be further optimized with user-selectable navigation modes via a simple ASCII command that can be stored in flash memory. These Navigation modes allow the selection of General Use, Automotive, Pedestrian and optimized Low Power modes to suit any application. Navman Wireless, Glenview IL. (847) 832-2367. [www.navmanwireless.com].
High Rate Lithium Ion Cell Panasonic has introduced a new high-rate lithium ion cell, the CGR26650A. The cell’s Nickel Manganese-based cathode material delivers higher discharge rates and better temperature stability than comparative cells, while Panasonic’s exclusive heat resistance layer (HRL) technology provides safer operation. CGR26650A cells are suitable for multi-cell applications. Specific features include: • High discharge rate capability—up to 40 Amps • 2650 mAH typical capacity • High recharge rate capacity—most capacity recovered within 20 minutes • Wide operating temperature range • Excellent cycle life The new cells are designed for applications from power tools to medical equipment, memory storage and back-up power to full scale UPS systems. For more information on this high-rate Lithium Ion cell, go to http:// www.panasonic.com/industrial/battery/oem/chem/lithion/index.html. Panasonic Battery Division, Rolling Meadows, IL. (877) 726-2228. [www.panasonic.com\batteries].
Intel Rolls out New SSDs, New Flash Technology At IDF this week Intel outlined product plans and timelines for its High-Performance SATA Solid-State Drive (SSD) Product Line, aimed at mobile and desktop clients and enterprise server, storage and workstation applications. Intel is dead serious about pushing its flash technology as a replacement for hard-disk drives (HDDs) in everything from data center mainframes to mobile Internet devices (MIDs). At the last IDF, Paul Otellini made it clear that flash would be a centerpiece of Intel’s strategy going forward. This week’s IDF only confirmed the point. According to Randy Wilhelm, VP & GM of Intel’s NAND Products Group, over the last 12 years CPU performance has increased 175x, while HDD performance is only 1.3x faster. Intel’s new SATA SSDs display a “write amplification”—the lag between the measured write from the host to the completed write to NAND of <1.1. Wilhelm doesn’t predict that flash will replace HDDs entirely—indeed, in high-end computers he sees HDDs using flash for cache—but in portable devices he can see flash eventually replacing HDDs. The newly introduced device that fits the portable (non-laptop) space is the Z-P140 PATA Solid-State Drive, which comes in an ultra-small package-on-package BGA solution designed for small mobile devices. Suitable to boot, load and store applications, the Z-P140 is scalable storage in 2 Gbyte, 4 Gbyte, 8 Gbyte and 16 Gbyte capacities, with a standard PATA interface. The recently introduced Z-P240 is a 54 x 4 x 38 mm multi-level cell (MLC) device offered as a “netbook platform solution.” Available in 4-, 6- and soon 16-Gbit MLC, the devices feature access times of up to 35/7 Mbytes/s R/W, 600 G/2ms shock and 2 MHr MTBF. Intel’s new flash products incorporate some innovative technology. Users will get a performance boost from the new Intel Turbo Memory with User Pinning. User pinning allows the choice of which applications or files are stored in non-volatile memory (NVM), giving users direct control of which applications to accelerate. Data-intensive programs will see the most benefit. Intel has also addressed two of the main drawbacks of NAND flash: slow write times and wear leveling, which are directly related. According to Al Fazio, director of Memory Technology Development at Intel, write times in NAND flash are notoriously slow because you can only write at the block level, not to individual cells. A single write request from the OS can result in numerous block writes in NAND; this in turn leads to oxide degradation, shortening the life of the memory. Intel claims to have found a way to write to individual cells without the need for a block copy/erase, thereby extending the life and reliability of the device. Intel Corporation, Santa Clara, CA. (408) 765-8080. [www.intel.com].
Permanent-Mount Antenna Offers Compact Size and Low Cost Antenna Factor has announced a 2.45 GHz version of its popular LP Series antenna family. The series is ideally suited for applications requiring a compact, low-cost, cosmetically attractive antenna solution. These 1/4-wave antennas mount directly to a product’s PCB via a single screw, thereby eliminating the cost of a connector and ensuring FCC compliance. The antennas are available in standard and custom frequencies which, with this latest introduction, now span the 315 MHz to 2.45 GHz range. LP Series antennas exhibit a 50 ohm characteristic impedance and a VSWR of less than 1.9. The antenna costs less than $1.10 in production quantities. Custom colors and logo options are available for volume OEMs. Antenna Factor, Merlin, OR. (800) 489-1634. [www.antennafactor.com].
38
PORTABLE DESIGN
Targeting the handheld market, Texas Instruments has announced the industry’s smallest and thinnest 500 mA, step-down DC/DC converter solution for space-constrained applications. The highefficiency power management integrated circuit (IC) is the first 6 MHz, 500 mA converter to achieve a 13 mm2 solution size with an ultra-thin 0.6 mm total height. The new TPS62601 converter achieves up to 89-percent power efficiency and only 30 μA typical operating quiescent current—all from a 0.9 mm x 1.3 mm chip scale package roughly the size of a flake of pepper. The synchronous, switch-mode device’s fixed frequency of 6 MHz allows the use of only one 0.47 μH inductor with a height of 0.6 mm and two low-cost ceramic capacitors, without compromising performance and efficiency. The TPS62601 can deliver DC voltage regulation accuracy of +/- 1.5 percent. In addition, the device’s excellent load transient response, wide input voltage range of 2.3V to 5.5V and 1.8V of output allows it to effectively support single-rail voltage requirements as designers add new features and functionality. The TPS62601 supports many applications, such as memory modules, GPS modules, Bluetooth and Wi-Fi modules or other wireless micro-modules used in ultra-thin smart phones, digital still cameras, portable disk drives and media players. The converter also applies energy-saving techniques to help maximize battery run-time. For example, the converter automatically enters a power save mode during light-load operating conditions via an automatic pulse frequency modulation and pulse width modulation switching feature. In shutdown mode, the device’s current consumption is reduced to less than 1 μA. The TPS62601 is available today in volume quantities. The device comes in a highly reliable, six-pin, wafer chip scale (0.9 mm x 1.3 mm) package and has a suggested resale price of $1.45 each in quantities of 1,000 units. The TPS62601EVM-327 evaluation module, application notes and TI’s online Power Management selection tool are available through www.power.ti.com.
