Category Archives: Hardware

IAR Systems updates development tools for ARM Cortex-M7 devices


IAR Systems shortens build times in leading development toolchain for ARM-based devices.


Version 7.40 of the incredibly-popular IAR Embedded Workbench for ARM has introduced support for ARM Cortex-M7 microcontrollers from Atmel. Beyond that, the tools now feature parallel build for shorter build times, as well as an integration of IAR Systems’ new tool C-STAT for powerful static code analysis.

template-screenschot-b-trace-timeline-crun-micrum

As you know, the ARM Cortex-M7 processor is the most recent addition to the ARM Cortex-M family. Not only focused on energy efficiency and high-performance, the MCUs are intended for use in a wide-range of applications including automotive, industrial automation, medical devices, and of course, the burgeoning Internet of Things.

The new version of IAR Embedded Workbench adds support for ARM Cortex-M7 devices from Atmel, including support for the double precision floating point unit. This covers the recently-revealed Atmel | SMART SAM E70, SAM S70 and SAM V70. In addition to these MCUs, support for a number of ARM Cortex-based devices from several other vendors have also been added.

In order to speed up build times, version 7.40 introduces parallel build. Users can easily set the compiler to run in several parallel processes and make better use of the available processor cores in the PC. This feature can have a major impact on reducing the build times of the compiler.

The add-on product C-STAT for powerful, integrated static code analysis is now available. Static analysis finds potential issues in code on the source code level and can be used to prevent errors such as memory leaks, access violations, arithmetic errors and array and string overruns. The analysis performed by C-STAT improves code quality and aids alignment with industry coding standards. It checks compliance with rules as defined by MISRA C:2004, MISRA C++:2008 and MISRA C:2012, as well as hundreds of rules based on CWE (the Common Weakness Enumeration) and CERT C/C++, for example. Users can easily select the rule set or individual rules to check their code against, and the analysis results are provided directly in the IAR Embedded Workbench IDE.

Interested? Head over to IAR Systems’ official page to learn more. Also, as of late last year, over 1,400 new example projects could be found in IAR Embedded Workbench, which supports Atmel’s entire portfolio of MCUs and MPUs.

DIY printing custom touch-sensitive displays


The PrintScreen lets anyone print their own custom displays.


In recent years, the industry has advanced by leaps and bounds when it comes to flexible touchscreens, albeit reproducing this technology for DIY projects has been a daunting task. Have you ever wanted to devise an interactive prototype with a display, like a smart plant that can reveal incoming calls or messages? Or, how about a touch-enabled greeting card for a loved one that could depict self-created symbols? Or, perhaps a postcard that illuminates in color when triggered? Despite countless innovative ideas, affordable displays for initial mockups can be quite bulky and only come in a few sizes.

printscreenAppcases

That may soon all change thanks to a group of researchers from Germany’s Saarland University who have developed a technique that could allow anyone to literally print their own custom displays, including touchscreens. The aptly-named PrintScreen has been designed to facilitate the next-generation of digital fabrication for customized flexible displays using thin-film electroluminescence (TFEL). Through the approach, an ordinary inkjet printer will pave the way for inexpensive yet rapid production of highly-customizable screens in low volume, ranging from a simple lab environment, a print shop or even at home. (It’s almost like temporary tattooing meets in-home graphic t-shirt making meets 3D printing.) The possibilities for endless as these TFEL displays can be integrated into almost every object in daily life, whether that’s affixed to paper objects, furniture, decorative accessories, bags or garments.

“We show how to print ultra-thin (120um) segmented and passive matrix displays in greyscale or multi-color on a variety of deformable and rigid substrate materials, including PET film, office paper, leather, metal, stone, and wood. The displays can have custom, unconventional 2D shapes and can be bent, rolled and folded to create 3D shapes. We contribute a systematic overview of graphical display primitives for customized displays and show how to integrate them with static print and printed electronics. Furthermore, we contribute a sensing framework, which leverages the display itself for touch sensing. To demonstrate the wide applicability of PrintScreen, we present application examples from ubiquitous, mobile and wearable computing,” the team of Simon Olberding, Michael Wessely and Jürgen Steimle writes.

Fab

How it works is relatively simple. Using a regular inkjet printer equipped with some conductive ink, Makers and designers alike can produce DIY displays from a digital template of a desired size and shape with programs such as Microsoft Word or PowerPoint and an editor like Adobe Illustrator. Making use of one of the two methods the researchers have been presenting, the template can now be printed out in about two to four hours, depending on the exact procedure, size and colors. Nevertheless, these results will be high-res displays that are just 1/10mm thick. To cover an entire standard printer page currently costs €20 ($21.69), with the most expensive part being the special ink that is required.

