Page 1

The magazine of record for the embedded computing industry

November 2009



New Directions for Mezzanines Enforce Design Rules for Reliability Small Modules in Even Smaller Places An RTC Group Publication


Spans Embedded Applications

43 Robust Dual Gigabit Ethernet Mezzanine Board with Long-Term Availability

44 FPGA-Based Industrial Ethernet Module Integreates a High-Capacity Switch-IP


46 Small Linux Networking Server Offers Low-Cost Edge Computing for OEMs



5Editorial Memory... Memories and the Fall of Civilizations 6

Industry Insider Latest Developments in the Embedded Marketplace

9 & Technology Newest Embedded Technology Used 42Products by Industry Leaders Small Form Factor Forum Happy Holidays from SF3


technology in context

Industry Insight

New Direction for Mezzanines

High Reliability

Generation of M-Modules and Increasing Reliability through PMCs Saves Development Time Automated Enforcement of Design 14 New 30 and Money Rules Barbara Schmitz, MEN Mikro Elektronik

Paul Anderson, Ph.D., GrammaTech

Solutions Engineering


MicroTCA Spans Applications

Advances in Small Form Factors

to Reality: An Application Atom SBC Targets PowerArchitecture with MicroTCA over-Ethernet Apps 18 Concept 34Tiny Tony Romero, Performance Technologies

Bandwidth and Small Form Factor MicroTCA for Flexible 24 High Embedded Design Sven Freudenfeld and David Pursley, Kontron

Eurotech - CPUs in Specialized Devices

Chris Lane, Jaco Electronics and Kelly Gillilan, Adlink Technology

Form Factors Harness Emerging Technologies to Enable 38Embedded Wireless Systems Jason Krueger, VersaLogic

Architectures Bring C Programmability into Formerly 10CPU Specialized Devices Tom Williams

Digital Subscriptions Avaliable at RTC MAGAZINE NOVEMBER 2009



Don’t let


tough specs

PRESIDENT John Reardon,

keep you

Editorial EDITOR-IN-CHIEF Tom Williams, CONTRIBUTING EDITORS Colin McCracken and Paul Rosenfeld MANAGING EDITOR Marina Tringali, COPY EDITOR Rochelle Cohn

from bidding

You know electronics.

We know electronics packaging. Give use your most challenging packaging problem and we can engineer it for you.

Art/Production CREATIVE DIRECTOR Jason Van Dorn, ART DIRECTOR Kirsten Wyatt, GRAPHIC DESIGNER Christopher Saucier, GRAPHIC DESIGNER Maream Milik, DIRECTOR OF WEB DEVELOPMENT Marke Hallowell, WEB DEVELOPER James Wagner,

Our electronics packaging is used in military, security, telecommunications, space research and many other fields.

Advertising/Web Advertising

Whether it’s exceptionally high EMI/RFI attenuation, FCC/VDE, EMP/Tempest, Mil-Spec 810 or 901, seismic resistance or protection against dust, shock and vibration we’ve probably engineered it.

WESTERN REGIONAL ADVERTISING MANAGER Stacy Mannik, (949) 226-2024 WESTERN REGIONAL ADVERTISING MANAGER Lauren Hintze, (949) 226-2014 EASTERN REGIONAL ADVERTISING MANAGER Shandi Ricciotti, (949) 573-7660 INSIDE SALES MANAGER Carrie Bowers, (949) 226-2029

Billing Maggie McAuley, (949) 226-2024

Equipto Electronics’ experienced enclosure engineers can work from your drawings, rough sketches, even conversations.

To Contact RTC magazine: HOME OFFICE The RTC Group, 905 Calle Amanecer, Suite 250, San Clemente, CA 92673 Phone: (949) 226-2000 Fax: (949) 226-2050,

Call us today for a free consultation. Together we can win your bid.

EASTERN SALES OFFICE The RTC Group, 3310 Twin Ridge Drive, Charlotte, NC 28210 Phone: (949) 573-7660

8 :200 nt 9001 ia ISO Compl s RoH

Editorial Office Tom Williams, Editor-in-Chief 245-M Mt. Hermon Rd., PMB#F, Scotts Valley, CA 95066 Phone: (831) 335-1509 Fax: (408) 904-7214

800-204-7225 •630-859-7840 • • Aurora, IL 60506


Untitled-3 1


11/9/09 9:53:14 AM

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.


Tom Williams Editor-in-Chief


Memory... Memories and the Fall of Civilizations

hat is memory? In its most basic form, it is the establishment of synaptic connections in the brain. But as we all know, especially as we get older, that is a temporary and unreliable means of actually preserving information. So over the centuries memory has taken the form of impressions in clay tablets, monumental stelae, hand-calligraphed manuscripts, runes, parchment volumes laboriously copied in monasteries, and ultimately, the printed word. The 20th century saw the arrival of photographs, film, video, phonograph, tape recording, and now a blistering variety of digital technologies. It is, of course, these latter that concern us most today. Without digital technologies there would be utterly no way to accommodate the flood of information that is cascading over the world, and yet sometimes a question arises: Archaeology has been able to unearth and interpret the information left by previous civilizations. I wonder if it will be that easy for future excavators or visiting aliens to uncover and decipher ours. In order to read our technology we need other technology, and how durable is digital memory over long centuries? As durable as stone? We have lost great repositories of information in the past such as the Great Library of Alexandria and through the wholesale destruction of the Mayan manuscripts. These speak to the vulnerability of centrally stored, non-redundant data. Today’s data is nothing if not decentralized and redundant in the form of the Internet. Most of the data, however, is on spinning media in server farms all over the world. The technology that makes it available is a web of wired and wireless communications held together by other data in the form of domain servers. Access to the data is by means of yet another technology known as search engines. As wondrous as all this is, there is still something about it that feels tenuous. At least it feels tenuous over the long term, and by that I mean centuries. The curve of the increase of information will continue to climb necessitating new technologies to contain it. What is the life in hundreds of years of a magnetic disk, of flash memory? If and when our civilization does collapse from war or climate change or an epidemic of gout—whatever—will our learning really be available to future civilizations—ones perhaps worthy of the name?

Some very minor glitches are starting to appear. We have all had, or will one day have, the experience of an older loved one passing away. Part of the experience is going through an attic or basement or closets and finding boxes of old letters, pictures and cherished objects. I remember finding a bundle of letters written by my father from the Pacific during World War II. That killed an entire afternoon. What would have happened if those had been emails? They would not have been accessible over the Internet or by any search engine. And yet such normally personal things are often key to insights into history. Caroline Alexander, author of The Bounty: The True Story of the Mutiny on the Bounty, of course researched the records of the British Admiralty and other official documents, but she also gained access to private letters still kept in homes by the descendants of sailors who had been aboard the Bounty. These gave priceless insight into unraveling the actual story of what happened. Dad also left behind a bunch of documents on 3 ½-inch diskettes. While there was nothing on them that was nearly as interesting as the letters from the Pacific, establishing that fact was a genuine hassle. Nobody had a computer handy with a 3 ½-inch drive. Ten years from now, it will be genuinely a hardship and beyond that might take a forensics lab. Useful data and records need to not only exist in public archives, but also in the attics and old media of ordinary people. With the rapid changes in media and interfaces, much of the access to such information may become increasingly difficult to the point that we may lose valuable insights into our historical and cultural heritage. Just finding things could be harder as well. There is no search engine or indexer for a closet. I’m not sure what the answer is from a technology standpoint. On one hand, vast amounts of data and knowledge have been made available to almost anybody in an instant. Technologies for accessing and organizing that data have advanced as well. And yet, the preservation of all of this does not seem to have been thought out. If this by no means eternal civilization continues, the data will be carried forward on new technology—at least the public data. The private, often more interesting stuff is a different matter. RTC MAGAZINE NOVEMBER 2009



INSIDER NOVEMBER 2009 CP-TA Introduces Self-Testing Program for xTCA Interoperability The Communications Platforms Trade Association (CP-TA) has introduced a self-testing program for the interoperability of xTCA (AdvancedTCA, MicroTCA, AMC) building blocks. Companies can now utilize the interoperability criteria, test procedures and test tools developed by CP-TA and its members to perform self-testing in order to verify xTCA product interoperability. CP-TA empowers the xTCA ecosystem by providing standardized, proven guidance and tools to lower integration costs without the added expense of a formal third-party certification program. In addition to its continued work toward easing xTCA interoperability challenges for traditional telecom central office applications, CP-TA has also expanded its mission to include xTCA communications platforms used in Military, Aerospace and data center applications as well. While originally designed for use in telecom network operating environments, xTCA has gained traction in additional markets due to the advantages of open standards, inherent reliability and power efficiency. Vendors supplying to each of these vertical markets share similar xTCA interoperability objectives. Today’s announcement comes in conjunction with the organization’s recent designation of Polaris Networks’ MicroTCA Tester and AMC Tester as suggested tools for use with CP-TA’s Interoperability Compliance Document (ICD) and Test Procedure Manual (TPM). CP-TA approved Polaris’ ATCA Tester earlier this year as an authorized test tool, and all three tools are harmonized with the ICD and TPM to test for manageability interoperability of xTCA building blocks. Furthermore, CP-TA has negotiated for certain discounts on xTCA test tools from Polaris Networks for member companies depending on the level of membership. Similarly, CP-TA is working with additional tools providers to develop thermal testing tools so that companies can test xTCA products for CP-TA B.x thermal compliance. To correspond with the goal of broadening its focus, CP-TA has restructured its membership levels and now offers a “Participant” level for $2,000. The new level offers members significant technical and marketing benefits at a relatively low cost, and provides an easy entry point for joining and participating in the organization. A list of membership levels and benefits is available on the CP-TA Web site

RTI DDS Selected for Upgrade of Grand Coulee Dam’s Control System

Real-Time Innovations has announced that the U.S. Army Corps of Engineers has selected RTI Data Distribution Service (DDS) for an upgrade of the Grand Coulee Dam’s control system. RTI’s middleware will integrate components of the dam’s nextgeneration Supervisory Control and Data Acquisition (SCADA) system, enabling more efficient and reliable hydroelectric power generation. The Grand Coulee Dam is the largest hydropower plant in the United States. The plant must carefully control how it supplies power to the grid. Failures that cause power spikes or dropouts are not



acceptable. The project to retrofit the control system is for the U.S. Bureau of Reclamation (BOR) in cooperation with the Bonneville Power Administration (BPA). The U.S. Army Corps of Engineers Hydroelectric Design Center (HDC) is the system integrator; HDC selected RTI Data Distribution Service to implement a highly available, high-performance and scalable control system. The dam network connects a 40,000-point SCADA system controlling 30 generators to the transmission switchyard. In the new architecture, RTI DDS will be used to integrate 50 to 60 Linux-based Programmable Logic Controllers (PLCs) on the central control network with Windows Human-Machine Interface (HMI)

systems and various database and health-monitoring servers. The system will run on standard, ruggedized commercial computing hardware. The architecture will eventually be deployed on a network of 12 dams in the Federal Columbia River Power System. Automated Control Systems (ACSI) is a contractor on the project. ACSI President Dan Perrier noted, “The dam control system processes waveform data, power metrics, temperatures, oil levels, vibration sensors and more. The current system uses many different protocols. DDS, with its flexible configuration and high performance, is easier to work with and is a better way to integrate the systems.” RTI DDS provides messaging for mission-critical distributed ap-

plications. It combines low latency, high throughput and fault tolerance in a scalable architecture for real-time systems. Systems built with RTI middleware are loosely coupled; they communicate easily but under strict Quality of Service (QoS) control. Developers can specify the exact needed timing and reliability for each communication path. Direct peer-to-peer connection ensures fast delivery with no single points of failure. Automated configuration allows hot-swap, non-stop redundancy. Subsystems can start in any order, join or leave, or insert upgrades without impacting existing operation. RTI reliably delivers the right data to the right place at the right time.

ISM and SUMIT-ISM Specs Bring SUMIT Expansion to 90 x 96 mm Stackable Modules

The Small Form Factor Special Interest Group (SFF-SIG) has announced the availability of both revision 1.0 of the Industry Standard Module (ISM) and the SUMIT-ISM specifications for small, rugged, stackable embedded systems. The SUMIT-ISM specification documents the use of SFF-SIG’s flexible Stackable Unified Module Interface Technology (SUMIT) interface on popular 90 x 96 mm stackable modules. The ISM specification provides an explicit form-factoronly definition upon which the SUMIT-ISM specification is built. Since the SUMIT specification itself defines only a board-to-board interface (connectors and pin definition), the ISM specification is necessary to define the form factor while the SUMIT¬-ISM specification defines how SUMIT is implemented on ISM modules. In order to support a wide variety of legacy 90 x 96 mm modules marketed as PC/104 or PCI-104 modules on a SUMIT-

ISM stack, the SUMIT-ISM specification offers a high level of flexibility to maintain compatibility. The SUMIT-ISM specification defines two legacy stack types, using slotted mounting holes on ISM modules to provide symmetry not found with PC/104. This enables SUMIT-ISM modules to be created with legacy support for either the PC/104 ISA bus or the PCI-104 PCI bus by allowing the module to be rotated 180 degrees as necessary to fit the legacy type required while maintaining the SUMIT interface. Legacy bus support can be supplied by the CPU and maintained up the stack, or it can be provided through a bridge module in the stack itself. With SUMIT-ISM, stackable I/O expansion is implemented using the SUMIT standard introduced by SFF-SIG in early 2008. Through the inclusion of one or two 52-pin SUMIT connectors, a SUMIT-ISM CPU can provide PCI Express (up to 6 x1 lanes or 2x1 and 1 x4 lane), USB 2.0, LPC, I2C and/or SPI interfaces to the SUMIT-ISM I/O modules. The SUMIT-ISM CPU designer has the flexibility to provide all or any subset of these interfaces. In anticipation of the release of these specifications, SUMITISM CPUs are already available from member companies ADLINK and VersaLogic. SUMIT-ISM I/O modules are available from member companies VersaLogic and WinSystems.

“Tree Power”—Bio-Energy for Forest Service’s Climate Sensor Network

Voltree Power has announced its first contract with the USDA Forest Service, following a rigorous and successful test in June of its Climate Sensor Network. The Climate Sensor Network complements the USDA Forest Service’s Remote Automated Weather Sta-

tions network. Its durable and inexpensive mesh network of lowpower sensory nodes was designed to operate underneath the forest canopy, and collect and report data on wind speed and direction, air humidity, and temperature. The data, which is transmitted back to fire officials, can be used to assist in the prediction of areas in the forest at higher risk for fires at the time the information is collected. The initial contract is for a total of five network systems, the first to be deployed in Boise, Idaho, and tested through the end of this year. In 2008, Voltree Power received a patent for its bio-energy harvesting technology, which collects the energy that is naturally produced by living trees and other large plants, and uses it to trickle charge and run low-power circuits. The USDA Forest Service is evaluating a wireless mesh network application of the technology, with nodes that consist of low-power transceivers, sensors and Voltree Power’s bio-energy harvesting technology. Data collected by the nodes would be transmitted from one node to another until they reach a central monitoring station. Each station provides a satellite microwave uplink connection that allows the collected information to be shared with government agencies and many others worldwide. The USDA Forest Service and the Bureau of Land Management are not the only major entities to recognize the potential of Voltree’s bio-energy harvesting technology. Texas Instruments has been a contributor to the development of the Climate Sensor Network, providing industry-leading, ultra-low-power embedded processing and power management technology that enables innovative, self-powered applications. In addition, work by a group at the University of Washington with no

ties to Voltree has confirmed the viability of operating low-power circuitry from tree power, a methodology covered broadly by Voltree’s issued patent and pending patent applications.

Giesecke & Devrient and Phison in Joint Venture for Secure Flash Solutions

Giesecke & Devrient (G&D) and Phison are joining forces to develop and market secure flash products. Their joint venture— Giesecke & Devrient Secure Flash Solutions GmbH—will be headquartered in Munich, Germany, and will have a second office in Taipei, Taiwan. It will develop and market hardware-based security solutions for mobile end devices, for example in the form of microSD cards and USB tokens. G&D will hold a 70 percent majority stake in the joint venture. Mobile end devices for communications and multimedia applications such as notebooks, cell phones, mobile TV equipment and games consoles have become an integral part of everyday life. These applications make it necessary to protect the digital identity of the user. At the same time, mobile transactions made using the equipment and data stored on it also need to be protected against unauthorized access. Giesecke & Devrient Secure Flash Solutions will offer security solutions that cover these market demands and are designed for special applications such as the encryption of emails or cell phone conversations, or mobile pay-TV. The Mobile Security Card already developed by G&D using flash controller technology from Phison is a good example of such solutions. In the future, the Mobile Security Card will be further developed by Giesecke & Devrient Secure Flash Solutions and

marketed by the joint venture. New generations of the Mobile Security Card, such as for contactless applications, and other secure flash products are already under development. Potential clients include manufacturers and other providers who are interested in integrating secure applications into mobile end devices. These may be PC and mobile equipment manufacturers, system integrators, retailers and/ or providers of access systems for mobile TV applications.

One Stop Systems, Bell Microproducts Sign Distribution Agreement

One Stop Systems has entered into an agreement with Bell Microproducts whereby Bell Micro will sell and distribute One Stop Systems’ portfolio of computer modules and systems. This new partnership enables One Stop customers to easily purchase products by phone, email or online, as well as make simple price and availability inquiries. Bell Micro’s call center is able to quickly respond to customer requests and handle them expediently, and a dedicated account manager oversees the customer/factory relationship. One Stop has long been designing and manufacturing custom computers for the industrial and commercial marketplace, and also offers 500 standard products. These products include NEBS servers, Direct Attached Acceleration Systems, Direct Attached components and expansion systems, CompactPCI/CompactPCIe and passive backplane/SHBe components and systems, disk arrays, and Ethernet and T1 boards. Steve Cooper, president and CEO for One Stop noted, “OSS has been ordering products from Bell Micro for years, and is happy to extend that relationship to selling OSS products and solutions to Bell RTC MAGAZINE NOVEMBER 2009



Micro’s customer base of OEMs, systems builders and resellers. We will look to Bell Micro’s wellestablished, knowledgeable sales force to help us expand our business over the coming years.� According to Gary Gammon, senior vice president of enterprise marketing for Bell Microproducts, “OSS has a large product offering and a talented engineering team, and by partnering with them, Bell Micro is able to further expand the market-leading solutions that we can offer our customers such as PCIe expansion, NEBS servers, RAID units and Mil-Spec servers/storage. This distribution agreement is a win-win situation.