USB 3.0 Protocol Analyzer LeCroy Corporation has announced the launch of its Voyager verification system, the world’s first protocol analyzer exerciser system for testing USB (Universal Serial Bus) 3.0 devices, systems and software. The Voyager, LeCroy’s sixth-generation verification platform, will help developers bring USB 3.0 products to market faster, while ensuring compatibility with an enormous installed base of USB 2.0 products. Based on circuitry developed for LeCroy’s industry-leading PCI Express platform, the Voyager protocol analyzer is an all-encompassing test solution for USB development. With data rates up to 10 times faster than USB 2.0 and a growing pallet of emerging applications, USB 3.0 is expected to be eagerly adopted by developers. The technology will target fast sync-and-go transfer applications in the PC, consumer and mobile segments that are necessary as digital media become ubiquitous and file sizes increase. For example, a typical 25 Gbyte high-definition digital movie will be able to be downloaded in as little as 70 seconds, versus hours for previous technologies. SuperSpeed USB will be backward-compatible and have the same ease-of-use and plug and play capabilities of previous USB technologies. In addition, the USB 3.0 specification will be optimized for low power and improved protocol efficiency. LeCroy’s Voyager protocol analyzer provides simultaneous protocol capture of both USB 2.0 and USB 3.0 signaling. The Voyager protocol analyzer uses the de facto standard CATC Trace display to illustrate USB 3.0 protocol and includes many enhancements to accelerate testing of USB 3.0 links. Advanced triggering, hardware filtering and Spec View are designed to help developers quickly understand and verify early USB 3.0 protocol behavior. LeCroy Corporation, Chestnut Ridge, NY. (845) 425-2000. [www.lecroy.com].
Texas Instruments Inc., Dallas, TX. (800) 336-5236. [www.ti.com].
Op Amps Eliminate EMI-Induced Errors in Analog Systems National Semiconductor Corp. has introduced three new operational amplifiers (op amps) with integrated electromagnetic interference (EMI) filters that maintain the accuracy of analog systems by reducing the effects of radio frequency (RF) interference. The LMV83x op amps deliver the industry’s highest EMI rejection ratio (EMIRR) of 120 dB, thereby eliminating EMI-induced errors. They also feature 3 MHz unity gain bandwidth while operating on only 240 μA of supply current. This yields a power-to-performance ratio of 80 μA per MHz, which places these devices among National’s PowerWise family of energy-efficient products. The LMV831 single, LMV832 dual and LMV834 quad EMI-hardened op amps reduce development time and board size by minimizing the need for metal shielding, filters and extra components. They are well-suited for use in phone accessories, medical instruments, precision weigh scales and other industrial electronic equipment that are sensitive to electromagnetic disturbances in noisy environments. For example, an injected RF signal in a weigh scale can result in as much as 1V of output offset, which would diminish a 10-bit analog-to-digital converter’s (ADC’s) resolution (1024 codes) to the equivalent of only 3 bits. Using an LMV83x op amp would only reduce resolution by 0.2 bits. A wide range of applications can benefit from EMI protection, including filters/buffers, photodiode preamp and piezoelectric sensors. The LMV831 single, LMV832 dual and LMV834 quad op amps feature CMOS inputs, a maximum input offset voltage of 1 mV, a rail-to-rail output stage and an input common-mode voltage range that includes ground. These devices provide a common-mode rejection ratio (CMRR) and a power supply rejection ratio (PSRR) of 93 dB. The LMV83x devices maintain stability for capacitive loads as large as 200 pF and offer a wide temperature range of -40° to 125°C. The LMV831 is offered in a 5-pin SC70 package, while the LMV832 is offered in an 8-pin MSOP package and the LMV834 is offered in a 14-pin TSSOP package. The LMV831 is priced at 55 cents each, the LMV832 is 79 cents each and the LMV834 is priced at $1.10 each in 1,000-unit quantities. Samples of all three EMI-hardened op amps are available now and production quantities will be available in third quarter 2008. National Semiconductor Corporation, Santa Clara, CA. (408) 721-5000. [www.national.com].