In order to light up a display segment, the team has crafted a controller that is tasked with applying a high-voltage, low-current AC signal between the upper and lower electrode layers. The luminance of a display segment or pixel is then controlled using pulse-width modulation (PWM), a standard method for controlling the luminance of LEDs. For mobile applications, their prototypical controller utilized a small driver IC that generates the high-voltage AC signal from a 1.0-7.0V DC power source. If a higher luminance is required, a stronger driver IC with a slightly bigger footprint can be implemented. Meanwhile, an ATmega2560 MCU triggers the optocouplers for multiplexing the high-voltage signal between display pins.

“The TFEL-specific ghosting effect in passive matrix displays can be significantly reduced by using a slightly modified controller design, thus further increasing the contrast of the matrix,” the team reveals.

printdisplay-3

“So far, nothing like this has been possible,” Olberding adds. “Displays were produced for the masses, never for one individual user.”

Printing a regular high-res display could be just the beginning, as the researchers hope PrintScreen can be used with other materials like leather, metal, stone, and wood in the future. Interested in learning more? The researchers will be exhibiting their solution at Cebit in Hanover, Germany this month. Meanwhile, you can read their entire paper here.

Henri wants to help designers create ambient UIs for smart devices


This interactive gizmo is making it easier for designers to explore the ambient UIs of tomorrow’s gadgets. 


Smart devices will undoubtedly continue to proliferate over the coming years. With billions of these connected gizmos expected to hit the market and ultimately make their way into our homes, this leaves one important question: How do you communicate with an Internet-enabled appliance when it doesn’t have a screen? Think about it: Your toothbrush. Your robotic vacuum. Your cooking utensils. Typically speaking, these sort of items emit luminescent cues that are used to catch your attention only when in need of a battery charge or some sort of malfunction. That’s exactly the conundrum design firm Method has set out to solve with what they’re calling Henri.

3043132-slide-s-3-a-box-for-designing-the-screenless

“[At the moment], there isn’t an easy way to design that. You need someone with fairly strong programming skills,” Daniel Nacamuli, Method’s lead interaction designer explains.

Instead, Henri is an interactive gadget that wants to make it easier for designers to explore the ambient user interfaces of IoT devices. The device has been developed to function as an abstract stand-in for a connected home product such your smart lights, thermostats or locks. Housed in a wooden enclosure, the system is comprised of a central box, two control panels and a desktop user interface. The main console is packed with an Atmel based Arduino, a round set of LED lights, and a built-in speaker.

3043132-slide-s-8-a-box-for-designing-the-screenless

Two control panels of steel knobs are linked to Henri, enabling users (even without any sort of coding background) to easily experiment with enchanted interface elements. With just a few turn of its dials, designers can devise a wide-range of patterns of lights and sounds with varying pulses, hues, intensities and durations (zero to 16 seconds), as well as watch them play back in real-time on its central hub.

The Arduino is tasked with recording the sequence and relaying it back to the desktop interface for storage. The main box also syncs with synthesizers so users can simultaneously create sound cues. Later, Henri can reprogram all of this into the final piece of hardware.

3043132-slide-s-5-a-box-for-designing-the-screenless

Additionally, Henri will certainly come in handy for designers across a plethora of industries like gaming. “You could use the controller to fine tune animation on a screen. Say your animation is of the sun setting and that sun is going to move down. Normally they’ll use some animation software with a timeline. What could feel more natural, is to turn the dial on the Henri, to turn the speed,” Nacamuli tells Wired. 

This interactive device is merely one component of Method’s overall efforts to raise awareness around the design of ambient user interfaces for the Internet of Things. The firm unveiled the Henri box as part of a workshop earlier this year in its Bay Area office during San Francisco’s IxDA conference. There, attendees were paired in teams and asked to program light and sound patterns on the Henri to communicate test scenarios.

outcome4-1380

“After initial brainstorming, it was clear we could use this opportunity to solve a set of problems and roadblocks that have inhibited us from designing non-screen based products in our own practice,” the team writes. “Henri allows those concepts to be tested real-time, and enhance the overall capabilities of both the designer and the product. It helped people literally think outside of the box, and be productive while playing.”

Intrigued? Head over to the project’s case study here.

Pleurobot is a lifelike robotic salamander


This bio-inspired robot may be the future of search-and-rescue missions.


Bio-inspired robotic locomotion is a fairly new sub-category of bio-inspired design, revolving around learning concepts from nature and applying them to the design of real world engineered systems. More specifically, this field is about making robots that are inspired by biological systems. When it comes to these bio-mimicking bots, it seems as though we’ve seen just about everything, ranging from bats to spiders to dogs. And while it may not be all that difficult to make a bot that looks like an animal, having it behave like one is an entirely different story.

Pleurobot_front-1425415109545

Meet Pleurobot. Recently developed by the BioRob at EPFL and NCCR Robotics, it is a salamander-inspired robot that is truly amphibious, meaning it is capable of walking, changing its gait to navigate uneven terrain, and even swimming. As the researchers note, the key to Pleurobot’s eerily-lifelike motion is its unique design, which was based on 3D X-ray movies of a real salamander walking, waddling and swimming. By tracking up to 64 points on the animal’s skeleton, the team was able to record movements of bones, and then deduce the number and position of active and passive joints needed for the robot to reproduce the 3D movements with reasonable accuracy.