Emerson Network Power and Kontron to Deliver Embedded Computing to Japanese Market

market. Uniting the two companies, Emerson Network Power will be the master channel partner for both Emerson and Kontron embedded computing products in Japan, including all sales, customer support and related activities under the Emerson brand. Customers will be served via an established network of direct and distribution channels, providing both sales and technical support using Emerson Network Power’s service center and partners. Kontron and Emerson will jointly provide all product engineering, application engineering and technical support. Through this collaboration in Japan, customers will have access to the best embedded computing technologies, products and solutions that both Emerson and Kontron offer.

Event Calendar 01/26/10


Real-Time & Embedded Computing Conference

Real-Time & Embedded Computing Conference

Santa Clara, CA

Huntsville, AL




Real-Time & Embedded Computing Conference

Santa Clara, CA

Robins AFB, GA


AFCEA West 2010 San Diego, CA

To list your event, email:

Emerson Network Power, a business of Emerson, has announced it will join forces with Kontron to serve the Japanese







Untitled-4 1


t     t t t t t t t t t t t






t t t t t t




11/11/09 3:47:09 PM



Colin McCracken & Paul Rosenfeld


was the night before shipping, the deadline was near Not a creature was stirring but one engineer His systems were staged in burn-in with care In hopes that a serial console prompt soon would appear Small form factors were nestled all snug in their racks While delusions of interoperability pervaded each stack Ops started the bake and then left for the night While the engineer settled down to consider his plight When out on the floor there arose such a clatter He sprang from his bench to see what was the matter Away to the chamber he flew with his flash Checking for failures, he re-booted the batch The glossy datasheets on the engineer’s bench Hinted at plug-and-play with the modules entrenched When what to his wondering eyes should appear But an LPC super I/O and eight COM-port gear With a Win32 driver, migration was quick He knew in a moment to use a USB stick More rabid than squirrels his vendors once came Selling I/O he had purchased by name: Now, ARINC! Now, PATA! Now, DO-160! On, COM ports! On, 1553! On A-to-D! On top of each stack, screwed to the wall A thermal solution that’s excessively tall As wide temp now dictates a heat transfer plate Both to a COM module and enclosure must mate So up like a smokestack, the modules they grew With a tray full of cables and pin/socket connectors too! He scratched his head and was turning around When the custom BIOS re-flashed with a bound He stressed out while awaiting re-boot His 5-vendor COMe RFQ had become moot A bundle of cables he had flung on his back And he looked like a peddler just opening his pack He thought back on how he got into this mess The “legacy-free” mantra pushed by COM Express The promise of inter-op’rability With the carrier design guide from the PIC-M-G

Happy Holidays from SF3 The words of his boss were doomed to repeat “Your stacking approach is way obsolete The size, weight, and power budget must shrink.” The consultant appeared before he could blink His eyes—how they twinkled! His design skills—how merry! His schematics like roses, his layout a cherry A wink of his eye and a twist of his head The consultant said we had nothing to dread He spoke not a word but went straight to his work And redesigned the baseboard; then turned with a jerk “In COM Ex Type III, IDE is not there This pinout type uses a differential pair…” He dipped into his bag of proven design tricks Then pulled out a heat spreader plate for x86 He spun the design, a right jolly old elf Now smaller than the stack built commercial off-the-shelf As he sprang to his sleigh, the consultant gave a whistle And away he flew like the down of a thistle Leaving behind our poor engineer To debug the pilot ’fore the end of the year Integration had our friend jumping through hoops Of the finest standards from many trade groups And then in a twinkling he heard in his head A voice reminder of an SF3 column he’d read About ISA vampires and legacy I/O The LPC COM ports need BIOS initialization to flow He plugged his flash drive into each system’s port With new firmware, the serial ports didn’t abort The characters were displayed through console redirection And he restarted the burn-in after one last inspection Then our hero exclaimed ere he drove out of sight “Happy troubleshooting to all during many a long night!” RTC MAGAZINE NOVEMBER 2009


ploration your goal k directly age, the source. ology, d products

editor’s report Eurotech - CPUs in Specialized Devices

CPU Architectures Bring C Programmability into Formerly Specialized Devices The quest to combine high functionality, high performance and configurability with ease of programming is moving CPU architectures into more deeply embedded and specialized device environments. by Tom Williams, Editor-in-Chief


iven all the hardware assists that knowledge as did graphics processors. are continually appearing to speed SoCs, while based on familiar CPU cores, specialized parts of applications required precise selection of specific onsuch as I/O, signal conditioning, graphics chip components to make them work as nies providing solutions now processing, signal processing, etc., in the desired. ion into products, and companies. Whether your goal is to research Gradually, the latest formtechnologies of coprocessors, FPGAs, Systemshowever, many of these ation Engineer, or jump to a company's technical page, the goal of Get Connected is to put you on-Chip (SoC) and more, most of us would extremely helpful technologies are being you require for whatever type of technology, really like for. to develop and express our tamed into the C corral. This is happening and productsstill you are searching actual application in one language—most both through the adoption of C constructs usually C—if possible. to the more specialized silicon and by way Traditionally, however, when these of mainstream CPU architectures moving new devices become available, they take into the parts, giving easily programmable specialized knowledge to wring out their capabilities to high-performance devices. promised performance. FPGAs once required Verilog but now have much Kudos to CUDA friendlier development environments, and You may have a massively parallel some even use a “C to Gates” approach general-purpose processor on your PC to program them. DSPs required special and not know it. There is a growing push for “general-purpose computing on graphics processing units” (GPGPU) and it is Get Connected centered on a widely used graphical archiwith companies mentioned in this article. tecture from Nvidia. Since 2006, Nvidia’s

End of Article



Get Connected with companies mentioned in this article.

graphical processors have incorporated an architecture called Compute Unified Device Architecture (CUDA) on their graphic cards. CUDA processors were designed to be highly parallel graphical engines programmable in C. Millions of units have already been sold for exactly that reason, but they turn out to be capable of much more. While the latest generation of generalpurpose, multicore processors is spurring the development of virtualization techniques and hypervisors to bring out their full potential, the CUDA architecture is programmable in C and has a runtime component that can spread sections of parallel algorithms over its complement of available cores. The performance gains can be impressive. GE Fanuc, which has entered a partnership with Nvidia to develop new hardware products using the CUDA architecture, is very focused on rugged devices for aerospace and military applications where size, weight and performance (SWaP) are critical. GE Fanuc’s estimates are that it can keep the same level of performance for a given system and implement it in about 10 percent of its current size and weight. This would translate to 1000 percent performance in the same size and weight. In an autonomous vehicle, the speed at which the system can process algorithms determines the speed at which the vehicle can move. Increases on the order of 15x have already been reported. This is now attracting the development of applications in fields as diverse as medical research and life sciences, fluid dynamics, signal processing and video processing, among others. For example in computer fluid dynamics (CFD) applications, the program algorithms model the physics and then for each iteration of the model, render a display image. Thus, high-speed modeling and graphics generation are taking place on the same set of processor cores. In video processing, the same GPU that is known for generating graphics is also able to take in that

editor’s report

video data and process it for things like face recognition and other image processing tasks. The CUDA architecture consists of a number of parallel general computing cores. There are different versions of CUDA processors, and performance varies depending on the number of cores in a particular model. The software model, however, is consistent across all versions of CUDA. CUDA still acts in the role of a coprocessor with a general-purpose CPU (at this point an x86) running the main program. The developer writing in C with extensions for CUDA decides which parts of the code are appropriate to run on the GPU and identifies them. When the main processor encounters a section of CUDA code, it diverts it to the GPU, which returns the results as appropriate to memory (Figure 1). OpenCL, a parallel compute model for parallel processing across heterogeneous devices, enables compliant drivers to be used to run the parallel code. As mentioned, GE Fanuc has entered a partnership with Nvidia, and as a major prime contractor in the military/aerospace industry, has evaluated the CUDA architecture in a radar system and found that performance improvement of 15x is achievable with minimal reprogramming effort. Other prime contractors are expressing substantial interest in the technology. In addition to having potential applications in radar, signals intelligence, and video surveillance and interpretation, GE Fanuc sees GPUs based on the CUDA architecture to represent big potential in other application areas including target tracking, image stabilization, SAR (synthetic aperture radar) simulation, pattern recognition, video encoding/decoding, graphics rendering, object recognition, cryptography, sensor processing and software defined radio. Traditional graphics applications will also benefit greatly from high-performance Nvidia GPUs. The first board will be 3U VPXbased and will benefit from a broad-

ranging supporting infrastructure of other GE Fanuc 3U VPX solutions and services, allowing customers to specify tailored multi-board solutions. Alternatively, a CUDA-enabled 3U VPX rugged graphics subsystem will also be available. A family of 6U VPX products will feature combinations of Intel dual core processors and Nvidia CUDA-enabled GPUs, allowing designers to create powerful multi-board solutions for demanding applications. This family, as well as the new 3U VPX platform, will additionally support OpenCL. Beyond that, the potential for other compute-intensive applications support exists for any number of vendors adapting CUDA or similar parallel architectures. Main Memory

Nvidia’s Web site offers a software development kit for CUDA including a set of drivers, a set of CUDA development tools and a library of code samples. The development environment is available for Windows XP, Vista and 7 as well as for Linux and MacOS from its OpenCL downloads page. These can be used for developing a wide range of general-purpose computing applications on the CUDA architecture.

Programming the Guts of the SoC

The System on Chip (SoC) emerged as an alternative to a general-purpose processor plus other components and a full-custom ASIC yet without having to resort to an FPGA. It tries to stretch the concept of general-purpose with enough


1 Copy processing data

Instruct the processing 4

2 Copy the result

Memory for GPU GPU ( GeForce 8800)

Execute parallel in each core 3

Processing flow on CUDA

Figure 1 In the CUDA processor architecture the code running in main memory may contain sections that have been designed to run algorithms on the parallel CUDA cores. The main CPU then diverts that code to the CUDA device, which processes it and returns the results to main memory. RTC MAGAZINE NOVEMBER 2009


editor’s report
























I2C FS USB 2.0












Figure 2 The three layers of the PSoC architecture are, except for the processor core, the same for both the PSoC3 and PSoC5 devices. I/O can be routed to selected external pins on the chip and to internal elements via the programmable routing and interconnect block.

specialization to make itself attractive to more than its initial target application. A number of semiconductor companies have contracted to develop a processor with a select combination of peripherals on-chip for a given customer’s needs with the proviso that they would then carry that design in their product portfolio and make it available to other customers. The result—especially in the 8-bit microcontroller arena but also with 32bit devices—is a vast number of devices all built around the same core. Just think of the number of 8051-based devices that are out there on the market. And there are many more built around other cores. Finding the right one with the desired mix of peripherals requires Web searching and poring over catalogs. Then, of course, if you want to scale up and move your work to a 16- or 32-bit platform, the fun begins again. Recently, Cypress Semiconductor has added a scalable 8- to 32-bit path with additions to its existing 8-bit line of programmable systems-on-chip: the PSoC device, in the form of PSoC3, based on an 8051 core, and the PSoC5, which is built around the ARM Cortex-M3 and is capable of running ARM’s Thumb 2 16bit code or 32-bit code. The two devices



represent a scalable, programmable platform for embedded development that incorporates a wide selection of digital and analog peripherals. The PSoC3 and PSoC5 architectures are differentiated mainly by the different processor cores but are essentially similar (Figure 2). The CPU subsystems consist of one of the two CPUs and several hard-wired blocks for CAN 2.0, USB 2.0 and I 2C interfaces as well as a DMA controller, EEPROM, flash memory and debug interfaces among others. The chip has an interface to external memory as well. The blocks on the CPU subsystem have connection paths to the programmable routing and interconnect block on the programmable routing and interconnect layer where they can be routed to external GPIO ports or peripheral blocks on the configurable analog and digital layer. The analog and digital layer consists of two sections. The digital subsystem contains an array of universal digital blocks (UDBs), which consist mainly of programmable logic that can be defined as digital peripherals such as UARTs, pulse width modulators, LCD drivers and controllers, timers, multiplexers and many more by means of the PSoC Creator de-

velopment tool. The UDBs can also be set up for custom logic such as “glue” logic or selected other functions. By the same token, the analog subsystem contains programmable elements for DACs and ADCs, op-amps, comparators and customizable analog blocks that can be defined from a library of pre-built, characterized components, again using the PSoC Creator tool. In addition, it contains a configurable digital filter block capable of two separate filter channels and up to four cascaded filters per channel. Analog channels are capable of up to 20-bit resolution with ± 0.1V reference voltage. Both PSoCs have a wide operating voltage of 0.5Va to 5.5V and can be set to operate with external devices at different voltage levels in up to four different voltage domains. In addition, there is a power management section with three power modes: active, sleep and hibernate. The latter consumes only 300nA in hibernate mode. Configuration of the PSoC 3 and PSoC5 is done with a graphical tool called the PSoC Creator, which contains a library of pre-defined, yet custom-configurable peripherals. To configure the chip, the developer simply drags a peripheral from the list onto the screen (Figure 3) and connects wires to other peripherals, the CPU, I/O and memory blocks or to GPIO ports and then to external pins on the device. Peripherals such as digital filters can be further configured by setting their parameters to the appropriate values. Peripherals can be given appropriate names, which then become part of the API and callable from application software. The PSoC devices additionally are supplied with a set of third-party compilers and IDEs for developing application code in C as well as on-chip JTAG debug and trace tools. Code, for example, that was developed for the 8-bit PSoC3 can be readily recompiled for use as 16-bit or 32-bit ARM applications. In addition, a peripheral configuration developed for the 8-bit device can readily be ported to the ARM-based version allowing upscaling of a given solution and the ability to enhance it on the new platform without redoing the earlier development.

editor’s report

ARM Joins the FPGA

The ARM Cortex architecture is also moving into the world of the FPGA for a new generation of programmable logic devices with embedded hard-wired processor cores. Xilinx, whose successful Virtex line of FPGAs that incorporate the PowerPC architecture has gone through a series of versions and enhancements is now adopting ARM Cortex processor IP, using performance-optimized ARM cell libraries and embedded memories for their future programmable platforms. In addition, ARM and Xilinx are working to define the next-generation ARM AMBA interconnect technology that is enhanced and optimized for FPGA architectures. The agreement indicates Xilinx’s intent to adopt ARM technology to provide flexible computing platforms where IP and software development can be shared and re-used on a broad scale. To help create a more optimal ecosystem serving both Xilinx FPGA and ARM IP technologies, the two companies are already working with a number of IP providers and EDA vendors including Cadence Design Systems Inc., CAST, Inc., Denali Software, Inc., Mentor Graphics Corp., Northwest Logic, OMIINO Ltd., Sarance Technologies Inc., Synopsys, Inc. and Xylon, to support an advanced version of the AMBA specification, the de facto standard for on-chip fabric communication, which is closely aligned to ARM processors. Not only will this new interconnect simplify and extend the capabilities of next-generation programmable platforms using 32-bit processor IP, but the definition of the standard is aligned with the “Socketable IP” aspect of the Xilinx Targeted Design Platform strategy. Because IP reuse is an essential component in reducing system development costs and timescales, this plug-andplay approach means that IP developed by Xilinx and its ecosystem can be easily used without requiring the user to make a huge investment in vendor support. The AMBA protocol for on-chip fabric communication details a strategy for the interconnection and management of functional blocks that make up a system-on-chip. (The new collaboration

Figure 3 The PSoC Creator tool lets developers select and configure on-chip peripherals by drawing them in a schematic as one would on paper or a white board.

with Xilinx optimizes this for FPGA implementations and introduces a new use model of being used with or without processors as well.) The move to ARM aligns with the Xilinx Targeted Design Platform Strategy announced earlier this year, when the company mapped out plans to roll out development platforms based on FPGA technology, boards and standardized IP. A key component of the Targeted Design Platforms is ecosystem participation enabling mutual customers to easily scale their product designs around a stable architecture. Neither company has been specific about concrete product plans at this point, such as whether the agreement will entail both hard and soft versions of the ARM architecture or what types of devices will be rolled out. The Xilinx line of Virtex FPGAs consists of several families optimized, for example, for high-performance logic, DSP applications, high-speed serial connectivity and the like. At this point there have been no products announced, but the penetration of the ARM architecture into ever more deeply embedded applications is certainly apparent and serves as yet another example of the ongoing

effort in the industry to find the “sweet spots” among performance, flexibility, integration and speeding time-to-market while bringing down costs. It is a neverending quest. NVIDIA Santa Clara, CA. (408) 486-2000. []. GE Fanuc Intelligent Platforms Charlottesville, VA. (800) 368-2738. []. ARM Cambridge, UK. +44 01223 400400. []. Cypress Semiconductor San Jose, CA. (408) 943-2600. []. Xilinx San Jose, CA. (408) 559-7778. [].