SEPTEMBER 2008
39
products for designers
Industry’s Thinnest 500 mA Power Converter Solution
products for designers
DC/DC Converter MOSFETs Toshiba America Electronic Components, Inc. (TAEC) has expanded it synchronous DC/ DC converter lineup with five new power MOSFETs targeted for use in mobile and desktop computers, servers, game consoles and other electronic devices to convert input voltage to the level required for various subsystems, such as the processor, memory and input-output devices. The TPCM8004 features drain-source voltage (VDSS) of 30V (max.), drain current (ID) of 24A (max.), RDS(ON) of 11 milliohm1 (mΩ, max.), low gate resistance of 1 ohm (Ω, typ.) and a compact, low profile TSSOP Advance package measuring 3.5 mm x 4.65 mm x 0.75 mm. The Toshiba TPCA8030-H and TPCA8031-H MOSFETs are targeted for use as high-side DC-DC converter MOSFETs with VDSS of 30V (max.), ID of 24A (max.), and RDS(ON) of only 11.0 mΩ1 (max.). The TPCA8030-H has typical gate resistance of 1Ω, while the TPCA8031-H has typical gate resistance rated at 3.4Ω. Both devices are available in low-profile SOP Advance packaging, which measures 5 mm x 6 mm x 0.95 mm. The Toshiba TPC8037-H and TPC8038-H MOSFETs are high-side MOSFETs with VDSS of 30V (max.), ID of 12A (max.), and RDS(ON) of only 11.4 mΩ1 (max.). The TPC8037-H has typical gate resistance of 1Ω, while the TPC8038-H has typical gate resistance rated at 3.4Ω. Both devices are available in SOP 8 packaging, which measures 5 mm x 6 mm x 1.6 mm. Samples of the five new Toshiba UMOS V-H DC/DC converter MOSFETs are available now, with prices in sample quantities starting at $0.52. Toshiba America Electronic Components, Inc., San Jose, CA. (408) 526-2400. [www.toshiba.com/taec].
Micropower 50mA Linear Regulator with Programmable Delay Linear Technology Corporation has announced the LT3011, a highvoltage micropower, low dropout regulator that delivers up to 50 mA of continuous output current with a low dropout voltage of only 300 mV at full load. The LT3011 features an input voltage range of 3V to 80V, delivering output voltages as low as 1.24V and up to 60V. The device’s PowerGood flag indicates output regulation. However, a single capacitor may be used to program the delay between this regulated output level and the flag indication. The 80V input voltage capability makes it ideal for automotive applications, 48V telecom backup supplies and industrial control applications. Low quiescent current of 46 μA (operating) and 1 μA (in shutdown) make it an excellent choice for batterypowered “keep alive” systems that require optimum run-time. Output noise is minimized at only 100 μVRMS over a 10 Hz to 100 kHz bandwidth, making the LT3011 ideal for noise-sensitive applications. For high-voltage applications that require large input-to-output voltage differentials, the LT3011 provides a very compact solution. Its thermally enhanced MSOP and DFN packages offer thermal resistance equivalent to much larger conventional packages. The LT3011E and LT3011I are available from stock in a proprietary 12-lead thermally enhanced MSOP package and a 10-lead 3 mm x 3 mm DFN package, both with operating junction temperature of -40° to +125°C. The LT3011H is available in the MSOP package only and is rated up to +150°C operating junction temperature. Pricing starts at $1.75, $1.93 and $2.26 each for the E, I and H grades, respectively, in 1,000-piece quantities. Linear Technology Corporation, Milpitas, CA. (408) 432-1900. [www.linear.com].
40
PORTABLE DESIGN
Constraint-Driven High-Density-Interconnect Design Flow for PCB With BGAs rapidly approaching 2000 pins and 0.8 mm pin pitch, the era of complex high-density interconnect (HDI) stackups in PCBs is upon us. Laying out a PCB of that complexity—especially if you include 2.4 GHz RF components—is beyond the capacity of most PCB design software. Cadence Design Systems has just announced a sweeping set of improvements to the Cadence Allegro and OrCAD families of products aimed at boosting performance and productivity through new features and functionality. Part of the Cadence SPB 16.2 release, the new technology helps deliver shorter, more predictable design cycles for PCB designs. With significant improvements for designers using high-density interconnect (HDI), the technology will be of particular value to customers in the high-end consumer electronics market, as well as those in segments such as computing and networking, where users are seeking a constraintdriven HDI design flow. According to Hemant Shah, product marketing director for Allegro PCB Products, Cadence introduced a constraint-drive design flow in 2000, but used it initially for backend functions. Cadence then migrated it to the front end to enable engineers to insert their design intent into the tool flow. Shah says the new technology combines manufacturing-aware data with a wide range of user-defined HDI design rules, automating the PCB design process while still providing real-time feedback on the constraints. Shah tells Portable Design that Allegro PCB SI will now provide an IBIS 5.0-compliant algorithmic modeling interface for fast and accurate simulation of high-speed serial links. New technology introduced in Allegro PCB for HDI designs includes new objects, an extensive set of new rules for micro-vias, an enhanced via-transition use model, and changes to the entire PCB design flow to enable a comprehensive constraint-driven HDI design flow. Design partitioning has been enhanced with new capabilities for partitioning the design horizontally and adding soft boundaries to allow users to work in parallel more efficiently, further shortening the design cycle. Customers can shorten their time-to-market and reduce development costs for high-frequency signals such as those found in PCI Express 2.0, Serial ATA II and SAS II. Using Allegro PCB SI users can quickly and accurately simulate and validate for BER compliance using new and advanced eye mask capabilities, high-frequency field solver technology. In addition, Allegro PCB SI provides simulation support for interoperable, multi-vendor IBIS 5.0 AMI-compliant transceivers. With the layout-driven RF PCB design capability introduced in the new release, users can eliminate the need to manually update schematics for RF circuit elements added into the layout. Combined with an improved bidirectional integration with Agilent’s ADS environment, the Allegro PCB RF option allows users to shorten time to create mixed-signal digital-analogRF designs. Finally, engineers can specify and embed physical and spacing constraints for critical high-speed nets in the design to improve chances of first-time success while eliminating traditional error-prone verbal, email and spreadsheet-based communication. This can help shorten design cycles and eliminate unnecessary iterations between hardware designers and PCB layout designers. SPB 16.2 will be available in Q4 2008. Cadence Design Systems Inc, San Jose, CA. (408) 943-1234. [www.cadence.com].