Technology in


New Direction for Mezzanines

New Generation of M-Modules and PMCs Saves Development Time and Money FPGA technology built into a flexible submodule that can fit onto popular mezzanine standards can speed time-to-solution for a variety of specialized I/O demands. by Barbara Schmitz, MEN Mikro Elektronik


ndustrial embedded applications with small or medium volumes need tailor-made control solutions that can be implemented as much as possible on the basis of standard components. The development of any mezzanine concept is focused on flexibility to achieve low design costs and a short time-to-market. This is true for the various system-onmodule (SoM) families, which accommodate the CPU functionality on a plug-on module, as well as for mezzanine standards like PMC or M-Module, which typically place I/O functions on the plug-on module. Despite mezzanine technologies, the differences between I/O requirements may be huge, so it is not always possible to meet demands through available standards. Cost-effectively mitigating development costs and production time has historically been a challenge for embedded designers. One approach to optimizing I/O flexibility is the Universal Submodule (USM) that uses FPGA technology, which has become inexpensive for even simple I/O demands. The functionality of the mezzanine module is defined solely through



Configuration CPLD

Flash 2 MB

50-pin USM 50 I/Os front Universal Submodule connector

46 I/Os

FPGA Cyclone II

PCI Interface 32-bit/ 33-MHz


Figure 1 The USM concept standardizes many of the complex components to limit development costs and time to market.

the combination of IP cores in the FPGA. In addition, FPGAs—and therefore any of the installed functions—are easily upgraded and readily available, which means they are unaffected by obsolete components, and are qualified for the extended temperature range of -40° to +85°C. USM was designed in such a way that it can be used for both PMC modules

and M-Modules, the two most popular mezzanine standards. PMC derivates, such as XMC, and the option of conduction cooling for MModules and PMC modules were also taken into account during the concept’s development. These mezzanine types are typically used in 19-inch systems such as CompactPCI or VME but also on stand-

The Final Piece is Your Application Performance Technologies furnishes Application-Ready Systems designed to accept your customization and go straight to market, saving you valuable time and resources. These pre-integrated systems allow you to focus on developing the most important piece of the product design puzzle – your application.

Each IPnexus ® MicroTCA® System comes complete with platform, NexusWare ® (Carrier Grade Linux ®), processors, communications, storage, APIs, middleware, web-based management, and more. These finely-tuned systems

Application-Ready Systems

are designed for long life-cycle availability, guaranteeing product sustainability.

To learn more visit:




Phone: 585.256.0200

© 2009 Performance Technologies. All rights reserved. All trademarks or registered trademarks are the property of their respective owners.

technology in context



cPCI Carrier



M-MODULE Carrier

Component Area M-MODULE Carrier (b)

Figure 2 By limiting development to the carrier board, USM is a time-saving development method for CompactPCI (a) and M-Module (b) that also provides scalability for future technology upgrades.

alone platforms. A 6U board can accommodate up to four M-Modules or three PMCs, while a 3U board can carry one PMC or up to two M-Modules.

Looking Forward with FPGA Technology

The Universal Submodule concept gives life to a new generation of M-Modules and PMCs that implement the desired functionality entirely in the shape of an IP core or a number of cores inside the FPGA. The FPGA itself is located on the “basic” PMC or “basic” M-Module and is a Cyclone II component from Altera in different versions starting at a size of 18,752 logic elements (which corresponds to about 225,000 gates). The hardware configuration of the FPGA is loaded during the boot phase from an external, 2 Mbyte NOR Flash using a CPLD, and is serially transmitted to the FPGA (Figure 1). The flash memory contains the source



code for a minimum configuration of the FPGA over a PCI bus connection. When the Cyclone II is programmed with the basic configuration, the hardware configuration itself can be loaded into the second area of flash memory via the PCI bus. The Nios soft processor, also implemented in the FPGA, can provide local intelligence where needed. Accordingly, up to 32 Mbytes of DDR2 SDRAM can be assembled on the basic “main module.” The cores inside the FPGA may have totally different functions from computer I/O, such as graphics, Ethernet and UARTs, to mobile or industrial communication with fieldbus connection, including CAN bus, Profibus, Industrial Ethernet, IBIS and MVB, up to typical industrial functions such as digital/analog process I/O, motor control and SSI. The configuration of a USM-based M-Module or PMC can be changed at any time through the implementation of different IP cores. Except for their FPGA contents, the main

M-Modules or main PMCs have a fixed hardware setup. On the front, the userdefined I/O signals from the FPGA are led to a 50-pin SCSI 2 connector, chosen because it fits on the mezzanine card with the smallest width, the M-Module. The line drivers related to the function of the IP cores are decoupled from the main module and are implemented on the USM. The USM simply plugs into the corresponding main PMC or main M-Module (Figures 2a and 2b). The mechanical setup of the USM provides maximum space for electronic and mechanical components. This not only includes the available surface of the USM but also the component heights. Maximum component heights are defined by the different mezzanine standards, and therefore vary. In order to reasonably place a USM within the very small dimensions of a PMC module, a special pass-through connection is used. As for the maximum surface, different mezzanine card formats and the additional space needed for the temperature interfaces on a conduction-cooled PMC module were taken into account. The electrical connection between the USM and its main M-Module or main PMC consists of two 64-pin plug and receptacle connectors, respectively. J1/P1 makes the connection that transfers the communication signals generated on the USM back to the main module. On this card, J1 is directly connected to the 50-pin front connector. The signal lines are routed in such a way that there is a sufficient isolation distance to develop a USM with four electrically isolated channels, for instance. The second connector, J2/P2, is the electrical connection between the main module’s FPGA and the USM. Among others, 12 ground pins are evenly distributed over the complete connector. The USM is supplied with 3.3V and 5V using two pins each. Since I/O functions are realized in the FPGA, the lifetime of a PMC module or M Module no longer depends on the availability of commercially available components. Even after 10 years and more, older IP cores can be updated and brought into a newer and maybe larger FPGA. Both MModule and PMC main modules and the USM plug-in modules are designed for an

technology in context

operating temperature of -40° to +85°C. To meet demands for increased shock and vibration resistance, the boards are equipped with soldered components and sturdy connectors. The first USM-based standard PMCs brought to market include a four-channel CAN bus interface, a four-channel RS422/485 interface, dual Fast Ethernet and a reflective memory PMC module. To implement these functions, four different USM plug-in modules could also be used to bring the same functionality to MModule mezzanines.

Broad Range of Applications

Due to the universality of the USM concept, there are no restrictions regarding the target markets and the type of application. A number of implementations are possible from machine control in industrial settings to test and simulation systems in automobiles or in telecommunication up to monitoring and control in mobile and safety-critical applications. However, the following are a few realworld examples of how the USM concept is being implemented across various applications (Figure 3). In one of the first applications to use M-Modules as a hardware platform, the USM-based M-Module was an addition to existing CompactPCI systems and also to industrial PCs where a number of other M-Modules were already being used. Since the desired functionality was very specialized but had to be combined with different functions on a single module, USM provided the fastest and least expensive solution. The FPGA contains a CAN bus controller already available as a standard core as well as a K-/L-Line interface. Apart from the CAN bus and K-/L-Line line drivers, the USM includes a FlexRay controller and digital I/O in hardware. The USM-based M-Module is a part of computers used to test anti-lock braking system (ABS) control devices inside vehicles and is used both in the lab and in series production as well as for reprogramming during maintenance. In an application for testing systems for the aircraft industry, PMC main modules are the platform of choice. Here, PowerPC-based VMEbus computers are used

Figure 3 Cost-saving USM-based modules are employed in a number of applications from light duty industrial to heavy, missioncritical.

in simulation, validation, test and maintenance. Interfaces such as ARINC429/629, MIL-STD-1553 and AFDX, which may vary from one application to another, are integrated using PMC modules. In order to guarantee that current projects are future-safe, controller components for airplane fieldbuses, such as ARINC429, are implemented only as IP cores in the FPGA, with the line drivers realized on a USM plug-in module. The performance of the FPGAs is sufficient for up to 32 channels in the system with a data rate of 100 Kbits/s each.

Comprehensive Development Package

A considerable range of functions implemented as IP cores are now available on the market, for example from For its own boards, MEN Micro uses a continuously growing number of IP cores designed on its own on the basis of Wishbone, which are also used to complete customized solutions. With the growing acceptance of FPGA-based approaches, more users can develop and integrate their own cores. To do this—as is the case with the USM concept—users should get support from the hardware manufacturer. For this reason, there is a comprehensive development package for M-Modules and PMC modules based on USM that enables the user to easily and quickly turn very specific I/O requirements or individual functional combinations not available as a standard configuration into series products.

The USM development package includes a main M Module (M199 equipped with an FPGA, 32 Mbyte DRAM and 8 Mbyte flash) or a main PMC module (P599 equipped with an FPGA, 32 Mbyte DRAM and 2 Mbyte flash). In addition, the package includes a bare USM plug-in module, a test board where I/O signals from the FPGA are led, a SCSI cable for connection between the main module and test board as well as an FPGA package. A debug interface for the Nios soft core is included on the main modules. The FPGA package comprises the Nios processor, memory control, connection to the PMC or M-Module and the Avalon/Wishbone bridges. For development of IP cores on the standard Wishbone bus, the Wishbone BusMaker tool from MEN Micro is included. In order to use the Nios core and to develop IP cores on the Avalon bus, users also need Altera’s Quartus II design environment including the SOPC builder.

USM as an Open Specification

To give access to the new concept to other vendors of PMCs and M-Modules, MEN Micro has published the entire USM specification. It documents the mechanical and electrical characteristics as well as environmental requirements for the different main modules and the corresponding USM plug-on cards. The driving requirements for the development of this concept included a simple design structure, low production and design costs, ruggedness for harsh and/or mobile industrial environments and a maximum size of the USM module. The same USM was designed to be as large as possible and still be pluggable onto M-Modules, PMC modules or XMCs and conduction-cooled PMC modules. It is even possible to use two USM cards on one Eurocard. For the user, new designs of future USM-based PMCs and M-Modules primarily involve the FPGA content, and therefore save significant development time and costs. MEN Mikro Elektronik Ambler, PA. (215) 542-9575. [].




engineering MicroTCA Spans Applications

Concept to Reality: An Application Architecture with MicroTCA The flexibility and capabilities of AMCs and MicroTCA to develop embedded applications are endless. Even more rewarding is that COTS-based application-ready systems allow faster time-to-market with less development cost. by Tony Romero, Performance Technologies



solutions engineering


uch has been written about the virtues of MicroTCA platforms and AdvancedMC (AMC) modules. Both of these technologies have been maturing and there are now numerous options and flexibility to configure a wide range of applications with products that exist today. So let’s explore in detail some application configurations. In the examples below, we will look at all the components that are necessary to configure an end-user application. On the hardware side, this includes the MicroTCA chassis (which includes the backplane, power and cooling), the fabric switching such as Ethernet, PCI Express and SATA, IPMI Platform Management, and the AMC payload boards. The payload boards can range from processors, to storage/video, to I/O access cards. On the software side, this includes the Operating System, the Development Environment and, of course, the application software. MicroTCA and AMC are COTS-based standards. As such, components from one vendor are specified to be compatible with other vendors components. This is highly beneficial, as it offers embedded designers with the freedom to choose components from as many vendors as necessary to configure their specific application. In fact, there are industry events occurring annually whereby multiple COTS vendors meet to verify compatibility between AMC modules and MicroTCA platforms, enabling cross-company product integration. While it is fundamentally true that the chassis, mezzanine and MCH components from different vendors are compatible with each other, often a significant amount of time and resources are required in order to complete the integration process to ensure final product compatibility and configuration. This gets more complicated when you need to incorporate an operating system, development environment, APIs, drivers, and other software stacks besides the application. So while on the surface, this seemingly wide range of options exists, there are several important nuances to AMCs and MicroTCA,

Figure 1 Screen Capture - MicroTCA Web-based remote management tool, NexusWare Portal.


DSP/Baseband PCI-e/SRIO


x4 DirectConnector or SRIO

CPU Enet

Platform Uplinks

ASN Gateway Ethernet

Figure 2 MicroTCA Base Station System of hardware and software components.

whereby a single vendor who offers a full suite of hardware pre-integrated with software components can maximize the value of a configured system, increase functionality and performance, and guarantee full compatibility directly from the onset. Listed below are several aspects of a fully integrated platform, called an Application-Ready System.

One is configuration. A single vendor who can integrate their components together—and right out of the box, can ensure that these components are configured properly, that the appropriate onboard switches are set, and test to make sure the hardware and software components are compatible and functioning properly. RTC MAGAZINE NOVEMBER 2009


solutions engineering

Trust a world-wide expert for your embedded critical network application

SWITCHES S & IP ROUTERS More than 3O models... VME, cPCI, VPX


3G Radio Access Network

Packet Processing

Control CPU Ethernet

Platform Uplinks

Connectivity Service Network (CSN)


Mobile WiMAX Radio Access Network

ComEth 4300a 10 Gig Ethernet ports Full wire-rate router

• 4 front Giga ports (copper or ber) • 2 front 10 Gigabit Ethernet ports • 20 rear Gigabit Ethernet ports • Compliant with PICMG2.16

VPX NEW switch 6U

ComEth 4340a 10 Gig Ethernet ports


Intel® & Freescale® p processors rocess

IC-De6-VMEb amazing processing unit

• Two MPC 8640 single or dual core • One embedded Ethernet Switch • One Open FPGA VIRTEX 5 • PMC/XMC, USB, RS232/RS422, IOs

New VPX Range • Intel® Core 2 Duo®, MPC8640, Virtex 5 VPX 3U Single Board Computers

For more information on our products and custom design services... +33 (0)298 577 176


Untitled-6 1


11/11/09 4:02:07 PM

Figure 3 A Wireless Gateway System.

Another important aspect is unified tech support. If any technical issue arises, a single tech support group can more quickly debug and isolate the issue through all the components that are integrated together. Dealing with disparate tech support groups can lead to delays, as they may not be able to replicate the same configuration, or not clearly understand a third-party board’s specific issues. Or worse, blame another vendor for the problem. Product uniformity may at first seem superficial, but this can range from the physical appearance, such as the labeling on the faceplates of the AMCs and platform, to functionality on AMC modules, software, or even firmware such as the sensor thresholds and FRU information that are programmed into the MMC management controller. Management software is an essential element, and several MicroTCA system vendors provide Web-based management portals (Figure 1) that are typically programmed to enable off-site development, systems and module management, as well as monitoring of system-wide status and operations. Finally, compatibility of the operating system suite with hardware must be assured throughout. In selecting a Linux software suite, make sure to consider a solution that provides all the necessary components such as Operating System, IDE, kernel builder, image builder, compliers, linkers, debuggers, cross platform develop-

ment, hardware drivers, network services, platform management and Web-based system management. In addition, a suite that is developed to be specifically compatible with a suite of hardware not only ensures compatibility, but also provides synergistic features for an overall solution. In all the application examples below, a 1U MicroTCA platform that houses six AMC slots is a perfect balance between having enough slots to house all the AMCs with room for scalability, but not more slots than will be cost-effective. What is also important to consider are the cost trade-offs between reaching the right level of high availability with going all the way to a fully redundant platform. Components that are most apt to fail, such as power supplies, fans, or hard drives, need to be redundant and easily serviceable. Electronic components on backplanes and/or motherboards have a high MTBF and offer a high level of reliability.

Example: Wireless Base Station

Data overload is already a reality with mobile operators, as consumers are rapidly adopting mobile smartphone devices and using them to download news and information from the Internet. One recent industry study involving millions of mobile device users indicated that these same users doubled their network usage in one year. 4G technologies, such as WiMax and LTE applications, are de-

solutions engineering

signed to handle heavy usage of mobile data. MicroTCA and AMC modules are a good fit with these types of applications, as they provide a carrier-grade platform with high-speed fabrics, high reliability, remote platform management and hotswap functionality in a cost-effective form factor. Even defense departments are keyed in on wireless solutions for Network Centric Warfare communications. Wireless applications require real-time signal processing with high-bandwidth communications on and off the boards. A 1U platform supports the four major functional blocks: 1) RF, 2) Baseband, 3) Control and 4) Transport. The AMC ecosystem offers AMC modules that support both the RF and Baseband functions on a single card or separate cards with a direct connection to the Remote Radio Head (RRH) for three-sector operation utilizing both a Xilinx FPGA and multiple TI DSP cores. The Control function is handled by a multicore processor AMC. And the transport for the backhaul is accommodated on Ethernet uplinks from the MicroTCA platform, or for legacy deployments, it can utilize a T1/E1 AMC module (Figure 2). The 1U MicroTCA platform is an ideal form factor with six AMC slots, as it can house all four functional blocks in a single shelf with room to scale to a second RF/Baseband AMC module if needed. Larger MicroTCA platforms that house 8 to 12 AMC slots are too large and are not cost-effective. For frequency and time synchronization of all the AMCs, the MicroTCA’s clock distribution can deliver a phase-synchronized Stratum 3 signal (e.g. 10 MHz or 30.72 MHz) to all the AMCs derived from an external Stratum 2 GPS Receiver outputting a 1PPS reference signal. In addition, if the GPS reference clock fails or is absent, a crystal oscillator can provide a holdover signal for a limited time until the GPS reference clock is back on line.

and aggregates traffic from base stations, manages the hand-off between the base stations, and provides quality of service. Here too, a carrier-grade architecture ensures hot-swap scalability, high availability and remote platform management. The two primary functional blocks for this

application are packet processing and a control CPU. A high-performance, dualcore processor AMC can handle both functions, and housed in the 1U shelf, can scale up to six processor AMCs. The typical backhaul network is Ethernet to the Connectivity Service Network (CSN), and

Example: Wireless and Media Gateways

A wireless gateway receives traffic from 3G Radio Access Networks or from a Mobile WiMAX Radio Access Network Untitled-4 1

5/12/09 2:17:42 PM



solutions engineering

the quad 1GbE uplinks on the MicroTCA platform provide ample bandwidth (Figure 3). Media Gateways such as the ones that convert media streams from a Public Switched Telephone Network (PSTN) to an IP-based Network can also take advantage of the MicroTCA architecture. T1/ E1 Communication Controller AMCs can receive multiple lines and perform highperformance and high-capacity processing via a PowerQUICC controller. These AMCs have onboard FPGA with the ability to transmit TDM traffic as I-TDM packets over the Ethernet fabric. In this example, the Media Gateway can scale up with more T1/E1 AMCs as subscriber traffic grows. Some modern Media Gateways also integrate Session Initiation Protocol (SIP)-based signaling and call control to function as stand-alone units for independent and intelligent SIP end-points. In addition to the operating system, the other software stacks that may be required are a SIP stack and/or SS7 MTP2 (Figure 4). Another type of application that can take advantage of MicroTCA is carrier-


T1/E1 PCI-e


Enet/ I-TDM


Enet/ I-TDM


Enet/ I-TDM

Platform Uplinks


Figure 4 Components for a MicroTCA Media Gateway System.

grade, high-reliability, IPMI-based managed service appliances. Service applications can range from IPSec, deep packet inspection, load balancing, Quality of Service, traffic management, IVR and other content-based services. Deploying in a MicroTCA platform allows for hotswap scalability by adding more blades

Need to match price and performance with tight time-to-market deadlines?

dynamically as more service processing is required. It can also provide redundancy if one processor fails. Finally, raw processor blades, which have no physical storage medium and no OS or application installed, can be provisioned in real time as they are added to the system. These AMCs can be installed in a platform and undergo a Preboot eXecution Environment (PXE) to boot via the integrated Ethernet fabric in the MicroTCA platform. One of the processors in the system can act as a DHCP server to assign an IP address and send the client AMC a list of Boot Servers. This same AMC can act as the TFTP server to provision the new client AMC module with its operating system, application software and any data files. Performance Technologies Rochester, NY. (585) 256-0200. [].