Atmel Corporation has announced the immediate availability of its AT91CAP7X-STK Starter Kit for the evaluation of its ARM7-based CAP family of customizable microcontrollers. CAP7 customizable MCUs allow designers to migrate ARM7-plus-FPGA designs to a low-NRE, single-chip solution with approximately 30% lower unit costs, with eight times better performance, 98% less static power consumption and 70% less active power consumption. The $399 CAP7 Starter Kit includes a printed circuit board with Atmel’s ARM7-based AT91CAP7S microcontroller, Altera CycloneII EP2C8F256C7N FPGA with EPCS4SI8N serial configuration memory, 2.8” TFT LCD panel, joystick, 64 Mbytes of SDRAM application memory, 10-bit analog to digital converter (ADC), 256 Mbytes of NAND flash and a 4 Mbyte DataFlash. External interfaces include USB full-speed device, four analog inputs, external bus interface (EBI), USART, SPI and Debug UART. The board also supports sensing applications with light and temperature sensors and potentiometers. The onboard Cyclone II FPGA contains 8,256 four-input Lookup Table (LUTs) logic elements equivalent to 66,048 CAP7 MP block gates—about one seventh the total available logic on a CAP7 device. It can be used to map and emulate application-specific IP blocks, in the CAP7’s embedded Metal Programmable (MP) block, along with the system software. The AT91CAP7S executes at clock speeds up to 80 MHz and can send to or receive data from the FPGA at this speed over the EBI. The AT91CAP7S can also interface the FPGA via peripheral IO (PIO). The CAP7 starter kit allows designers to determine how the software, MCU and new IP will play together in a customizable MCU, without incurring any NRE charges or spending thousands of dollars for a full development board. Designers can accurately gauge the superior power consumption and performance characteristics of the CAP7 device, evaluate the ADC resolution, and experiment with the peripherals on the device. Smaller designs can be emulated and evaluated on the AT91CAP7X-STK. However, for development of a customized CAP7 device, the fully featured AT91CAP7-DK is recommended to allow customers to take full advantage of the CAP7’s multi-layer high-speed bus (AHB) and peripheral DMA. There are 32 general-purpose I/O connections on the AT91CAP7S, and 75 I/Os on the FPGA to support application-specific external interfaces. Debug is facilitated by the provision of an ICE-JTAG interface for CAP7 JTAG programming, and a USB-Blaster-JTAG interface for Cyclone II JTAG programming. Atmel Corporation, San Jose, CA. (408) 441-0311. [www.atmel.com].
Integrated 3D EM Simulation Solution for RF Module Design Agilent Technologies Inc. has announced an integrated design flow solution that includes full 3D electromagnetic (EM) simulation for RF Module Design. Called EMDS-for-ADS, the full 3D EM simulator is integrated into Agilent’s Advanced Design System EDA software platform and eliminates the need for stand-alone EM tools. EMDS-for-ADS helps designers accurately predict the 3D EM interactions of embedded passive components in RF modules while co-simulating with active circuits to maximize the wireless subsystem performance. The integrated 3D EM simulator improvements allow designers to analyze larger circuits faster without leaving their familiar design flow. This capability increases productivity in the overall design and verification process. EMDS-for-ADS accounts for the finite dielectric boundaries of RF modules. In addition, it is useful for verifying the accuracy of faster Planar EM simulators such as Agilent’s Momentum, which assumes infinite dielectric planar layers in its analysis. EMDS-for-ADS also features a new finite element mesher and high-capacity iterative solver that delivers better accuracy, speed and capacity for RF SIP (system-in-package) and RF module designs. The most common applications for EMDS-for-ADS are RF modules based on LTCC (low temperature co-fired ceramics) and laminates with embedded passive structures. They are found in almost all RF modules produced today. EMDS-for-ADS saves time when drawing these structures with its planar RF layout macros. These macros automatically draw RF components such as spiral inductors and meander lines—which are timeconsuming to construct using a generic 3D drawing and simulation tool. EMDS-for-ADS is part of the new Agilent Advanced Design System 2008 Update 2. Advanced Design System is a powerful electronic design automation platform, offering complete integration to designers of consumer and commercial wireless electronic products such as mobile phones, wireless networking and GPS, as well as radar and satellite communications systems, and high-speed digital serial links.