Problem solved… Schroff’s ATCA and μTCA solutions incorporate your unique design requirements, plus optimize system performance. Through our global manufacturing infrastructure, engineering, technical service and support network, we’ll help you get your product to market sooner with a price/performance ratio that will accelerate your path to market leadership.

Partner with Schroff …


Untitled-5 1


IP Network

11/17/09 11:53:54 AM

ER! F F O TIME the latest D E T LIMI 2,000 on Module Save $ are Radio Softw

We’ve Hatched the Next Generation of Software Radio Solutions

Pentek delivers board and system-level SDR products that include digital downconverters, upconverters and transceivers. Since most of these products are FPGA-based, Pentek offers powerful factory-installed IP cores plus the GateFlow FPGA Design Kit for custom development. These software radio solutions are perfect for applications in communications, SIGINT, radar, direction finding and many more. And now the latest software radio module, Model 7156 with Dual 400 MHz A/Ds, 800 MHz D/As and Virtex-5 FPGAs, is now available in a bundled package offering you a $2,000 savings! Call 201-818-5900 or go to for your FREE Software Defined Radio Handbook, technical datasheets and to request pricing.

s s s s s s s


Pentek, Inc., One Park Way, Upper Saddle River, NJ 07458 Phone: 201.818.5900 Fax: 201.818.5904 Worldwide Distribution & Support, Copyright Š 2009, Pentek, Inc. Pentek and GateFlow are registered trademarks of Pentek, Inc. Other trademarks are properties of their respective owners. Prices are subject to change. Offer ends December 31, 2009. Offer applies to the Model 7156 only.


engineering MicroTCA Spans Applications

High Bandwidth and Small Form Factor MicroTCA for Flexible Embedded Design Modular AdvancedMC modules are instrumental to the success of MicroTCA in multiple markets including telecom, networking, medical and other key embedded markets.

by Sven Freudenfeld and David Pursley, Kontron

Embedded systems must meet the growing demand for ruggedness, flexibility, mobility and high-end processing, and these attributes should be designed into standards-based solutions that can be developed and implemented quickly and within budget. Beyond applicationspecific requirements, many embedded system designs may also require features that address low-power, size and scalability challenges. MicroTCA, comprised of modular COTS components such as Advanced Mezzanine Cards (AdvancedMCs), offers some significant design advantages to solve these challenges. AdvancedMCs are the building blocks that make MicroTCA an ideal platform for a broad range of high-availability, highbandwidth embedded designs. Telecom and medical are two market examples that have a diverse set of application requirements, with each demanding a certain level of design flexibility. MicroTCA possesses the practical and cost-effective versatility to meet these demands. Initially designed as a hot-swappable component within AdvancedTCA (ATCA)



Figure 1 A series of AdvancedMC GbE/10GbE Intelligent IO modules is based on the next-generation Cavium OCTEON multicore packet processor. The AM42xx AMC modules are optimized for layer 4 to 7 data and security processing, targeting access and service providers with 3G/4G BTS, RNC, xGSN and Media Gateways.

platforms, AdvancedMC modules brought a modular, open-standard approach at the mezzanine level for telecommunication

equipment manufacturers. AdvancedMCs are used to add specific features or functions to a particular application. Typically

solutions engineering

integrated in tandem with an ATCA carrier board or within AdvancedMC slots in a CPU or switch blade, AdvancedMCs allow for hot-swap and redundant system management within an ATCA system, or provide additional processing functions on any single ATCA node or switch blade. Another common use for AdvancedMCs is in redundant and non-redundant MicroTCA platforms. Here the form factor can vary between single or double-wide AdvancedMCs and MicroTCA platforms, and provides full-scale PCI Express, GbE, 10GbE or Serial Rapid I/O infrastructure to the system being implemented. A growing number of applications are also using AdvancedMCs with low-cost backplanes and a reduced-functionality infrastructure. This retains the standard AdvancedMC implementation but reduces implementation costs, allowing OEMs to only pay for what they really need. AdvancedMCs have evolved from single or double-wide modules, and based on the PICMG specification, MicroTCA can scale from four non-redundant to up to 12 AdvancedMC modules with redundancy in a MicroTCA backplane. This delivers extensive flexibility in a smaller footprint—a critical feature for MicroTCA-based designs in demanding applications such as telecom or medical imaging. Low-cost AdvancedMC system implementations (designed with direct interconnectivity between two to four AdvancedMC modules via the backplane) also deliver design flexibility in cost and functionality.

AdvancedMC Evolution

Different usage models for AdvancedMC modules have further broadened their use and economies of scale for embedded systems. For example, TEMs can free up valuable ATCA slots by combining general-purpose processing with packet processing functions through the use of AdvancedMC slots with PCI Express, GbE and even 10GbE on a single ATCA blade. This scalability and flexibility have become essential in helping manufacturers manage system upgrades while simultaneously protecting their initial ATCA platform investments. Since AdvancedMCs could be used for a wider range of system-level func-

Figure 2 The Kontron OM6120 brings significant cost improvements compared to conventional MicroTCA platforms. Cost optimization includes Power Management and Fan Control on the backplane, and pluggable Power Supply Units instead of MicroTCA Power Modules. By accommodating a high number of multicore AdvancedMC modules and allowing a tight coupling of processors over high-speed communication links (over the backplane), the system is also well suited for image processing applications in single-channel or dual-channel architectures.

tions, it became apparent that their use could extend well beyond ATCA-based telecommunication applications and into other high-performance computing environments such as medical, industrial and even transportation applications. AdvancedMCs plug directly into the MicroTCA backplane—fundamentally turning the mezzanine itself into a blade, which enables designers to take full advantage of a flexible platform to readily create a small yet highly powerful and scalable system. Consequently, AdvancedMCs can be used as a main controller, data server, traffic processor, image or signal processing engine, security appliance, or network processor—giving designers a cost-effective solution to accelerate to market a finished system that can feature multiple applications on the same platform.

New Telecom Application Enablers

Intelligent networks are being deployed in enterprise, datacenter, broadband and access networks—a new genera-

tion of systems that must effectively handle highly distributed, service-related or legacy architecture applications. As a result, AdvancedMC modules have become an attractive design alternative, meeting needs for deep packet inspection processing (vital to monitor and modify network activities), TCP/IP packet processing and load balancing, content aware applications, security processing, compression/ de-compression and other new and evolving communications services. By providing information about typical activities such as traffic profiles, users, sources or destination, carriers and wireless providers can introduce new tiered services, as well as quality of service (QoS) and enhanced network efficiencies that enable smoother operations and new revenue opportunities. Multicore AdvancedMCs are integral to handling this range of packet and security processing functions including forwarding, load balancing, traffic management and IPsec in both ATCA- or MicroTCA-based environments. AdvancedRTC MAGAZINE NOVEMBER 2009


solutions engineering

Common Options MCH1Fabric [A] to AMC Port O

Common Options MCH2 Fabric [A] to AMC Port 1

AMC Port 2 AMC Port 3

Fat Pipe MCH1 Fabric [D:G] to AMC Port [4:7]

Extend. Fat Pipe MCH2 Fabric [D:G] to AMC Port [8:11]






















Figure 3 The port mapping and use of clocks show how a MicroTCA backplane can be used to enable the synchronization of processor-based AdvancedMC modules to an internal or external clock for applications such as real-time processing or complex robotic motion-control.

MCs were typically released as Gigabit Ethernet modules, but now more multicore packet processor modules support 10GbE for higher bandwidth applications. This has spurred new growth opportunities for MicroTCA platforms to be designed for media servers, signaling gateways, CMTS-cores, radio network controllers, and more recently, new LTE base station solutions that deliver nextgeneration broadband wireless technology for wide area networks. Sheer cost, power and size efficiencies make MicroTCA ideal to mass deploy, for example, radio access eNodeB systems into the field. The platform approach can include configurations of I/O, general-purpose processor and DSP (Digital Signal Processor) AMC modules (Figure 1).

AdvancedMCs as the Foundation

Selecting the right AdvancedMC for


Untitled-7 1


11/10/08 10:01:37 AM

a system design depends entirely on the features and functions required for the system. Designers have the choice to select an AdvancedMC based on processing performance, communications bandwidth, storage or specific I/O options. For example, processor AdvancedMCs are employed to host microprocessors and network processors. Storage AdvancedMCs are used to host flash memory cards or perhaps a solid-state drive if the design requires critical data storage and support for extended temperature ranges. Since AdvancedMC modules offer a flexible combination of size, performance options and features, designers can create MicroTCA systems with as few as one or two AdvancedMC modules, or as many as the 12 allowed by the MicroTCA specification. This makes MicroTCA cost-effective and extremely flexible, with each slot populated with a different type

solutions engineering

xTCA Expands its Reach with a Thriving Ecosystem

Take RF Measurements Up to 10X Faster

Originally developed to fulfill telecom and carrier-grade requirements for system-level projects designed by telecom equipment manufacturers, the Advanced Telecommunication Computing Architecture (ATCA) standard has experienced an ongoing fine-tuning process. Gaps in the standard have led special interest groups such as the SCOPE Alliance to define a common profile for standard hardware, software components and systems. Since hardware components were standards-based, a seamless integration into a multi-vendor environment was not guaranteed, thus necessitating a common methodology for hardware component testing and verification, which was taken on by the Communication Platform Trade Association (CP-TA). Today, xTCA (ATCA and MicroTCA) is an expanding ecosystem—comprised of component providers, blade vendors, CG OS vendors and system integrators—and includes software vendors offering platform management and middleware solutions. TEMs and NEPs have the freedom to select the best performing general-purpose components and platforms, preventing the “lock-in” approach with second sourcing. Due to the evolution of Gigabit Ethernet technology and the future availability of new switch silicon, recent and ongoing revisions to the PICMG specification are opening the door to other telecom networking applications in core networks that require 40GbE. Future 40GbE-designed ATCA platforms could be used for optical line terminals (OLT) or point-to-multipoint GPON network infrastructures, ideal for the delivery of video services, quality voice and high-speed Internet access. GPON allows one single-feeding fiber from the provider’s central location to deliver data via passive splitters to multiple homes and small businesses.

of AdvancedMC module. For example, a 1U six-slot design could feature a mixed and redundant configuration of processor, storage and packet processor modules to be used for central office SIP Server, SSL offload, content-aware processing and QoS over Ethernet applications. When looking at 12-slot MicroTCA platforms, the resulting design possibilities can make it popular for any number of compute-intensive applications such as medical imaging and screening, and diagnostics and therapy. As a configuration of such a high number of multicore AdvancedMC processor modules that are tightly coupled over high-speed communication links (i.e., over the backplane), this cube form factor is well suited for a range of access and edge network elements for voice, data and video networks. Two of these MicroTCA cubes could even be mounted together in a 19” frame, thus providing space for 24 AdvancedMCs in a 5U 19-inch form factor, achieving an unparalleled density of computing in terms of AdvancedMC space per total chassis volume (Figure 2).

Introducing the 6.6 GHz RF Test Platform

MicroTCA and AMCs in Medical Applications

Many medical applications require high bandwidth and overall processing power, and are an ideal environment for the power and flexibility of a MicroTCAbased platform. Image processing—with emphasis on the processing—is growing as a valuable medical diagnostic tool in the analysis of images and related patient information. Pattern recognition, rendering of organs, and volumetric analysis of a range of image types not only demand heavy processing power themselves, but are frequently tied back to specific patient information in medical databases. Familiar medical procedures such as X-rays have been using digital technologies for some time now—and the processing demands are only increasing. CAT Scans, for example, provide three-dimensional images based on X-rays, ultrasound or nuclear medicine functions such as Positron Emission Tomography (PET) or Magnetic Resonance Imaging (MRI). Coming soon are live images, delivered and used in real-time therapy or diagnostic applications. Consider mass screening


Test multiple wireless standards at rates up to 10X faster Harness industry-standard PC technologies such as multicore processors and PCI Express Increase flexibility and cost-effectiveness with a modular, softwaredefined solution

View product demonstrations at 800 891 8841

©2008 National Instruments. All rights reserved. National Instruments, NI, and are trademarks of National Instruments. Other product and company names listed are trademarks or trade names of their respective companies. 2008-10524-101-D

Untitled-2 1



9/25/09 10:15:34 AM

solutions engineering

for skin cancer by using spectroscopy in combination with image processing. Advancements in multicore processing, represented by the latest Intel processors and chipsets, have in turn led to the development of new imaging applications that might not have been previously possible due to limited platform bandwidth. Today’s image processing requires many cores, frequently on multiple blades, either tightly coupled over high-speed buses like Serial Rapid I/O or loosely coupled over network protocols such as 10GbE or GbE. Image rendering further demands high-performance graphics boards, which interconnect over high-speed PCI Express lanes. Designers have plenty of options, and high-level imaging can be achieved by stacks of servers or industry standard server blades. However, compared to isolated servers, systems based on server blades include the benefit of higher computing density and the tight coupling of processors over the backplane. In MicroTCA platforms that incorporate multicore AdvancedMCs, the Ethernet network infrastructure is built right into the system. With the ability to utilize up to 12 AdvancedMCs, medical systems can also support effective system management, power management, Ethernet switching and other transport systems, in addition to the processing performance required for compute-intensive imaging applications (Figure 3).

Modular, Powerful and Effective

Ultimately, AdvancedMCs provide diverse applications across a range of communications, networking and medical implementations. Integrated with a MicroTCA 2U platform, AdvancedMC modules can be configured, along with other processor and storage AdvancedMC modules, for powerful integrated security services. When that MicroTCA system builds to 3U or perhaps 4U—and every blade has incorporated a multicore processor—the resulting systems could be operating today with as many as 24 cores. Even with that extremely high level of processing power, such a system would still maintain a very small footprint, an achievement that many designers consider MicroTCA’s most powerful benefit.


Untitled-5 1


10/12/09 3:07:36 PM

AdvancedMC modules have demonstrated their versatility and flexibility as COTS building blocks for AdvancedTCAand MicroTCA-based systems, solidifying their role as a cost-effective solution to manage a variety of next-generation networking functions. Further, AdvancedMCs have translated that ability to handle high-performance, high-availability processing across a variety of embedded markets including telecom, networking and medical implementations. As a wide range of embedded systems and markets continue to drive technology toward open, standards-based multicore computing platforms, AdvancedMCs will continue to deliver ongoing cost and design advantages in a high-bandwidth, high-performance small form factor. Kontron Poway, CA. (888) 294-4558. [].

Extreme performance across the board. NEW! XV1™ (Quad-Core Intel Xeon-based VME SBC) • Quad-Core Intel® Xeon® 2.13 GHz processor • Up to 8 GB ECC DDRII SDRAM memory • CompactFlash™ slot • Up to two mezzanine slots on board • Up to three Gigabit Ethernet ports • Four USB ports and three SATA II ports • VITA 41 compliant • Solaris™ 10, Linux® and Windows® support • Up to 30G shock

TC2D64™ (Intel Core 2 Duo-based VME SBC) • 1.5 GHz and 2.16 GHz Intel® Core™ 2 Duo Processors • Up to 4 GB ECC SDRAM Memory • CompactFlash When your mission-critical • Two Gigabit Ethernet ports • Two SATA ports applications require high • Up to four PMC slots performance, turn to Themis SBCs. • On-board graphics controller • Four USB and four serial ports In mission critical applications, there’s no • Solaris 10, Linux and Windows support • Up to 30G shock substitute for high performance. The Themis

family of single board computers includes Quad-Core Intel Xeon with the Intel 5100 MCH San Clemente chipset, also Penryn compatible, in addition to our leading UltraSPARC® products on VME and CompactPCI. So we can support applications in Solaris, Windows, Linux and UNIX®.

All Themis products offer maximum configuration flexibility and life cycle support for your technology refresh cycle process, reducing your Total Cost of Ownership. So when mission success depends on higher performance, you can rely on Themis. Across the board. (510) 252-0870

Transformational. © 2008. Themis Computer, Themis, Themis logo, TC2D64 and XV1 are trademarks or registered trademarks of Themis Computer. All other trademarks are property of their respective owners.


insight High Reliability

Increasing Reliability through Automated Enforcement of Design Rules Advanced static-analysis tools are good at finding serious programming errors. They can be customized to find violations of domain or projectspecific rules faster and more efficiently, but also to enforce compliance with coding guidelines and quality standards. by Paul Anderson, Ph.D., GrammaTech


n the ongoing effort to improve software reliability, advanced static-analysis tools have proven to be effective at finding serious programming errors. These tools use sophisticated symbolic execution techniques to detect bugs such as buffer overruns, null pointer dereferences, race conditions and many others. They excel at finding generic flaws—places where the fundamental rules of the language are violated, such as division by zero errors. They are also good at locating defects that arise because a common library is being used incorrectly, such as a memory or other resource leak. These tools can also be extended to find domain-specific defects. There are many benefits to customizing these tools—improving reliability by finding defects unique to the code base is the main advantage, but it can also reduce the cost of manual code inspections, and help reduce compliance costs by checking adherence to coding standards. Almost all projects have rules about how code should be written, even if they are not explicitly stated. These rules may cover superficial aspects of the code, such as layout, identifier naming and syntactic



Source Code

Model Extraction

Names Database/Symbol table Name copy_item item_cache color header.h

Intermediate Representations (IR)

Abstract Syntax Tree (AST)


Control Flow Graph (CFG)


Call Graph

Kind Location function item.c:25 variable item.c:10 parameter pallette.c:23 file shapes.c

Figure 1 The architecture of an advanced static-analysis tool.

restrictions (e.g., no expressions nested more than four levels deep). They may also address semantic properties of the code such as the order in which it is legal to perform certain operations, or how to handle the detection and logging of run-time errors. Although there are lightweight tools (such as lint) for detecting violations of the superficial rules, only

the advanced static-analysis tools can effectively find violations of the complex semantics-oriented rules. Often, a domain-specific rule is just another instantiation of a rule already checked by a static-analysis tool. Typically, tools can be easily configured to detect violations of these. For example, leaks of project-specific resources can be detected by

industry insight

configuring the tool to recognize the functions that allocate the resource and those that release it. However, a project-specific design rule can often be sufficiently different that a checker for it can only be written by programming a special-purpose extension to the tool. Whether it is possible to do so depends on the nature of the rule itself, and the degree of extensibility that the tool provides. Before explaining how this can be done, it is worthwhile to briefly describe how advanced static-analysis tools work.

Advanced Static-Analysis Tools

Static-analysis tools work very much like compilers. They take source code as input, which they then parse and convert to an intermediate representation (IR). The IR typically comprises the program’s abstract syntax tree (AST), the symbol table, the control-flow graph (CFG) and the call graph. Figure 1 shows a block diagram of the process. A compiler uses these to generate object code and then discards them. However, staticanalysis tools retain the IR, and checkers are usually implemented by traversing or querying the IR looking for particular properties or patterns that indicate defects. A checker looking for a simple syntactic property (e.g., goto statements) would search the abstract syntax tree for constructs that match that pattern. The advanced tools get their power from sophisticated symbolic execution techniques that explore paths through the control-flow graph. These algorithms keep track of the ab-


Untitled-3 1


stract state of the program and know how to use that state to exclude consideration of infeasible paths. This level of complexity is required in order to find the serious bugs while keeping the level of false positives low.

Custom Checkers

No two tools provide exactly the same interface or techniques for implementing custom checkers, but two mechanisms are common. The first allows users to essentially annotate their code. The second is an API that gives access to some aspects of the IR, typically using a visitor pattern. We explain the first mechanism using examples written for CodeSonar. Other tools have analogous mechanisms. Suppose, for example, that there is an internal function named foo that takes a single integer parameter, and that it is dangerous for that parameter to have the value -1. A check for this case could be implemented by adding some code to the body of foo as shown in Code Block 1.

The #ifdef construct ensures that this new code is not seen by the regular compiler. The call to csonar_trigger is never actually executed, but is instead analyzed by the tool. If the analysis considers that the trigger condition may be satisifed, then it will issue a warning with the given message. Of course in most cases it is not appropriate to require that programmers interleave these annotations with the code, so there is an alternative way to implement this kind of check that allows it to be written in a separate file, which avoids the need for the code to be edited. This approach would also be useful when the source code for foo is not available, such as when it is in a third-party library. To do this, the author of the checker would write a replacement function as shown in Code Block 2. void csonar_replace_foo(int x) { csonar_trigger(x, “==”, -1,

void foo(int x)

“Dangerous call to foo()”);



#ifdef __CSURF__ csonar_trigger(x, “==”, -1, “Dangerous call to foo()”); #endif __CSURF__ … }

Code Block 1


Code Block 2

When the analysis sees the definition of csonar_replace_ foo, it treats all calls in the code to foo (except the one inside csonar_replace_ foo) as if they were calls to csonar_replace_ foo instead.

11/11/09 3:45:15 PM

industry insight

CompactPCI Plus ®

This trigger idiom is good for checking temporal properties, particularly sequences of function calls. Suppose there is a rule that says that bar should never be called while foo is executing. A check might be implemented as shown in Code Block 3. static int foo_is_executing = 0; void csonar_replace_foo(int x) { foo_is_executing = 1; foo(); foo_is_executing = 0; } void csonar_replace_bar(void) { csonar_trigger(foo_is_executing, “==”, 1, “Call to bar from foo”); bar(); }

Code Block 3

Note that a global state variable is used to record whether or not foo is active. Before entry to foo, it is set to one and then reset to zero after foo returns. This variable is then checked in the trigger in bar, and if set to one, a warning will be issued. The example in code block 3 shows how a global property might be checked. The same mechanism can be used to write checks for properties of objects. For example, it is illegal to read from a file object after it has been closed. In such a case the state being checked must be associated with the file object. The extension API allows users to specify attributes that can be attached to objects. These can be thought of as state variables that can be associated with objects. An attribute can be used to encode whether a file object is open or closed, and the checker for a reader function can test the value of that attribute and issue a warning if the file is in fact closed. This approach allows users to author static checks almost as if they were writing dynamic checks. The API for this kind of check is small, and the language is regular C, so there is a shallow learning curve. This simplicity is deceptive because the technique can be used to implement fairly sophisticated checks. In CodeSonar, many of the out-of-the-box checks are implemented this way.

The second way to implement a custom check is to use the API that gives access to the underlying IR through a visitor pattern. A visitor is a function that is invoked on every IR element of the appropriate type; there are different visitor types for different IR constructs. When they are present, visitors piggyback on the various traversals carried out by the analysis. Visitors can be defined for files, identifiers, subprograms and nodes in the controlflow graph (which correspond roughly to program statements), and the syntax tree. The interface to a visitor allows for pattern matching against the construct. This way a checker author can easily search for constructs without having to know exactly how they are represented. For example, suppose there is a rule that no variable is allowed to contain uppercase characters. The checker for this would be best implemented by writing a function that takes an identifier as input, checks that it represents a variable, scans it for uppercase letters, and issues a warning if one is found. This function would be registered as a visitor for the table of identifiers. CodeSonar has an additional kind of visitor, which is invoked during the path exploration of the control-flow graph. At each step along the path, the check can query the abstract state of the program, so the implementation can ask questions such as “what is the value of this variable at this point?” This allows the checker author to write sophisticated checks that leverage the built-in program analysis capabilities of the tool.

The future of CompactPCI is serial...


F19P – 3U CompactPCI® PlusIO SBC with Intel® CoreTM 2 Duo

MEN Micro leads the way again with advanced serial I/O to PICMG’s newest specs:

CompactPCI® PlusIO PICMG 2.30 ■ ■ ■ ■

100% compatible with parallel CompactPCI® PCI Express®, Ethernet, SATA, USB Support of 4 peripheral slots Fast 5 Gb/s connector

CompactPCI®Plus PICMG CPLUS.0 ■ ■ ■ ■ ■ ■

GrammaTech. Ithaca, NY. (888) 695-2668. [].

Star architecture Full Ethernet mesh No bridges, no switches Support of 8 peripheral slots Fast 12 Gb/s connector Proposed CPLUS.0 CompactPCI® Plus specification currently under development

Count on MEN Micro to get you to the future of harsh, mobile and mission-critical embedded technology first!

MEN Micro, Inc. 24 North Main Street Ambler, PA 19002 Tel: 215.542.9575 E-mail:

Untitled-11 1



9/17/09 3:29:03 PM

System Integration

Known formally as IEEE 802.3 or 802.3af, PoE can supply up to 15.4 watts of power to remote devices, although devices must be designed with the assumption that only 12.95 watts are available due to cable losses. With a communications heritage, typically 48V DC is supplied at 350 mA. Some early Atom offerings are targeting 802.3at, dubbed “PoE Plus,” to extend the power available to 25W allowing for remote peripherals, storage devices, LCDs and backlights to be powered as well. But advances in backlight technology, coupled with a renaissance of standardized lowpower x86 single-board controllers, are A low-profile x86 solution combines the lowest-power paving the way to the ideal solution for Atom Z-series and chipset, LED backlight and ISM 12.95W PoE applications. Industry Standard Module (ISM) is a form factor for a complete embedded computer with pure 3.6 x 3.8” (90 x 96 mm) form factor peripherals and SSD within a 4” square footprint and only that addresses shortcomings of existing 12.95 watt power budget. small form factors by: • Fitting x86 circuitry well within the board outline • Freeing designers from restrictive by Chris Lane, Jaco Electronics top- and bottom-side component Kelly Gillilan, Adlink Technology height zones • Decoupling expansion interfaces from form factor definitions • Re-capturing the space wasted by EMs who build the tiniest, lowest-power devices tend to large legacy parallel bus connectors use RISC processors and microcontrollers at the heart of • Enabling flexible I/O connectorization their systems. While the x86 architecture can reduce de• Allowing bus combinations that were previously undefined velopment costs and risk due to familiar programming, perforand unnamed mance and off-the-shelf boards ecosystem, the size and power consumption usually squash any such considerations. A solution SBC manufacturers have struggled to fit modern legacy-free is needed for those who want the best of both worlds. platforms onto small form factors without protruding beyond the While legacy EBX and PC/104 form factor SBCs continue allowed board outlines with “wings.” Now that the ISA bus, setheir acceptance in truly rugged applications, there is a new sher- rial ports and even the PCI bus are no longer integrated into lowiff in town: ISM. An acronym for Industry Standard Module, this power chipsets, it takes extra circuitry to meet the needs of many new form factor from the Small Form Factor SIG is allowed to embedded applications. Applications like PoE generally don’t be legacy-free and expansion-agnostic. This opens up brand new need legacy buses. markets for embedded x86 designs that are not saddled with the The ISM Specification defines the board size, four fixed added cost, power and board space of parallel bus bridges and mounting holes, component height limits and flexible “expansion old-generation PC peripherals such as PS/2 keyboard and mouse zones” for I/O and/or bus connectors. The fixed corner mountand printer ports. ing holes allow reuse of enclosures without modifications in the future. ISM allows many combinations of bus and I/O connectors PoE Reduces Cabling as long as those interfaces reside within the defined “expansion Previously beyond the reach of high-power x86 processors zones.” ISM also offers a choice of using right-angle connectors and chipsets, Power-over-Ethernet (PoE) has gained traction in that overhang the board edges, or extending the circuit board if VoIP phones, wireless access points, LAN switches and factory vertical / non-overhang connectors are used. automation applications where costs and convenience are improved without separate power wiring and wall power transform- Legacy-Free ISM CPU ers. Innovative ultra-low-power processors like Intel’s Atom, Figure 1 shows the Ampro by Adlink CoreModule 730. Ofcombined with the 3.6 x 3.8” ISM form factor, now make it pos- fering a choice of Atom Z510 processor at 1.1 GHz or Atom Z530 sible for PoE to enter the broad space of embedded computing. processor at 1.6 GHz, the CoreModule 730 provides the basics Possibilities are endless in this new frontier. of an embedded controller in a small ISM size without the over-

Advances in Small Form Factors

Tiny Atom SBC Targets Power-over-Ethernet Apps




system integration

Figure 1 The Adlink CoreModule 730 is an example of an Atom Z Series-based ISM module that consumes only 3.6 watts.

kill of larger motherboards like Mini-ITX. The single-chip Intel US15W chipset provides modest expansion buses and I/O sufficient for controller applications. The board’s feature set includes a Gigabit Ethernet port, several USB ports, eight GPIO pins and a CompactFlash socket on the bottom of the board. Since space is at a premium, pin headers are used for flexible cabling. Only the I/O used by an application is cabled out to the enclosure, and in the locations where desired. SBCs with fixed I/O blocks are convenient but inflexible. Computer-on-Module architectures require custom carrier boards to be designed in order to bring out the I/O, consuming development budgets and extending project risk and schedule. Solid-state drives (SSDs) are gaining in popularity for mass storage. They offer small size, low power, lower MTBF and no spin-up time compared to rotating disk drives. Windows XP Embedded brings the familiar desktop user interface in a modular footprint (image size)—even down to 1 Gbyte and below. For many applications, a 1 Gbyte CompactFlash device is very mainstream and cost-effective, well below the steep part of the flash price curve. A short copper plate with fins spreads the modest heat from the Atom processor and single-chip US15W chipset over the full breadth of the board. The SODIMM RAM memory is also cooled by this heatsink. The thermal solution has a low thermal resistance with minimal airflow inside the system enclosure, allowing operation to +70°C air temperature just above the heatsink. The heatsink can be attached to the outer enclosure in places, further removing heat from the tiny SBC (Figure 2). At the front of Figure 2 are pin headers for two USB ports and one Gigabit LAN port, as well as an IDE connector that can be used in other applications. These connectors are not covered by the copper heatsink so they are easily cabled to the outside of the enclosure.

PoE System Solution

The Jaco Electronics JDS104CA01 (Figure 3) is a new x86based PoE solution with full LCD and touch screen capability. Designed as a wall-mounted user interface to a professional au-

dio control system, the unit includes: 10.4” SVGA LCD, resistive touch screen and 1.6 GHz Atom CPU powered Ampro by an Adlink CoreModule 730. This panel PC system uses only 11 watts, landing safely under the IEEE 802.3 PoE power requirements. The combination of x86 and ultra-low power is the best of both worlds that many have been waiting for. JDS104CA01’s configuration creates a reference design that is suitable for application in industrial, medical and other arenas using LCD sizes up to 12.1” diagonal. Many of today’s panel PCs strive for a fanless, low-powered, high-performance product. While these panel PCs may be low powered compared to some standards, they do not come close to meeting the 12.95 watt PoE requirement. Table 1 compares a typical low-powered panel PC to JDS104CA01. A low-powered SBC is only the first step to designing a highperformance system that is PoE-compliant. As one can see from Table 1, an LED backlit LCD also plays a major role in power savings. The emergence of LEDs in LCD backlighting has allowed for numerous advantages for today’s applications. New high-efficiency LEDs have allowed displays to use less power for the same luminance. For example, an AUO G104SN02 V2 10.4” display with 400 nits of luminance uses 5.2 watts of total power. This is 32% less power as compared to similar displays utilizing CCFL backlights.

Figure 2 The thin heat sink on the ISM board draws heat off the major components and can also be attached to the chassis, allowing operation in +70° C ambient air temperature.

The power savings due to LED backlights does not stop there; LED backlit panels require even less power due to the differences in drive requirements compared to their CCFL rivals. The change from a DC-AC inverter to an LED driver with a constant current circuit allows for a simpler, more efficient solution that does not inherit the EMI problems that accompany CCFL inverters. In fact, many of the latest LED backlit LCDs embed these constant current circuits; this eliminates one more part in the design and simplifies the total solution. The AUO G104SN02 V2 is one example of an LCD with a built-in LED driver circuit. This driver circuit is included in the 5.2W of total power from the LCD and allows the system to eliminate an additional 1.2W that is lost by the inefficiencies of the CCFL inverter. If you add RTC MAGAZINE NOVEMBER 2009


system integration

Typical Solution

Power Consumption


Power Consumption

10.3 Watts

ADLINK CoreModule 730, Z530 1.6 GHz Atom CPU

3.6 Watts

AUO 10.4� SVGA 400 nit LCD w/ CCFL backlights and Inverter

7.7 Watts

AUO 10.4� SVGA 400 Nit LCD w/ LED Backlights and LED driver

5.2 Watts

Resistive Touch Screen and Controller

0.2 Watts

Resistive Touch Screen and Controller

0.2 Watts

PoE Power DC-DC Conversion Board.

1.9 Watts

Total Power

10.9 Watts

3.5� SBC, N270 1.6 GHz Atom CPU

Total Power

Figure 3 This 10.4-inch LCD display from Jaco Electronics including the ISM module, the display and its LED controllers consumes a total of 10.9 watts­—well within the limits of power-over Ethernet.

up the total power savings in a solution with LED backlights, you can save your system about 2.5 watts. A PoE power conversion board is the next essential part of this system design. The power conversion board connects a power-injected Ethernet cable and separates the power and communication signals. The board converts the 48 volt input to the 5V and 12V levels that the SBC and LCD need for operation. The

Conduction Cooled VME Solid State Disk Phoenix International’s VC1-250-SSD Conduction Cooled Serial ATA (SATA) based Solid State Disk VME blade delivers high capacity, high performance data storage for military, and y, aerospace p industrial applications requiring rugged, extreme emee envi eenvironmental i ron ronmen me tal and secure mass data storage.

LLow ow w Ope OOperational Op per pperational e r aatio era tio ti ioo nal al TTemperature Te Temp em emp mperat mp pper e rraatu er ature tur tu urre re --4 -40 -40° 40° C


High Operational Hi Temperature +85° C

Operational Altitude to 80,000 feet

An ISO 9001: 2000 CertiďŹ ed Service Disabled Veteran Owned Small Business



Table 1 Power consumption comparison between earlier Atombased solution for the LCD display and the solution using the lower-powered Z-Series and an LED-backlit display.

PoE power board is designed to receive an input of up to 12.95W and supply an output of up to 11.00W. The CAT5 cable entering the unit is the only external cable needed to run the system, making for a clean and elegant solution. The PoE board used in the 10.4� design is 85% efficient and has a PC/104 mounting scheme that can stack on top of the ISM Atom board. Although all low-power components are selected, heat dissipation is still an important design consideration in such an application. Many PoE applications require fanless solutions in sealed enclosures that do not allow for airflow through the unit. Heat sinking and proper enclosure design become critical to the longterm reliability of the product. The Atom-based SBC is designed to use the outside enclosure to help dissipate the heat generated from the SBC. The low-power and customizable solution also allows for the system to be designed to withstand harsh environments or meet many unique enclosure requirements. Although not essential, touch screens often complement many of the current PoE applications in today’s market. Whether it is resistive, capacitive, projected capacitive or any other touch screen technology, touch screens provide a convenient user interface to operate interactive applications. The touch screen also allows for a sleek look by having no buttons on the faceplate, one single in/out wire and no mouse or keyboard. Jaco Electronics (877) 373-5226. Hauppauge, NY. []. ADLINK Technology San Jose, CA. (408) 360-0200. [].