Compact Buck-Boost Regulators with Ultra-Fast Switching Analog Devices has announced a compact buck-boost regulator, the first to support switching frequencies at speeds up to 2.5 MHz. ADI’s ADP2503 and ADP2504 step-up/step-down DC/ DC regulators are designed to regulate voltages in portable devices for RF, audio and motor applications. The ADP2503 and ADP2504 operate at input voltages ranging from 2.3V and 5.5V, which meets the requirements for single Li-Ion, Li-Ion polymer cell and multiple alkaline/NiMH cell applications. Fixed output voltages range from 2.8V to 5V. The ultra-fast switching speed of the new regulators allows designers to use low-cost multilayer inductors that are half the size of other solutions. In addition, the total external component count has been reduced to three, resulting in a total PCB area of less than 13 mm2 and a height of less than 1 mm, making the solution ideal for space-constrained applications such as wireless handsets, digital still cameras, portable audio players and USB-powered consumer and industrial devices. The ADP2503 and ADP2504 regulators are based on ADI’s new proprietary current-mode buck-boost architecture, achieving glitchless mode transitions and outstanding transient performance across line and load. This high level of output stability is critical for powering sensitive analog and digital circuitry. The devices also feature one of the industry’s lowest no-load quiescent current (Iq) levels—38 µA in power-save mode—which extends stand-by time in portable electronics and/or increases the power budget for the inclusion of additional features. The proprietary H-Bridge buck-boost architecture improves the efficiency by more than 10 percent versus legacy cascaded boost-buck architectures by reducing switching losses. The ADP2503 and ADP2504 are available now in sample quantities and come housed in a 10-lead 3 mm × 3 mm thin LFCSP package. The buckboost regulators are priced at $1.40 per unit in 1,000-unit quantities. Analog Devices, Norwood, MA. (781) 329-4700. [www.analog.com].
Agilent Technologies, Santa Clara, CA. (877) 424-4536. [www.agilent.com].
SEPTEMBER 2008
41
products for designers
Starter Kit for ARM7-based Customizable Microcontrollers
design idea figure 2
NiMh/NiCd Switchmode Battery Charger Has dV/dt Charge Termination by A lfredo H. Saab and Shasta Thomas, Maxim Integrated Products Inc., Sunnyvale, CA
All battery chargers can be regarded as constant-current power supplies, but they differ from power supplies in two important respects: battery chargers (by design) block all discharge paths from battery to charger, under any conditions. They also include circuitry that decides when the battery has taken a full charge (signaling when the full-charge current must be reduced), and when the charging process should be terminated. Several techniques are available for deciding when a NiCd or NiMh battery is fully charged. The most common of these relies on terminating the charge when the battery terminals reach a particular voltage level, based on a characteristic increase in the positive slope of voltage versus time. This is not the best method, because the absolute value of termination voltage depends strongly on the ambient temperature and the charge rate (the “C” rate). The final result can therefore be an under- or over-charged battery. Overcharging a NiCd or NiMh battery is not as serious as for lithium batteries, which are much more sensitive to damage. NiCd and NiMh batteries are more rugged devices. Undercharging is a problem, simply because the store of charge will be less than expected.
figure 1
40
Cell Voltage (V)
1.55 1.50
V
1.45
∆V ∆t Cutoff
T
35 30 25
1.40 0
Cell Temperature (°C)
NiCd Battery Charging Characteristics at C Rate
30 60 90 Charge Time (Minutes)
These curves show the voltage and temperature characteristics of a NiCd cell as it approaches and passes the fully charged condition.
Most of the better termination methods rely on the fact that charging transforms electrical energy into stored (potential) chemical energy. Charging is an endothermic process. That means the battery temperature not only doesn’t rise; it actually falls slightly during a charge. When a battery reaches full charge, the reactions that transform electrical energy to chemical energy cease. Any further electrical energy forced into the battery by the charger transforms to heat, which increases the battery temperature. At that point charging should stop, because the battery has stored 100% of its capacity. You can sense this temperature increase and use it as a signal for termination of charging, but that measurement implies a thermal sensor in intimate contact with the battery—not always a feasible arrangement. You can also sense the temperature increase by changes in the battery’s terminal voltage, which is a sensitive indicator of internal temperature changes. Thus, 42
PORTABLE DESIGN
This switchmode charger for NiCd and NiMh batteries uses their dV/dt behavior as an indicator for charge termination.
charging a battery produces a positive slope in the plot of voltage versus time. The positive slope turns negative when a NiCd battery reaches full charge, and goes to zero (flat) when a NiMh battery reaches full charge. Figure 1 shows the terminal voltage vs. time for a NiCd battery under charge. The time scale for change of slope, which can range from minutes to tens of minutes according to the battery size, depends on the thermal time constant of the battery and its enclosure. It also depends on the charge rate (i.e., the charging current), because the temperature increase and its rate of increase are functions of the battery’s thermal capacity and of power delivered (which, in turn, is a function of charging current). To detect slope changes, the charge controller must run a detection algorithm that makes sequenced voltage measurements at long time intervals and then stores the results for comparison. This capability, which cannot be implemented in analog form, must be performed by a combination of ADC, memory, timer and sequencer. IC battery chargers such as the MAX712 (for NiMh batteries) and MAX713 (for NiCd batteries) run dV/dt slope-sensing charge-termination algorithms. The power section of these devices is linear and has limited efficiency, but is adequate for smaller-capacity batteries. The circuit of Figure 2 includes a MAX712 or MAX713 battery charger, and also a switching regulator (MAX5089) that handles the power portion of the charger. This regulator operates at higher efficiency and a higher switching frequency (2 MHz), which in turn enables construction of smaller-size chargers capable of higher fast-charge rates, yet with little need for heat-sinking. The MAX5089 step-down regulator is controlled by the MAX712 or MAX713, acting as a battery-charge controller. It produces outputs of 7V to 16V and operates from any DC source capable of delivering the desired fastcharge current, with this condition: the output voltage must equal 1.7V multiplied by the sum of (two plus the number of batteries to be charged in series). The Figure 2 circuit charges a pack of 1 to 8 cells with fast-charge currents as high as 2.5A, and retains all programmable features described in the MAX712/MAX713 datasheets. When charging is complete, the MAX712/MAX713 devices go to a state called “trickle charge,” in which they inject a small fraction of the full charge current (to compensate for the self-discharge always present). You can therefore leave the battery connected to the charger, and find it charged to +100% when needed. Datasheets for the MAX712/MAX713 and MAX5089 are available at www.maxim-ic.com.