Untitled-6 1

18.2 Watts

10/16/09 11:43:57 AM




Concept Development, Inc.

A Perfect Fit Your Vision — Our Capabilities

Homeland Security

Computer Systems Design

Medical Equipment

Concept Development, Inc. was established to assist companies in meeting their needs for timely and cost effective product development. We accomplish this by collaborating with the customer’s existing engineering group to effectively create an extension of the customer’s engineering resources. CDI has established itself as a flexible and responsive partner for many companies in a wide range of product technologies for an expanding set of markets.

In-Flight Entertainment

Aerospace Applications

Glass Cockpit Scanning & Analysis

A Complete Development and Support Process Define the Requirements

Engineering and Design

Production Phase












Please Call Us 949.623.8025

1881 Langley Ave. | Irvine, CA 92614 | 949-623.8000 | f 949.623.8055 |

System Integration Advances in Small Form Factors

Embedded Form Factors Harness Emerging Technologies to Enable Wireless Systems Developments in low-power processors, highly integrated peripherals and new interconnects are shrinking the size and power consumption of modules to allow their deployment in wireless network applications.

by Jason Krueger, VersaLogic


ireless connectivity is merging with technological advancements in silicon, signaling, mass storage and software to meet the high-performance, ultra-low-power requirements for next-generation wireless systems. Embedded form factors, seeking to utilize these developing technologies to the best advantage of system designers, continue to evolve by facilitating emerging standards and providing enhanced capabilities while simultaneously reducing form factor footprints. As wireless connectivity becomes increasingly ubiquitous, the volume of embedded systems that utilize wireless is expanding, thereby perpetuating the demand for higher processing power with minimal power draw and size. Embedded computing systems have long been tethered by wires to a stable infrastructure providing myriad connectivity options and a continuous source of power. For applications where the data is beyond the infrastructure, the solution has often been to build the infrastructure out to the data. Though this approach is feasible in locations where a wired infrastructure exists and the distance to be wired is relatively short, wired connectivity quickly becomes prohibitively expensive as the distance to be wired increases; furthermore, building out infrastructure is an



entirely improbable solution in situations where there is no established infrastructure to begin with. The solution to the limitations imposed by wires is simple—remove the wires. Though this dream of wireless connectivity has existed since the first network packet was sent over a wire, the technology required to facilitate reliable, high-bandwidth, secure data transmission has only recently emerged as a truly viable solution. The continuing evolution and adoption of wireless protocols, including Wi-Fi, WiMax, CDMA, UMTS, LTE, Zigbee and Bluetooth, along with free worldwide access to Global Positioning System (GPS) data, has unleashed an explosion of wireless consumer devices, including smartphones, mobile Internet devices (MIDs), e-books, portable media players and navigation systems. The success of these products has proven the potential for wireless applications while simultaneously fueling the expansion of the infrastructure necessary to sustain them. This has in turn opened the door for sophisticated wireless applications in the embedded systems space that take advantage of increased processing power and expanded system capabilities to revolutionize both existing and emerging embedded applications.

Design Restrictions

The freedom afforded by wireless carries with it inherent design restrictions. Lacking a constant supply of wired grid power, wireless systems typically rely on batteries. Faced with a scarcity of power, stringent control of power consumption is no longer an option but a necessity. Size is also a major design factor, as smaller sizes afford greater mobility. Support for standard interfaces and peripherals is also required to facilitate design implementation, speed time-to-market and keep system costs to a minimum. Wireless systems must also communicate over standard networking protocols and utilize advanced data compression and security functions to minimize bandwidth while maximizing data integrity and security. Wireless systems are often deployed in demanding environments subject to extreme physical and environmental stresses, which necessitates the need for ruggedized solutions with unquestioned reliability, especially in the case of remote applications. RISC architectures, including ARM and PowerPC, have been the preferred choice for wireless embedded devices due to their ultra-low power consumption and superior performance-tocost ratio in the light of application requirements and production

system integration

volume. Success naturally breeds competition. While RISC technology continues to drive incremental increases in processor performance, CISC architectures (primarily x86), have been aggressively moving into the wireless embedded systems space through radical decreases in power consumption and package size. Intel and VIA have been aggressively transitioning x86 technology into the wireless embedded systems space. Initially targeted at the netbook market, then at even smaller, lower-power nettop applications, x86 technology continues to find its way into smaller and smaller embedded applications. The second-generation Atom Z5xx processor series (Figure 1) illustrates Intel’s commitment to the exacting needs of nextgeneration embedded systems. Optimized for low power, thermally constrained, fanless, small form factor (SFF) solutions, the Atom Z5xx series offers system designers a true embedded processor solution with processing performance up to 1.6 GHz, extended temperature operation (-40° to +85°C) and long life-cycle manufacturing support. The Atom Z5xx, along with the System Controller Hub (SCH) US15WP(T), provide the processing performance required to deliver next-generation wireless applications incorporating advanced graphics, display, video and audio capabilities, while support for PCI Express (PCIe), USB 2.0, SMBus, I2C, LPC, IDE (PATA) and GPIO provide a wealth of I/O options to support diverse system applications. Advanced power management capabilities enable power to be removed from the processor core and caches while minimizing leakage to significantly reduce idle power. Intel is pushing deeper into the wireless embedded systems space as it transitions its System on Chip (SoC) designs to an Atom processor core, providing designers with a broad array of purpose-built processors that share a common core, yet have differentiating features that can be mapped to system requirements for speed, power and temperature range. SoCs further minimize space by combining typical two- and three-chip solutions into a single integrated chip. Originally targeted at MIDs, the addition of Intel QuickAssist Technology will appeal to wireless applications through the acceleration of cryptographic and packet processing to facilitate virtual private network (VPN) gateways, firewalls and Unified Threat Management (UTM). Further size and power reductions to the x86 architecture will continue to be achieved by building on the success of 45nm process technology, with the promise of 32nm process technology looming just over the horizon.

including PCI Express (PCIe), ExpressCard, USB 2.0, SMBus, I2C and LPC, as well as control and power management signals. Additionally, interconnects must be ruggedized for high shock and vibration environments. Smaller connector heights enable denser designs, further reducing the overall system volume. Finally, interconnects must be low cost in order to be viable for original equipment manufacturers (OEMs) in production quantities. Several new interconnect standards have emerged to meet these demands for wireless embedded systems. Some connectors, such as MXM, have been adapted to the embedded space after having proven themselves in the notebook sector. Others, such as Stackable Unified Interconnect Technology (SUMIT) from the Small Form Factor Special Interest Group (SFF-SIG), have been built from the ground up to meet the exacting needs of embedded systems development. Wireless embedded systems typically use a solid-state drive (SSD) for mass storage. Compared to hard disk drives, SSDs offer substantial power savings. Furthermore, as SSDs have no moving parts, they are less susceptible to the effects of shock and vibration. Secure latching connectors further increase reliability in mission-critical applications. SSD capacities are continually expanding, while extended temperature and conformal coating options are emerging to satisfy the needs of demanding environments. SSDs can be interfaced via parallel ATA (PATA), serial ATA (SATA), USB or PCIe, offering system designers the ability to target their system throughput requirements. CompactFlash, which has long been the standard for embedded systems, is now beginning to be displaced by these new offerings, including Disk on Module, eUSB, SSDDR and the new MiniBlade specification from the Small Form Factor-SIG.





Interconnects and Storage

As silicon technology advances to increase the amount of processing power available to embedded systems, so does the sheer amount of data that needs to flow in and out of these systems. Multiple simultaneous inputs (GPS, video cameras, microphones, sensors, accelerometers, etc.) and outputs (wireless data transmission), along with a rich user interface (high-definition video and audio), combine to put demanding I/O throughput requirements on wireless embedded systems. To meet these needs, advanced I/O connectors are emerging that offer an array of both high- and low-speed signals to meet the diverse needs of wireless embedded systems. Robust signal support is required, typically

Figure 1 Relative size comparison of COM (yellow) and mainboard (blue) form factors. [a] nanoETXexpress [b] QSeven [c] Pico-ITXe [d] SUMIT-ISM



system integration

ing that takes advantage of advancements in power efficiency, thermal management, and feature integration to enable ultracompact, stackable board solutions. The Pico-ITXe specification, originally proposed by VIA, has been adopted by the SFF-SIG specifically for the development of energy-efficient, fanless, x86 systems with a complete range of standard signaling on a 100 mm x 72 mm form factor. Pico-ITXe (and its complementary Pico-I/O specification) utilizes the SUMIT interface while also providing a standardized, modular, SFF expansion specification. The SFF-SIG is also bringing the benefits of SUMIT to the industry-standard 90 mm x 96 mm footprint popularized by PC/104, thereby opening the door to high-speed signaling via PCIe and SATA, with the added versatility of ExpressCard, while retaining legacy signal support. A size comparison of several major form factors is shown in Figure 2.


Figure 2 SUMIT-ISM SBC featuring Intel Atom Z530P processor, US15WP chipset, SUMIT-AB connector pair, and an IDE Disk on Module (DOM) socket.

Form Factors

Evolving form factors are synthesizing the hardware requirements for wireless embedded systems into a standardized development platform, thereby enabling innovation, reducing risk and speeding time-to-market. Taking advantage of reduced processor package sizes, form factors continue to decrease in size, enabling increasingly smaller embedded systems for wireless applications. Computer-on-Modules (COMs) allow a modular systems approach that decouples the processor complex from the remainder of the system, thus alleviating system obsolescence as imposed by Moore’s law. The benefits of a modular approach have led to a proliferation of specifications. The nanoETXexpress specification is a variant of the COM Express standard PICMG, which adds support for PCIe, Gigabit Ethernet (GbE), USB 2.0 and SATA while retaining legacy signal support on a reduced 55 mm x 84 mm form factor utilizing a standard COM Express Type 1 connector. Another variant of the COM approach is the Qseven specification from the Qseven Consortium, which utilizes standard high-speed MXM connectors to provide high-speed serial signals on a 70 mm x 70 mm form factor. The SFF-SIG, on the other hand, has taken a fresh approach to COM development with the introduction of its Computer on Module Interconnect Technology (COMIT) specification. Rather than focusing on a particular form factor, COMIT is a modular, processor-independent connector system that supports both emerging and legacy signaling common to modern low-power chipsets by way of a rugged, highdensity 240-pin connector pair that meets the bandwidth requirements for next-generation PCIe 2.0 and USB 3.0 signaling. A new generation of mainboard form factors is also emerg-



The critical role that software plays in the control of power consumption cannot be understated. The Advance Configuration and Power Interface (ACPI) specification is an open industry standard that provides industry-standard interfaces for application control of hardware sleep, performance and throttling states. By proactively controlling the sleep states of hardware via software, system developers are able to reduce power consumption to a fraction of typical system levels. Because entering and exiting deeper power-saving states increases system latency, the ability to invoke the full range of both processor and system sleep states afforded by emerging processors enables power benefits to be adaptively fine-tuned to application performance as needed. ACPI also enables systems to control device states for powering down auxiliary peripherals when idle, further reducing system power consumption for maximum efficiency and battery life. The open source community, through projects such as Mobile Linux and LessWatts, is contributing best practices and creating standards to realize even greater power savings and accelerate wireless embedded development. As wireless connectivity proliferates and the processing power available to SFF embedded systems increases, new opportunities are emerging to deliver innovative solutions that harness the true potential of wireless applications. To facilitate these solutions, embedded form factors are evolving to provide open platforms that meet the demanding needs of wireless applications while also providing the flexibility required to adapt to a diverse array of solution possibilities (Table 1). The future of wireless is at hand—all that remains is to realize its full potential for the benefit of humankind. VersaLogic Eugene, OR. (541) 485-8575. [].

Data Acquisition Boards Showcase Featuring the latest in Data Acquisition Boards technologies

PCIe-DIO-48S: PCI Express 48-Channel Digital I/O Card with Change of State Detection 48 or 24 channel high-current TTL digital I/O lines Change of State (COS) detection and interrupt capabilities Compatible with industry standard 8255 PPI All 48 digital I/O lines buffered with 32mA source / 64mA sink current capabilities 5V VCCIO (3.3V optional) available on each I/O header Molex PC-style connector for maximum ACCES I/O Products, Inc. 5V VCCIO current sourcing capability E-mail: Phone: (858) 550-9559 Web: Fax: (858) 550-7322

PCIe-COM232-8: PCI Express Eight-Port Low Profile RS-232 Serial Communication Card

ACCES I/O Products, Inc.

Eight-port PCI Express RS-232 serial communication Low profile PCIe MD1 form factor, MD1 defines the shortest standard card length available High performance 16C950 class UARTs with 128-byte FIFO Supports data communication speeds up to 921.6kbps +/- 15kV ESD protection on all signal pins

Phone: (858) 550-9559 Fax: (858) 550-7322

E-mail: Web:

Setting the Standard for Digital Signal Processing

Model 7158 – PMC/XMC Software Radio Transceiver with Two Virtex-5 FPGAs

Pentek, Inc. Phone: (201) 818-5900 Fax: (201) 818-5904

Two 500 MHz 12-bit A/Ds Digital Upconverter with two 800 MHz, 16-bit D/As Two Virtex-5 FPGAs with XMC gigabit serial I/O Built-in clock synthesizer with low phase noise Clock/sync bus for multimodule synchronization Factory-installed cores for complete SDR interface solutions Also available in cPCI, PCI, and PCIe formats E-mail: Web:

Mini PCI Type IIIB - 1553 Controller Controller of dual redundant (A/B channel) 1553 communications 32-Bit/33MHz PCI Interface DDC Chip BU-65864 Programmable Bus Controller, Remote Terminal, or Monitor Terminal modes Multiprotocol Support of MILSTD01553A/B Notice 2 and STANAG 3838 +3.3V Operation and Logic Long or short stub support Low power consumption On-chip transceivers

ALPHI Technology Corp. Phone: (480) 838-2428 Fax: (480) 838-4477

PC/104 Dial-Up Modems

Radicom Research, Inc. Phone: (408) 383-9006 Fax: (408) 383-9007

Channel Express Family x8 PCIe/XMC FPGA Configurabale ADC/DAC

Red Rapids Phone: (972) 671-9570 Fax: (972) 671-9572

Dual 14-bit 400 MHz ADC/DAC Dual 16-bit 160 MHz ADC/DAC Quad 14-bit 250 MHz ADC Dual 8-bit 1.5 GHz ADC Xilinx Virtex-5 LXT or SXT FPGA Optional 32 Mbytes of QDR SRAM Optional 128 Mbytes of SDRAM Includes FPGA reference design with source code Prices start at $4,730 (LX50T-2, no memory) E-mail: Web:

E-mail: Web:

2-wire leased-line and dial-up connections -40C to +85C operating temperature Size: 3.77” x 3.55” x 0. 57” up to 56K bps data rate 14.4K bps fax, voice Speakerphone & microphone input AT command DTMF, ring and Caller ID detection Transferable FCC68, CS03, CTR21 telecom certifications c/UL approved CE marking E-mail: Web:


Themis Computer Phone: (510) 252-0870 Fax: (510) 490-5529

Ideal for compute-intensive, multithreaded applications Sun UltraSPARC six or eight core T2 processor Up to 32GB ECC DDR2 FBDIMM Memory Three 10 Gigabit Ethernet Ports, two on VPX J1 CPU independent Board Management Controller On-board 1.8 in SATA HDD support Built in Hypervisor with LDOM Support Shock: 35G peak, 20ms OS Support: Solaris 10 E-mail: Web:

products &

TECHNOLOGY FEATURED PRODUCTS “Energy Optimizer” Lets Data Centers Conserve Energy While Meeting Efficiency Goals A wireless monitoring system provides continually updated information on a data center's electrical usage and thermal status, giving users the precise knowledge to take energy-conservation measures while maximizing the operational efficiency and reliability of their servers and other computing equipment. The Arch Rock Energy Optimizer-Data Center Edition (AREO-DC) measures both power and cooling system efficiency in the data center using non-disruptive Internet Protocol (IP)-based wireless sensor networks. It lets data center managers: identify the factors that cause energy waste (e.g., short-cycling, missing blanking panels, sub-floor obstructions) or "hot spots" and other harmful conditions, and correct these problems to reduce existing power consumption or allow the facility to support more equipment. They can then optimize available capacity for new servers based on the actual electrical/cooling consumption of equipment, rather than the overstated "nameplate" ratings. AREO-DC works by deploying wireless sensors to measure electrical, thermal, flow and pressure conditions on power circuits, server racks, computer-room air conditioners (CRACs) or air handlers (CRAHs), chillers and underneath the raised computer-room floor. The sensed data is then transmitted via wireless sensor networks to a graphical, multiwindow dashboard that shows the electricity load (and associated utility rate-adjusted spend rate) of various equipment, electricity usage by physical or functional area over user-selected time intervals; temperature and humidity data from CRACs, CRAHs, server racks and chillers over time; chiller water-flow rates; "heat maps" superimposed on a floor plan; and key performance indicators such as the Green Grid organization's Power Usage Effectiveness (PUE) Level 3 standard. From the AREO-DC dashboard, users can drill down to specific data centers and specific racks within a data center, and bring up side-by-side views of various factors, such as energy usage vis-à-vis indoor and/or out-



door temperature. Alerts can be generated when heat- and energy-use thresholds or userdefined financial thresholds are exceeded. AREO-DC can help data center managers, for example, pinpoint improper mixing of hot and cold air such as conditions that may lead to insufficient cold air pressure at rack inlets, top-of-rack hot spots, or cooling-plant overcompensation. It can verify the existence of short-cycling or conditions that cause suboptimal CRAC/CRAH operation. In addition, it can obtain an accurate PUE figure in mixeduse environments by measuring the cooling load of the data center relative to that of the entire facility, and combining this with measurement of the central cooling plant's electrical consumption. And it can measure outside air temperature and humidity to identify air or water free-cooling opportunities through the use of HVAC economizers or cooling towers rather than costly use of compressors and refrigerants. Data from the various sensor nodes is

processed and displayed by the Energy Portal, which is a Web-based application that displays detailed energy and thermal usage data in graphical and tabular formats. All wireless nodes are battery-operated with an optional AC wall adapter, and support over-the-air (OTA) software upgrades. An example starter configuration including the Energy Portal software (one-year subscription to hosted service), one PhyNet Router, three IPpower Nodes (monitoring up to nine circuits), four IPthermal-XT nodes (up to 28 temperature measurement points and four humidity points), one IPthermal-HT node (up to four combined thermal and humidity measurement points), two IPpressure nodes (up to six differential pressure measurement points), and two IPrelay (extender) nodes, is priced at $9,995 (U.S. list). Arch Rock, San Francisco, CA. (415) 692-0828. [].


Atom-Based Series of Fanless, Cable-Free Rugged I/O Platforms

A series of rugged and fanless I/O platforms based on the Intel Atom N270 are targeted for providing rugged I/O solutions to the market. The Matrix MXE-1000 and MXC-2000 from Adlink Technology are the result of combining Adlink’s experience in x86 platform design, versatile I/O function development and thermal design to push fanless systems to a higher standard, including a -20° to 70°C temperature range, 5G vibration and cable-free durable structure. The MXE-1000 and MXC-2000 series are also the first to integrate Adlink’s current I/O design into a computer platform for many application-specific uses. For instance, the compact design and comprehensive I/O of the MXE-1000 make it suitable for intelligent transportation, facility management and environment monitoring applications, and the MXC2000—fully compatible with Adlink’s full line of PCI/PCI Express data acquisition, digital I/O, motion control and image capture cards—is suited for industrial automation, factory control and test instrumentation applications. The MXE-1000 and MXC-2000 series include a specifically designed single board computer to fit the respective fanless chassis in which all heat-producing components come in direct contact with the aluminum shell. This allows for the widest operating temperature range among all off-the-shelf fanless computers. To increases reliability and durability, all connectors and components are mounted directly onto the PCB so there is no internal wiring. The design of the MXC-2000 series enables very simple card installation and maintenance compared to other similar offerings on the market. PCI and/ or PCI Express cards can be effortlessly installed or removed simply by loosening one screw. In addition to general I/O connectors such as Gigabit Ethernet (GbE), COM and USB ports, the MXE-1000 series also provides dedicated GbE and 1394b interfaces to support cameras for outdoor or in-vehicle video/imaging applications. The MXC-2000 series includes configurations of two PCI slots or one PCI slot and one PCI Express slot, supporting Adlink’s full lineup of PCI/PCI Express cards, in addition to third-party plug-in cards. The Matrix series is currently available at list prices starting at $550 and $750 for the MXE-1000 and MXC-2000, respectively. ADLINK Technology, San Jose, CA. (408) 360-0200. [].

Robust Dual Gigabit Ethernet Mezzanine Board with Long-Term Availability

A new dual Gigabit Ethernet mezzanine board enables all VPX, VME, CompactPCI and custom boards offering a PMC or XMC slot to be extended by two Gigabit Ethernet interfaces, which can be individually routed to the front or to the backplane. Equipped with the high-end Intel 82571 Dual Channel Ethernet Controller, the XMC-ETH2 mezzanine board from Kontron delivers very high data throughput. The new mezzanine board comes in two versions: air-cooled (Standard Commercial (SA)) or conduction-cooled (Rugged Conduction Cooled Version (RC)), in order to suit the individual needs of different system environments. Kontron’s dedicated long-term supply (LTS) program offers customers 10+ years of guaranteed supply, making the Kontron XMC-ETH2 a suitable Ethernet mezzanine for infrastructure programs. Typical application areas can be found in markets such as medical, transportation, military, aerospace, energy and automation. Because the XMC-ETH2 mezzanine board supports both XMC and PMC extension slots, it can be used universally in different system designs. Designed as a Plug & Play solution, the board automatically selects the data bus according to whether the power supply is present on the P5 connector (x4 PCI Express). Otherwise, it switches to 32-bit PCI 33 MHz or 66 MHz on the PMC connectors. An integrated micro DIP switch allows for manual override of the host bus selection. Another micro DIP switch controls Ethernet channel routing to the front panel (RJ-45) or to the backplane (P4). Each channel can be routed separately. The Kontron XMC-ETH2 dual Gigabit Ethernet mezzanine board supports Linux, VxWorks and Windows XP and is available now: the SA version for air-cooled system designs in a temperature range of 0° to 55°C (VITA 47Class AC1) and the RC version for robust conduction-cooled systems in the extended temperature range of -40° to 85°C (VITA 47-Class CC4). Kontron, Poway, CA. (888)-294-4558. [].

EPIC-Sized SBC Features Core 2 Duo and SUMIT Connector

New performance levels in an EPIC-sized single board computer (SBC) have been achieved with the integration of the new Intel Core2 Duo (model P8400) 45nm processor in the Komodo SBC from VersaLogic. The module performs at 2.27 GHz with midrange power consumption. Komodo also deploys the new Stackable Unified Module Interconnect Technology (SUMIT) interface, making this SBC an appropriate solution for OEM designers in demanding defense, aerospace and medical markets. Komodo fits the industry standard EPIC footprint of 4.5 x 6.5 inches and is one of the first EPIC-sized products to feature the SUMIT expansion scheme developed by the Small Form Factor Special Interest Group (SFF-SIG). The SUMIT interface supports a variety of expansion options, including up to three PCIe x1, four USB, an SPI and an LPC connector. Standard onboard features include Gigabit Ethernet, up to four Gbytes of DDR3 memory, four USB 2.0 ports, two SATA interfaces, two RS-232 COM ports, two RS-232/422/485 COM ports, PS/2 keyboard and mouse, HD audio, as well as dual LVDS and analog VGA support. Featuring the new MiniBlade module, a rugged, removable, solidstate drive (SSD), and the robust MiniBlade socket, which together comprise the SFF -SIG’s MiniBlade specification; the Komodo offers versatile expansion and application storage. As with other VersaLogic embedded computer products, the Komodo features Phoenix Technologies’ Embedded BIOS with OEM enhancements. The field-reprogrammable BIOS supports custom defaults and the Firmbase software developers kit (SDK) for security, remote booting and other application functions. Fully RoHS-compliant, Komodo is compatible with a variety of popular operating systems, including Windows, Windows Embedded, Linux, VxWorks, QNX and others. The Komodo is customizable, even in low OEM quantities. Options include conformal coating, BIOS customization, revision locks, custom labeling, high-G shock and vibration treatment, and custom testing and screening. VersaLogic, Eugene, OR. (541) 485-5712. [].




Wireless Mezzanine Combines Wi-Fi, Zigbee, GPS and Cryptography

A multi-function mezzanine card combines wireless, GPS and cryptography to deliver portable, secure in-the-field wireless connectivity. The lightweight, small form factor XMC-660 from Curtiss-Wright Controls Embedded Computing is a solution for quickly and easily adding high-performance trusted wireless communications to VME, VPX and CompactPCI embedded systems for applications including luggable computers, manpacks and secure laptop computers. Designed for rugged environments, the XMC-660, based on the VITA 42 XMC standard, combines support for WiFi 802.11 n/a/b/g communications, Zigbee 802.15 asset tracking, and GPS location services on a single plug-in mezzanine card, to deliver a solution for systems integrators building embedded wireless networks. Power dissipation for the card is 7 watts (typical)/ 8.4 watts (max). It requires only a 5V power supply from the basecard. All other necessary voltages are generated on board the XMC-660. Although targeted for multi-function wireless applications, the XMC-660 also provides advanced back-end AES/3DES cryptographic support for secure wireline communications over PCI/PCIe to a base card. As a modular building block, the XMC-660 can, for example, be mounted on a Curtiss-Wright Controls’ VPX3-1100 Atomic base card (using an ATOM 1.6 GHz processor), resulting in a 100 mm x 160 mm x 2.54 mm lightweight solution embeddable in a manpack or a ground vehicle as a wireless computing end node. The XMC-660 features a dual band MIMO 2T3R chipset with PCI interface. The chipset consists of two highly integrated RF and baseband/MAC IC modules that are fully compliant with the IEEE 802.11n draft specification, as well as IEEE 802.11 a/b/g standards that operate in the 2.5 GHz and 5 GHz bands. The GPS module is a small high-sensitivity solution comprising a 20 channel GPS receiver (SiRF Star- III) capable of providing accurate location, altitude, velocity and direction data of moving objects. The Zigbee interface is provided by a single chip TI CC2480 processor running Texas Instruments’ Zigbee Z-Stack. The application running on the base card can make use of this ZigBee stack through APIs provided on the base card driver. The XMC-660 hardware interface for ZigBee includes UART for data communication and GPIO lines for configuration and control. Software support includes drivers for the Windows XP, Linux GPPLE and VxWorks 6.7 (roadmap) operating environments. Curtiss-Wright Controls Embedded Computing, Leesburg, VA. (613) 254-5112. [].

FPGA-Based Industrial Ethernet Module Integrates a High-Capacity Switch-IP

An industrial Ethernet module (IEM) implements a costefficient and flexible solution to adapt to various Industrial Ethernet technologies. The IEM from IXXAT is based on an Altera Cyclone III FPGA, which in addition to an integrated 32-bit CPU, PHYs, magnetic modules and RJ45 jacks, includes the new 3-port switch IP for Profinet, EtherNet/IP and Modbus-TCP. The switch IP has been developed by IXXAT in close cooperation with one of the world’s leading experts in switch technologies. The IP implements a store-and-forward switch and is explicitly designed for real-time Ethernet networks. An implementation as cut-through switch is already in preparation. The switch IP makes use of priority information to transfer real-time Ethernet frames with lowest latency and jitter—in the range of a few microseconds. The switch IP comes with integrated MACs that support 10/100 Base-T, half and full duplex. To make use of features like static or dynamic address table, aging mode, routing mode, etc., the customer can configure the IP by using its management interface. On top of this, the internal port supports filters and backpressure to reduce the load of the connected system. The switch IP is also available as a product, based on Altera Cyclone III (further FPGA types on request). Furthermore, IXXAT also offers the development of fully customized adaptations of the switch IP on a project basis. IXXAT, Bedford, NH. (603) 471-0880. [].



ETX 3.0 Module Combines Core 2 Duo Performance with Legacy I/O

A series of ETX 3.0-compliant computer-on-module (COM) products is based on Intel’s LV and ULV Core 2 Duo and Core Duo processors. The ETX-945 series modules from Diamond Systems operate over an extended temperature range of -40° to +85°C, making them suitable for performance and reliability-critical applications, including aerospace, defense, transportation, energy management and industrial automation. To simplify the migration of PC/104-Plus stacks to these high-performance Intel processors, the ETX-945 is available as a pre-integrated “Pluto” module with a PC/104-Plus expansion interface. The highly compact (4.5 x 3.7 inches; 114 x 95 mm) ETX-945 is built around Intel’s 945GME and ICH7M chipset, and provides an onboard SODIMM socket for up to 2 Gbytes of high-speed DDR2 system DRAM. The ETX-945’s high-resolution display controller works with analog and LVDS-interfaced CRTs and LCDs at resolutions up to 2048 x 1536 pixels and supports dual independent displays. Other onboard I/O interfaces include one 10/100 Mbit/s Ethernet port, two SATA interfaces (support 1 drive each), one IDE interface (supports 2 drives), four USB 2.0 ports and two serial ports. Supported operating systems currently include Windows XP and Linux 2.6, with support for additional OSs and RTOSs (real-time OSs) available on request. The ETX-945 CPU module is priced at around $650. Diamond Systems, Mountain View, CA. (650) 810-2500. [].


PCI Express XMC Module with Dual Channel 1 GS/s 12-bit Digitizer

An XMC I/O module features dual channels of 1 GSample/s 12-bit digitizing with a Virtex5 FPGA computing core, DRAM and SRAM memory, and eight lane PCI Express host interface. The X5-G12 from Innovative Integration includes a Xilinx Virtex5 SX95T or LX155T with 512 Mbyte DDR2 DRAM and 4 Mbyte QDR-II memory to provide a high-performance DSP core for demanding applications such RADAR and direct RF digitizing. The close integration of the analog I/O, memory and host interface with the FPGA enables real-time signal processing at rates exceeding 300 GMAC/s. The X5 XMC modules couple Innovative’s Velocia architecture with a high-performance, 8-lane PCI Express interface that provides over 1 Gbyte/s sustained transfer rates to the host. Private links to host cards with >1.6 Gbyte/s capacity using P16 are provided for system integration. The X5 family can be fully customized with VHDL and Matlab using the FrameWork Logic toolset. The Matlab BSP supports real-time hardwarein-the-loop development using the graphical, block diagram Simulink environment with Xilinx System Generator. Software development tools for the X5 modules provide comprehensive support including device drivers, data buffering, card controls, and utilities that allow developers to be productive from the start. At the most fundamental level, the software tools deliver data buffers to your application without the burden of low-level real-time control of the cards. Software classes provide C++ developers a powerful, high-level interface to the card that makes real-time, high-speed data acquisition easier to integrate into applications. Software for data logging and analysis are provided with every X5 module. Data can be logged to system memory at full rate or to disk drives at rates supported by the drive and controller. Get IP Connected withthe technology and IP logic cores are also available for SDR applications that provide from 16 to 4096 DDC channels. These cores transform X5 modules providing solutions nowfor host your application. Software tools into versatile receivers using proven logic cores from R-Interface and Innovative, ready for integration into companies development include C++ libraries and drivers for Windows and Linux. Application examples demonstrating Get the module features andresource use arefor provided, Connected is a new further exploration into products, technologies and companies. Whether your goal including logging A/D samples to disk.

Ad Index

is to research the latest datasheet from a company, speak directly with an Application Engineer, or jump to a company's technical page, the goal of Get Connected is to put you in touch with the right resource. level of serviceDesign you requireFlexibility for whatever type of technology, Bendable LightWhichever Guides Allow Get Connected will help you connect with the companies and products A line of flexible light guides you are searching for.

Innovative Integration, Simi Valley, CA. (805) 578-4261. [].

Ruggedized Atom-Based COM Express Module

A low-power, long-lifecycle computer-on-module (COM) is based on the cost-effective, ultra-low-power Intel Atom N270. The Ampro by Adlink Express-ATR is specifically designed from the top down using the Ampro by Adlink Extreme Rugged design methodology for applications that require MIL-STD-202F shock and vibration compliance and an extended -40° to +85°C operating temperature range. The Express-ATR is a 95 x 95 mm low-power consumption COM that features the Intel Atom processor and embedded 945GSE Express chipset to provide a low-power, low-cost alternative for applications that require only single core performance levels. The Intel Atom processor N270 features a thermal design power of just 2.5 watts at peak levels while supporting Intel SpeedStep Technology. Although much smaller in size, Intel Atom processors share the same architecture as Intel Core2 Duo processors. Atom processors support Hyper-Threading Technology, which allows more than one code thread to be executed at the same time on a single core. The Express-ATR is an entry-level COM Express module for systems that require ruggedness and a full set of graphics features. The module supports an extended operating temperature range of -40° to +85°C and comes with integrated support for a high-resolution CRT, single/dual channel LVDS and TV-Out (SDTV and HDTV). In addition to the onboard-integrated graphics, the chipset’s SDVO bus can connect to DVI, TMDS, or additional LVDS or TV-Out device controllers on a custom designed carrier board. The graphics support is well-suited for in-vehicle systems, lab equipment, test and measurement applications, video preprocessing and industrial control. The Express-ATR supports up to 2 Gbytes of DDR2 533 MHz memory on a single SODIMM socket. The Mobile Intel 945GSE Express chipset integrates an Intel Graphics Media Accelerator 950 that provides CRT, single/dual channel LVDS and TV-Out (SDTV and HDTV). The board also allows connection of up to three additional PCI Express x1 devices on the Intel I/O Controller Hub, ICH7M Southbridge. List price is $325.

can be used to facilitate a wide range of adaptable panel indicating options. The flexible light guides from Elma Electronic can be bent in 90 degree or other angles. This allows the LED to be mounted to a PCB while fixed to a panel at a right angle. Available in various lengths, the Get Connected with technology and companies prov LED mounting can be done in a wide array of posiGet Connected is a new resource for further exploration into pro tions, locations, around corners or components, and datasheet from a company, speak directly with an Application Engine more. in touch with the right resource. Whichever level of service you requir Connected will help you connect The flexible lightGet guides have common sizes of 2with the companies and produc mm diameter, or 150 mm length, and black and clear tube colors. However, the light guides are customizable to the customer’s requirements. They are designed for -40° to 85°C ambient temperatures and housings and lens material to UL94 flammability standards. Elma also offers surface mount design (SMD) LEDs that can be used in conjunction with the flexible light guides. Pricing starts under $0.75 depending on volume and configurations.


Elma Electronic, Fremont, CA. (510) 656-3400. [].

Get Connected with companies and products featured in this section.

Get Connected with companies and products featured in this section.

ADLINK Technology, San Jose, CA. (408) 360-0200. [].




Small Linux Networking Server Offers Low-Cost Edge Computing for OEMs

A new, extremely small network server provides customers with a powerful engine for deploying advanced applications at the network edge, all in an integrated, thumb-sized package. The XPort Pro from Lantronix is the newest addition to the company’s XPort family of embedded Ethernet networking and compute modules. In addition to having the identical form factor as the original XPort, the XPort Pro’s leading-edge architecture, 32-bit processing power and ample memory allow resource-intensive applications to be deployed on a single platform. Advanced networking and security features enable machine-to-machine (M2M) edge computing with unlimited customization and application hosting, providing developers with faster time-to-market and unprecedented application development options. XPort Pro is available running Linux and IPv6, bringing the product to a global community and providing Linux developers with a tiny, powerful compute platform, along with an industry standard development environment. XPort Pro is also available with the Evolution OS turnkey operating system and software development kit. This product also includes Lantronix’ patent-pending Virtual IP (VIP) Access technology, which allows seamless integration with the company’s ManageLinx remote services enablement platform. XPort Pro provides over five times the processing power, and 32 times the memory of its predecessor. It also features SSH and SSL security and encryption, and supports a variety of protocol conversions. Lantronix is offering free XPort Pro evaluation kits to qualified customers, for a limited time. Lantronix, Irvine, CA. (949) 453-3990. [].