product feature Early Dynamic Power Analysis and Pre-RTL Exploration Tool Incisive Palladium Dynamic Power Analysis for Cadence Incisive delivers the promise of CPF across the tool chain. by John Donovan, Editor-in-Chief All designs involve trade-offs between the conflicting demands of performance, power and price. Decisions made at the architectural level have far more impact than those made post-RTL, but to date RTL has been a chasm that’s been hard to cross. If you got down to gate-level simulation only to discover that you’d blown a hole in your power profile, your options were tantamount to rearranging the deck chairs on the Titanic. One in four SoC designs fails because of power-related issues. The major EDA vendors have taken direct aim at system-level power estimation: Cadence with its Common Power Format (CPF) and Mentor, Synopsys, Magma et al with their Unified Power Format (UPF). Cadence was the first to commercialize their approach, integrating it across their Incisive tool suites and Encounter Power System with the Cadence Low-Power Solution. Now you could generate a power intent file at the system level and ensure that the implementation and verification tools would work within the parameters you set earlier. What was still lacking was the ability to do fast, cycle-accurate power estimation and exploration at the system level. Cadence has now rolled out Incisive Palladium Dynamic Power Analysis (DPA), which enables SoC designers, architects and validation engineers to quickly estimate the power consumption of their system during the design phase, analyzing the effects of running various software stacks and other real-world stimuli. Cadence’s acquisition earlier this year of Chip Estimate with its InCyte tool is quickly paying off. The new DPA offering includes Cadence InCyte Chip Estimator, which now provides lowpower planning capabilities, including automatic creation of CPF files. This allows designers to perform accurate pre-RTL estimation of die size, performance and cost, enabling early exploration of the design impact of various low-power techniques. The InCyte Chip Estimator can be used to author and explore CPF scenarios and interfaces into downstream Cadence implementation, RTL simulation and emulation tools that drive low-power strategy through the design methodology. Palladium Dynamic Power Analysis represents a methodology shift for power budgeting of electronic devices with system-level implications. With a focus on productivity improvement, DPA helps to quickly identify the average and peak power consumption of SoC designs running real software in various operational scenarios. Leveraging Palladium III’s built-in memory and RTL Compiler power estimation engine, Cadence provides the first high-performance, cycle-accurate integrated solution delivering full-system power analysis of designs, including both hardware and software.
InCyte Chip Estimator and Palladium Dynamic Power Analysis are available immediately. The Palladium Dynamic Power Analysis product is being sold as an option for the Palladium III system. This means you have to buy Cadence’s expensive simulation/emulation toaster in order to get it. But with Palladium III claiming verification speeds of 100-1,000,000x faster than RTL simulation and 10x the debugging speed of FPGA-based emulators, there’s an argument to be made for that in any case.
Cadence’s Dynamic Power Analysis product finally brings the promise of CPF to their entire tool chain, which goes a long way toward addressing the number one problem facing designers of portable products. It’s fair to assume that the UPF camp will catch up, bringing similar capabilities to users of competing EDA tools. In the meantime it’s “advantage Cadence.” In recognition of their achievement, Portable Design names Cadence Incisive Palladium Dynamic Power Analysis its September 2008 Product of the Month. Cadence Design Systems Inc, San Jose, CA. (408) 943-1234. [www.cadence.com].
SEPTEMBER 2008
43
ceo interview Necip Sayiner Silicon Labs
While Austin-based Freescale and AMD have been taking their lumps of late, Silicon Laboratories (SiLabs) quietly continues to be a major Austin success story. It’s admittedly easier when you’re not taking aim at the big dogs—and thus not winding up in their cross hairs—but SiLabs has continued to innovate its way to healthy growth for a number of years now in highly competitive wireless markets. SiLabs is a fabless semiconductor company focused on high-performance, analog-intensive, mixed-signal ICs. Their biggest revenue nd generators are their VoIP and embedded moer exploration dem products, the latter being widely used in ether your goal high-definition set-top boxes. Their CMOS FM speak directly tuners are designed into mobile handsets by ical page, the virtually every manufacturer, and they’ve reght resource. technology, cently introduced a 3x3 mm AM/FM receiver. es and products In the horizontal space, they’ve shipped over 100 million low-cost, mixed-signal MCUs and ed 100,000 development kits. This mixed vertical/horizontal strategy has resulted in a GAAP gross margin of 63% for the latest quarter and a 38% increase in revenues over the same quarter last year. In March 2007 SiLabs sold its AERO RF companies providing solutions now transceiver line to NXP for $285 million. Usexploration into products, technologies and companies. Whether your goal is to research the latest datasheet from a company, ofresource. the cash from level the of sale, in January mp to a company's technical page, the goal of Get Connected is to put you in touching withsome the right Whichever gy, Get Connected will help you connect with the companies and products you areSiLabs searching bought for. Integration Associates, who onnected make ASSPs for short-range wireless and audio subsystems. Much of the rest of the cash is earmarked for SiLabs’ extensive R&D pipeline. Portable Design sat down recently with SiLabs’ CEO Necip Sayiner to ask about its products, its business model and its plans for the future.
End of Article Get Connected
with companies mentioned in this article. www.portabledesign.com/getconnected
44
Portable Design: Silicon Labs’ strength has long been in analog-intensive, mixedsignal ICs. This is a valuable skill set, but not entirely lacking in your larger competitors. How has Silicon Labs managed to do so well in the face of considerable competition?