Fanless Box Computer—Ruggedness in Mobile and Extreme Apps

A new rugged box computer complies with the EN 50155 railway standard, fulfills IP67 and has an E1 approval from the German Federal Motor Transport Authority. The RC1 from Men Micro provides up to 1.6 GHz of processing power and combines a robust 8.66 inch x 5.12 inch x 3.39 inch (220 mm x 130 mm x 86 mm) aluminum enclosure with solid wall-mounting equipment to provide ruggedization in even the most extreme applications, from hot, dry and dusty to extremely humid or cold environments. It is specifically designed for mobile applications, including commercial vehicles such as agricultural or construction machines, trains, buses or cable cars, but can also be used in avionics, medical engineering and industrial automation applications. The RC1 is available as standard either with a 3.5" 4:3 color TFT LCD touch panel with a resolution of 640x480 pixels, 262,144 colors and touch functionality for service purposes, or without a display. All I/O interfaces, including two Fast Ethernet interfaces and a service interface with USB and RS-232, are accessible at the front via robust 8-pin M12 connectors. Two slots are available for additional I/O options implemented using special SA-Adapter kits. The onboard FPGA allows the integration of further interfaces such as CAN bus, RS-485, IBIS or binary I/O. A second power input makes it possible to connect a backup power source, such as a battery, to add reliability in the event of a power failure. As a power class S2 unit, the RC1 remains functional despite power interruption for up to 10 minutes. The 512 Mbyte RAM and 2 Gbyte flash memory offer sufficient space for all wireless functionality like Bluetooth, WLAN, WIMAX, GSM/GPRS or UMTS via two optional N-Type antenna connectors available at the front panel. The use of Intel's Atom processor guarantees fast performance and a minimum product availability of seven years. Pricing starts at $1,413 per unit without display and at $1,839 with display. MEN Micro, Ambler, PA. (215) 542-9575. [].



High-Resolution Digitizer Line Provides High Accuracy, Low Noise, High Performance

Three new PCI digitizers offer sampling rates of 10 MSample/s, 20 MSample/s and 40 MSample/s respectively. The newest members of the PCI-98x6 series from Adlink 98x6 family of digitizers include the PCI-9816, PCI-9826 and PCI-9846 4-channel, 16-bit high-resolution digitizers. The PCI-98x6 series combines three characteristics that make them optimized for various applications: high 16-bit accuracy, low noise and high dynamic performance over wide range of frequency domain, making them suitable for applications such as radar/lidar design and test, ultrasonic imaging, non-destructive testing, spectral monitor and automatic electronic testing. For example, the typical dynamic performance for the PCI-9816 includes 12.6bit effective number of bits (ENOBs) and a 78.4 dBc signal-to-noise ratio (SNR) at an input frequency of a 1 MHz sine wave. For applications that need high-density channels in one system, the PCI-98x6 series provides a specially designed system synchronization interface (SSI) to synchronize up to four cards. Each member of the PCI-98x6 series also includes up to 512 Mbytes of onboard memory for ample data storage to allow extended acquisition times. Adlink provides both legacy drivers for program development in Microsoft C++ and Visual Basic and task-oriented drivers, such as DAQPilot, to accelerate the development cycle. The PCI-98x6 series of digitizers can also be configured and operated in NI LabVIEW by using DAQPilot’s Express VIs and Polymorphic VI. Adlink also offers customized input ranges or higher bandwidth options on the PCI-98x6 series for OEM applications. The PCI-98x6 series of digitizers is currently available for a list price starting at $1,699. ADLINK Technology, San Jose, CA. (408) 360-0200. [].


AMC Carrier for Three IndustryPack Modules

A standard double-width mid-size or full-size AMC.1-compliant carrier can handle three single-size IndustryPack (IP) modules. The TAMC200 from Tews Technologies can be used to upgrade well-known and well-proven IndustryPack I/O solutions to a high-performance form factor, and to provide AMC users a large selection of over 200 off-the-shelf IndustryPack modules for analog, digital, communications, motion control, CAN and other various functions. All IP interrupt request lines are mapped to PCIe INTA. Alternatively, message signaled interrupts (MSI) can be used. The TAMC200 also provides a special IP interrupt status register for fast interrupt source detection. The IP power lines are fuse protected by self-healing fuses and RF filtered. Easy I/O cabling is facilitated with the use of three 68-pin SCSI-V type connector (VHDCI/Champ) that provides access to all IP I/O lines. Designed for demanding environments, the TAMC200 operates from -40° to +85°C and has a five year warranty. TEWS Technologies, Hastenbeck, Germany. +49 (0) 4101 4058-0. [].

Smart Camera for Single Point Machine Vision Inspection

A highly integrated smart camera comprises all of the elements of an industrial machine vision system. Powerful and quick-to-deploy, the BOA from Dalsa is targeted for automated quality inspection applications and factory automation. The all-in-one machine vision solution is smarter, easier to use, and more flexible than previous generations of smart cameras. It is the first smart camera in its class to incorporate multiple processing engines. This enables algorithm optimization via DSP, application management via CPU, and sensor management via FPGA. In addition, this smart camera offers truly embedded application software, which is easily set up through a standard Web browser. No software to install, no version control problems. BOA delivers flexibility via a rugged, easy-to-mount enclosure, built-in factory communications and a low deployment cost. The ultra-small BOA has been specifically designed for industry. Its 44 mm cube form factor is aimed for tight-fit applications. The IP67-rated housing means that the camera can be directly deployed in harsh, wash-down environments. This is particularly useful to meet cleanliness standards in the food and pharmaceutical industries, and eliminates the need for a separate, costly protective enclosure. BOA’s small form factor and easy mounting capabilities allow it to integrate easily into existing production lines, machinery or moving equipment. Dalsa has evolved and embedded the complete functionality of iNspect software in BOA. iNspect is suitable for both first time and experienced vision users; it offers inspection capabilities that can be readily applied across a multitude of applications. This easy-to-use, graphical point and click interface, allows users to rapidly prototype and deploy solutions, and it is available with a fully featured emulator for offline application development and debugging. The BOA vision system is available immediately in a monochrome version (model BVS0640M). A color version will be available later in the year.

Ad Index

Expansion Modules Add SUMIT Connection and Flexibility to PC/104-Form Factor Designs Get Connected with technology and

Two new expansion modules companies providing solutions now feature the Stackable Unified Module Get Connected is a new resource for further exploration Interconnect Technology (SUMIT) into products, technologies and companies. Whether your goal is to research interface scheme developed by thethe latest datasheet from a company, speak directly an Application Engineer, or jump to a company's technical page, the Small Form Factorwith Special Interest goal of Get Connected is to put you in touch with the right resource. Group (SFF-SIG). These two expanWhichever level of service you require for whatever type of technology, sion cards from VersaLogic, dual Get Connected will help you connect with the companies and products Gigabit Ethernet expansion, and for. an you are searching in-stack DC/DC power supply, add additional flexibility to the company’s SUMIT-based line of products. The PC/104-sized expansion modules are suitable for systems that need a small (3.55” x 3.78”) footprint. The SUMIT interface makes itGet easier for deConnected with technology and companies prov sign engineers to add bothGet standard and is a new resource for further exploration into pro Connected custom expansion boardsdatasheet to the system, from a company, speak directly with an Application Engine since it offers both low- in and high-speed touch with the right resource. Whichever level of service you requir Get Connected will help you connect with the companies signals. This includes PCI Express (PCIe) lanes, USB ports, LPC, SPI and produc and SMB. To support legacy applications, these new accessory cards are also equipped with a PC/104 ISA feed-through connector. The VL-EPMs-E1 Ethernet card features both high-speed Ethernet channels and optional wireless functionality via a PCI Express Mini Card interface. The VL-EPMs-PS1 power supply accepts +9V to +40 Vdc input and provide +5V and ±12 V output (up to 50 watts) for powering the embedded board stack. Customization is available including conformal coating, BIOS customizations, revision locks, custom labeling, high-G shock and vibration treatment, custom testing and screening, etc. VersaLogic’s accessory modules can be used to complement any SUMIT-enabled SBC to form a complete SUMIT ecosystem, and can be mixed with legacy PC/104 modules. They are with companies and compatibleGet withConnected any SUMIT-based SBC, including VersaLogic’s Z5xx products featured this the section. Atom PC/104-sized board,inand Core2 Duo EPIC-sized SBC. Pricing in OEM quantities for the VL-EPMs-E1 is around $250, and around $200 for the VL-EPMs-PS1.


VersaLogic, Eugene, OR. (541) 485-5712. [].

Dalsa, Bellerica, MA. (978) 670-2000. []. Get Connected with companies and products featured in this section.




SBC Supports Twin AMD Six-Core Opterons

A PICMG 1.3-style single board computer (SBC) supports twin, AMD six-core Opteron processors to deliver 12 cores of processing power. The MB-80020 from Win Enterprises features the AMD RS5690 chipset, FireWire, LSI chip and PCI-X. In addition to AMD Six-Core Opteron Processors, the MB-80020 accommodates AMD Dual- and Quad-core Opteron processors for scalability. The device supports 32- and 64-bit data-intensive computing that enables OEMs to easily upgrade high-performance product lines. The MB-80020 supports up to 20 lanes of PCI Express providing unsurpassed throughput. An MXM-II connector supports an ATI Radeon e2400 mobile MXM-II Graphics Card that enables graphic support for applications such as medical imaging, scientific computing, oil and gas, surveillance and other intense applications. Optional cards are available to provide HD Audio output and Firewire support. Other key features include the AMD SR5690 / SP5100 Chipset and 4 Gbyte memory. Temperature, fan speed and voltage are monitored, and a watchdog timer generates software selectable timeouts and system resets. The board provides dual 10/100/1000 Ethernet LAN ports with dedicated PCI Express, and a HyperTransport Link supports a stackable IP-90330 CPU module with second processor and additional memory. In addition, dual RS-232 serial ports and four USB 2.0 port headers are provided. An optional graphics card allows onboard DVI-I and LVDS ports to pass dual-head video signals. The single board price for the motherboard is $722, less CPU and memory. WIN Enterprises, North Andover, MA. (978) 688-2000. [].

Dual PMC Expansion on Core2 Duo VME Carrier Card

Many long-life VMEbus systems require high-performance expansion using PMC or XMC modules. Xembedded has announced the XVME-9076 dualPMC 6U carrier module. The XVME-9076 single-slot dual PMC carrier provides support for two additional PCI/PCIX sites to the Xembedded XVME-6200 Core2 Duo VMEbus processor. The XVME-9076 expands the XVME-6200 processor to support up to three PMC modules or two PMC modules and one XMC module for functions such as FPGA, Ethernet, SCSI, serial port, digital I/O, analog I/O and special-function PMC modules. The PCIe x4 is a high-speed connection in each direction between the XVME-9076 and the XVME-6200 processor. The XVME-9076 is compatible with PMC 2.0 specifications for IEEE P1386 modules. The XVME-9076 offers two 32/64-bit, 66/133 MHz sites, one with rear I/O out P2 of the carrier and the other site with rear I/O out the optional P0. Both PMC sites are capable of providing 14 watts of power. The XVME-9076 expansion module is available in commercial and air-cooled versions. Xembedded, Ann Arbor, MI. (734) 975-0577. [].



COM Express Module with Intel Core 2 Duo Provides Flexible Configuration

A flexible COM Express design features the Intel CPU Socket 479, which enables support for Intel Core 2 Duo, Core Duo or single core Pentium M and Celeron M processors. The MB-80070 from Win Enterprises offers further flexibility in storage options by providing parallel and serial ATA interfaces. In addition, the unit enables easy system expansion by supporting interfaces that include 5x PCI Express, 32-bit PCI 2.3 and LPC bus. The board supports 24-bit dualchannel LVDS for high-quality video and is appropriate for cart-based medical, test & measurement, industrial automation, interactive client, military and other applications. Key features include support for Intel Core 2 Duo, Core Duo, Pentium M and Celeron M processors and the Intel 945GME with ICH7M chipset. The MB-80070 is PICMG COM Express R1.0-compliant, using the standard form factor of 95 mm x 125 mm; 3.8� L x 5�W. It supports one DDRII SO-DIMM socket and an Intel 82573L GbE Ethernet interface. System memory supports DDR II 400/533/667 up to 2 Gbytes. Integrated graphics support includes dual SDVO, Analog VGA, 24-bit LVDS, TV-Out interface. The board supports COM Express standard features including USB, SATA, ATA, GPIO and AC97 and is RoHS-compliant. Pricing begins at $280 in OEM quantities and does not include CPU. WIN Enterprises, North Andover, MA. (978) 688-2000. [].

with an Application Engineer, or jump to a company's technical page, the goal of Get Connected is to put you in touch with the right resource. Whichever level of service you require for whatever type of technology, Get Connected will help you connect with the companies and products you are searching for.

Advertiser Index Get Connected with technology and companies providing solutions now Get Connected is a new resource for further exploration into products, technologies and companies. Whether your goal is to research the latest datasheet from a company, speak directly with an Application Engineer, or jump to a company's technical page, the goal of Get Connected is to put you in touch with the right resource. Whichever level of service you require for whatever type of technology, Get Connected will help you connect with the companies and products you are searching for.




ADLINK Technology America, Inc........................................................................................2................................................................................................ Birdstep Technology..........................................................................................................26............................................................................................................ Cogent..............................................................................................................................28.......................................................................................................... End of Article Products Concept Developement . ...................................................................................................37.............................................................................................................. Elma Bustronic Get Connected with companies and Get Connected products featured in .this section. Equipto Electronic Corp. with companies mentioned in this article. Interface Keil Software.....................................................................................................................52..................................................................................................................

Get Connected with companies mentioned in this article. MEN Micro, Inc.................................................................................................................33......................................................................................................... National Instruments.........................................................................................................27.....................................................................................................................

Get Connected with companies and products featured in this section.

One Stop Pentair Electronic Pentek, Performance Technologies.................................................................................................15...................................................................................................................... Phoenix International.........................................................................................................36........................................................................................................... Data Acquisition Boards Showcase....................................................................................41......................................................................................................................................... Real-Time & Embedded Computing Conference..................................................................49................................................................................................................ Red Rapids, Themis Computer..............................................................................................................29.............................................................................................................. UEI.....................................................................................................................................8............................................................................................................... VersaLogic Corporation.....................................................................................................51..........................................................................................................

RTC (Issn#1092-1524) magazine is published monthly at 905 Calle Amanecer, Ste. 250, San Clemente, CA 92673. Periodical postage paid at San Clemente and at additional mailing offices. POSTMASTER: Send address changes to RTC, 905 Calle Amanecer, Ste. 250, San Clemente, CA 92673.



Need Flexibility?


ad page 51


ersaLogic’s new generation of embedded de st of both single board computers gives you the best IT™ interface worlds. The new industry-standard SUMIT™ s well as legacy supports both next-generation speeds, as bus convenience.

• • •

SUMIT includes high speed buses like lik USB and d PCI Express E

SUMIT extends the life of your design platforms

SUMIT includes low speed buses like SPI and LPC SUMIT enables stackable, expandable, exible designs that are easy to update

VersaLogic’s new SUMIT SBC Series enables you to build compact highperformance systems around both high and low speed buses. Simplify system expansion, especially for custom add-on boards. VersaLogic SUMIT boards also support plug-in expansion with legacy PC/104 modules. Free white paper. Visit to download the VersaLogic SUMIT Technology White Paper. With more than 30 years experience delivering extraordinary support and ontime delivery, VersaLogic has perfected the ne art of service, one customer at a time. Contact us today to experience it for yourself. Call (800) 842-3163 for more information. Recipient of the VDC Platinum Vendor Award for 5 years running!

1-800-824-3163 1-541-485-8575

PICTURED: VersaLogic’s SUMIT-enabled Intell® Core2™Duo “Komodo”, Intell® Atom™ “Ocelot”, VL EPMs PS1 power supply EPMs E1 Dual VL-EPMs-PS1 supply, and VL VL-EPMs-E1 Gigabit Ethernet embedded computer products. SUMIT™ is a trademark of the SFF-SIG.

Development Solutions for ARM®, 8051 & XE166 Microcontrollers C and C++ Development Tools

� � � �

Best-in-class Keil and ARM compilers for small, fast code �enuine Keil μVision® IDE/Debugger/Simulator Fast development and verification using Device Simulation Easy device configuration with Device Database support for more than 1700 devices

Microcontroller Development Kits Royalty-Free RTX RTOS μVision Device Database & IDE μVision Debugger & Analysis Tools

Examples and Templates

C and C++ Compilers

A/D Converter

I/O Parts

Debug Run-Control

Timer/ Counter

Interrupt System

Debug Channel



Complete Device Simulation


RTOS and Middleware Components


ad page 52



Real-Time Clock


SD/MMC Interface

Flash ROM



ULINK® USB Adapters � � � � � �

�TA�, S�D & S�V support Debugging and Flash download ARM device support Data Trace for Cortex-M� ARM Instruction Trace �ULINKPro� � and 16-bit support �ULINK��

Evaluation Boards

TCPnet Networking Suite Flash File System USB Device Interface

Examples and Templates

RTX RTOS Source Code

CAN Interface

Keil PK51, PK166, and MDK-ARM support more than 1700 microcontrollers

RTOS and Middleware � � � � �

Easy implementation of complex applications Royalty-free RTX Real-Time Kernel TCP/IP Suite with Server Applications File System for ROM and Memory Cards Direct Support for USB and CAN interfaces

Download FREE Evaluations from 1-800-348-8051

Keil provides a wide range of evaluation boards for 8, 16 and 32-bit devices

Keil RL-ARM and ARTX-166 highly optimized, royalty-free middleware suites

RTC Magazine  

November 2009

Read more
Read more
Similar to
Popular now
Just for you