PORTABLE DESIGN
Get Connected with companies mentioned in this article.
Sayiner: We definitely have tremendous mixed-signal analog design talent in the company. One of our core competencies is to be able to design what traditionally has been solved in pure analog by using digital-centric approaches, which lend themselves well to cost reduction and achieving high performance with small die size. We have been able to successfully combine RF and mixed-signal with digitalcentric architectures. Another important component in our success has been our value system that encourages innovation among our design engineering community. They are encouraged and recognized by approaching problems in unique ways to come up with interesting solutions. As a result, if you look at our products you will see that most have very unique approaches to these problems. We are able to achieve high performance and enable our customers with better products, at the same time maintaining the low cost that many of our target applications require. Portable Design: Your R&D expenses for the six months ending July 5, 2008 amounted to 23.7% of revenues. This is down considerably for the same period a year before (31.4%), but still considerably higher than for the semiconductor industry as a whole. Why are you spending so much on R&D, and is spending at that level sustainable? Sayiner: On an adjusted basis—excluding the stock compensation expense—our target model calls for about 21% of our revenues to be applied to R&D; this is roughly where we are today. We have been just under 20% the prior quarter and guided about this range for the current quarter. We think that this level of R&D is appropriate for a company at our stage of growth. We still consider ourselves a growth company. We have been able to grow our revenues substantially higher than the industry in a weak macro environment. We feel that we can continue to grow our revenues and outperform the industry, but that requires injecting a significant amount of R&D into the business. We’ll continue to expand our R&D investments commensurate with sales growth and roughly keep it around 20-21%. Portable Design: In what technology areas are you placing your bets, and how do you expect to see them develop over the next few years? Sayiner: We think of our investments in two major categories. One is vertical markets, where
we have a number of established businesses where we have gained leading market share. Examples of that would be embedded modems for set-top box applications; our SLICs for voiceover-IP (VoIP) applications; and our radio products for broadcast audio applications. We’re going to continue to pursue a roadmap to maintain our position in the marketplace. Another application in the vertical category would be broadcast video, where we have decided that our RF core capability to develop our products can serve the video market going forward. We anticipate certain integration paths that will be available to companies like us who can design RF in CMOS. On the horizontal business side we have a relatively low share in a very large market supplied by a number of incumbents. These are markets, such as microcontrollers and timing, where we continue to grow our revenues at substantial rates, but we still see a lot of runway ahead of us given where our share is in these large markets. One area of new growth for us would be in timing, where we have brought to market several very innovative products over the last couple of years, both in clocks and oscillators. There is a lot in the R&D pipeline for us for the next several years. We’ve also recently started investing in shortrange wireless applications, again bringing the core capabilities that we have to markets where we see good synergy with our other existing product lines; this is consistent with our strategy of building a broad-based business. Especially with the recent acquisition of Integration Associates, short-range wireless in particular is going to be a notable growth sector for us. Portable Design: The sale of your Aero transceiver line to NXP last year certainly strengthened your bottom line while helping NXP to score a major design win with Nokia. Following that deletion from your RF portfolio, how have you reshaped your wireless strategy? Sayiner: When we sold the cellular business, we retained a good portion of our skill sets as well as a substantial amount of IP. We felt that wireless in general would be an area where we want to invest in different major trends in that space. We have been using our RF capability in broadcast audio; this has helped us to gain leading market share. We’re now looking to duplicate that in video, utilizing some of the RF resources in broadcast video. More recently we’ve also started initiatives in short-range wireless; that is an application
where we think our RF know-how and the capabilities we have in integrating RF with some of the digital circuits such as MCUs in small footprints are going to enable us to have competitive products that other companies will find difficult to match. So it’s been audio and video on the broadcast side, and short-range applications attacking wireless in multiple forms. Portable Design: In the first six months of this year you made 88% of your sales outside the U.S.—primarily in Asia, where you also do most of your manufacturing. Do you see this as a trend—or current reality—for the semiconductor industry as a whole, and if so, how do you see it playing out? Sayiner: In the past decade there have been a couple of noteworthy developments: One, the OEMs in Asia have outperformed their peers in North America and Europe. Whether we’re talking about the cellular industry or networking, you now see a number of giants in China, Korea and other places in Asia. So that certainly has shifted some of the center of gravity toward Asia from an OEM perspective. Secondly, the ODMs, working for the OEMs, have gained more power and capability over the past decade. What used to be just a manufacturing strategy now has turned into a partnership with those ODMs, where the ODMs have more decision-making power. What started as just choosing from a list of components dictated by the OEM, now they have a lot more freedom in designing the boards and making choices of the components. So as a supplier, we actually look at them as our customer in many cases. This is the current reality, and I think it’s also found in the industry globally.
Portable Design: Do you see IP vendors moving up in any numbers to become fabless semiconductor companies? What about fabless companies falling back on being IP suppliers? Sayiner: I don’t see either of those happening. I think there are very significant elements in developing a product that goes beyond IP in terms of product definition and customer insight. I personally think that it is going to be difficult—with a couple of notable exceptions—for IP providers to continue to grow and survive, because the IC suppliers are developing that IP as it becomes core to their clients; and I think those IP providers will find it difficult to move up the chain and develop the required customer relationships to be essentially IC suppliers. Most of the time they don’t have the infrastructure or the culture to achieve this. Portable Design: Finally, what would you like Silicon Labs to look like 3-5 years from now? Sayiner: In many ways the way it is today. We certainly would like to see the revenues and the business grow significantly over that period of time, and our challenge is to achieve those financial goals while maintaining the set of values within the company that has helped us become successful: our agility, our focus on commercial success, on execution and many of the values that people in the hallways here embrace. We want to be able to keep them intact. Silicon Laboratories Austin, TX. (512) 416-8500. [www.silabs.com]
Portable Design: All but the largest semiconductor manufacturers are going ‘fab light’ at some process node, while others are looking to joint ventures and partnerships to share fab CAPEX. Do you see the fabless model as the ‘wave of the future’? Sayiner: Yes, I think that’s a foregone conclusion now. Many IBMs offered to build partnerships at 65 nm nodes, and at 45 nm those partnerships are falling apart. With a couple of notable exceptions, nobody really has the need to build a 45 nm fab—or has the revenues of the two or three IBMs who are capable of building foundries going forward.
SEPTEMBER 2008
45
The RTC Group is a media services company specializing in bringing companies and their products to a focused group of electronic and computer manufacturers. RTC is proud of its track record of blazing new trails in search of marketing value for our clients. Portable Design magazine is the newest addition to RTC Group’s collection of publications.
event calendar 10/01/08
advertiser index Altera Corporation www.altera.com
27
EDA Tech Forum Boston, MA www.edatechforum.com
ARM Developers’ Conference www.arm.com/developersconference
33,34,35
10/07-09/08
ARM Developers’ Conference Santa Clara, CA www.rtcgroup.com/arm/2008
EmbeddedCommunity.com www.embeddedcommunity.com
4
10/01/08
EDA Tech Forum Denver, CO www.edatechforum.com
Express Logic www.expresslogic.com
47
10/21/08
Real-Time & Embedded Computing Conference San Diego, CA www.rtecc.com/sandiego2008 10/23/08
Real-Time & Embedded Computing Conference
Linx Technologies, Inc www.linxtechnologies.com
Maxim Integrated Products www.maxim-ic.com
4
48
Long Beach, CA www.rtecc.com/longbeach2008 10/26-30/08
Mentor Graphics www.mentor.com
2
Microchip Technology, Inc. www.microchip.com/16bit
9
Embedded Systems Conference Boston, MA www.cmp-egevents.com 11/04/08 – Election Day
Vote & Visit Real-Time & Embedded Computing Conference Reston, VA www.rtecc.com/reston2008 11/13/2008
Real-Time & Embedded Computing Conference Detroit, MI www.rtecc.com/detroit2008 If you wish to have your industry event listed, contact Sally Bixby with The RTC Group at sallyb@rtcgroup.com
46
PORTABLE DESIGN
Mouser Electronic www.mouser.com
23
! W
NE
BenchX
™
Express Logic’s new, complete, cost-effective development tools solution for embedded systems Eclipse-Based IDE for ARM, ColdFire, MIPS, and PowerPC Express Logic’s BenchX™ Integrated Development Environment (IDE) is a full– featured, Eclipse-based development tools solution for embedded systems. The Eclipse community’s efforts have ushered in a new generation of IDEs and tools that can be adapted for use with embedded systems. Express Logic’s BenchX, its own new Eclipse-based IDE, is tailored for embedded development, and supports the ARM, PowerPC, ColdFire, and MIPS processor architectures. Best of all, BenchX is very affordable, and requires no license keys.
LOAD TO GO IN JUST 4 MOUSE CLICKS!
• • • • • • • • •
Eclipse-based IDE for Embedded Development GNU C/C++ Compilers and Libraries GDB Debugger With Graphical Interface Project Builder, Editor, Browser, Simulator Hardware Debug Probe ThreadX Kernel-aware Debug New Project Wizard Easy-to-Learn Tutorials No License Keys
For a free evaluation copy, visit www.rtos.com • 1-888-THREADX Copyright © 2008, Express Logic, Inc. ThreadX is a registered trademark, and BenchX is a trademark of Express Logic, Inc. All other trademarks are the property of their respective owners.
High-efficiency, lowest EMI, 10W Class D amplifier provides I2C or analog volume control ESS SUCC
F RSRO G YEGAIN EE IN EN
Ideal for battery-powered applications 3.3V
4.5V TO 14V
AUDIO INPUTS
FILTERLESS CLASS D SPEAKER OUTPUT
SHDN MUTE ANALOG OR I2C VOLUME CONTROL
MAX9768
EMI WITH FERRITE BEAD FILTERS (VDD = 12V, 0.5m CABLE, 87 LOAD) 40
AMPLITUDE (dB V/m)
35 30 OVER 15dB MARGIN TO EN 55022B LIMIT
25 20
MAX9768
15 10 5 30
60
90 120 150 180 210 240 270 300 FREQUENCY (MHz)
U Mono 10W output into 8Ɵ U I2C or analog 64-step volume control U High efficiency eliminates heat sink U Excellent click-and-pop performance
DIRECT
™
www.maxim-ic.com/shop
www.avnet.com
U Low-EMI modulation passes EMC with inexpensive ferrite-bead filters U Wide 4.5V to 14V operation U Space-saving, 4mm x 4mm x 0.8mm TQFN package
www.maxim-ic.com/MAX9768-info
For free samples or technical support, visit our website or call 1-800-998-8800. The Maxim logo is a registered trademark of Maxim Integrated Products, Inc. © 2008 Maxim Integrated Products, Inc. All rights reserved.