According to its research proposal, the team first created a snake robot Kulko with tactile sensors in order to test the suggested control framework. This was comprised of a serial connection of 10 identical ball-shaped joint modules, along with a smooth surface to avoid getting stuck against obstacles. Each of the joint modules had 2-degrees of freedom (pitch and yaw), and had used servo motors as its actuators. On each side of every joint module, there were four force sensing resistors tasked with measuring contact forces — these were the only contact points between outer shell and inner structure.

“The current layout of the sensors can only measure horizontal forces which is enough for the application. The total force is estimated by summing forces measured by each FSR on the module. Each module also contains two Lithium-Ion batteries, angle sensors (magnetic rotary encoders), voltage regulation card, battery charger card and microcontroller card. The microcontroller card is based on the Atmel microcontroller AT90CAN128 and is continuously measuring position of the motors and controlling them with a PWM signal. Modules communicate with each other over CAN bus,” BioRob’s Tomislav Horvat writes.

Filming_140small

By design, Pleurobot provides torque control for all the active joints, which enabled its creators to apply their neural network models of the spinal cord neural circuits of the salamander and to activate virtual muscles to replicate the recorded animal movements along with realistic viscoelastic properties. This was imperative when obtaining a fundamental understanding of vertebrate motor control.

What’s more, Pleurobot is also waterproof. While this feature actually proved to be the project’s most daunting assignment, the salamander-like project is currently using a water-repellant skin suit. Moving forward, the team hopes to improve upon this layer with aspirations that one day, the bio-mimicking robot will have a role in search-and-rescue efforts, with shallow waters for example. Its amphibious nature will enable it to go where humans cannot.

In the future, the team says it plans to use Pleurobot’s design methodology to bring early tetrapods to ‘life.’ So sure, we can write about it all day, but watching it in action is so much better! Those wishing to read up on the bio-mimicking project can download the team’s detailed proposal here, or head over to its official page for an abbreviated version.

VIPER is a cross-platform Python IoT design suite


VIPER is a smart object development suite that brings cloud and IoT connectivity to your projects with just a click of the mouse.


New York City-based startup ThingsOnInternet has launched a Kickstarter campaign for their new easy-to-use development suite for interactive Internet of Things (IoT) designs. As its name implies, VIPER — or “Viper Is Python Embedded in Real-time” — makes it possible for Makers and embedded designers to create their next connected project in Python for Arduino, UDOO and Spark, all in in real-time. And, unlike other solutions that already exist today, this collection of products is platform-agnostic and compatible with all sensors and kits.

photo-1024x768-1

The idea was first conceived after conducting some detailed market analysis, where the company discovered that designers, Makers and programmers all faced a similar set of challenges. In hopes of simplifying how “things” are brought onto the Internet, VIPER converged a series of components to better streamline the process. This included an IDE to manage and program the boards, a Virtual Machine to serve as its operating system, a plug-and-play TOI Shield, an extensive library of ready-to-use functions, and a mobile app to act as the interface for smart objects. On top of that, it’s also cloud-ready. With just a little coding, users can develop a wide-range of IoT applications, ranging from interactive storefronts, to home and industrial automation systems, to art and museum installations, to smart farming.

Arduino-Shield1

“Designers aim to create behaviors that can co-exist at the same time and they are often frustrated by ‘anomalous’ and unexpected results on their projects. Makers, who have higher competences, invest a lot of time in understanding how to code multithreaded behaviors, how to manage interrupts in C++, etc. and sometimes their code become really hard to be maintained. Programmers are frustrated by executing ‘boring’ tasks for their customers, one of them is related (again) to multithread, interrupts, callbacks, timers and exceptions,” ThingsOnInternet writes.

Since millions of developers already know Python, VIPER decided to make the programming language readily accessible for commercial interactive products as well, therefore amplifying the potential for smart objects to be as pervasive as mobile devices in their ease of design interactivity. To do this, VIPER provides a browser-based, minimal-installation development environment where users can write code with extensive library support and have it executed on any Arduino-like board. What’s great for designers is that, with VIPER, it leaves them able to focus on the features and functionality, not the tediousness, along with a mobile app to control their creation for free.

Arduino-Shield

“Devices like Arduino, UDOO, ST Nucleo, Spark Core, Photon and the last released Spark Electron are microcontroller boards that revolutionized the world of amateur and semiprofessional electronics. They allowed hundreds of thousands of people around the world to give objects a new life by making them interactive, able to communicate and interact with humans,” the team explains. “Unfortunately, programming them is quite easy for engineers and computer scientists, while most of the users are able to exploit only part of the huge potential of such incredible boards. Here comes the idea of TOI to extend the world of smart and interactive object design to everyone. VIPER allows in a few clicks to convert a common lamp in a smart assistant that reminds us to take the umbrella, turn on the air conditioning while monitoring the house for intrusions.”

In order to use the suite, Makers and developers simply download a one-time package from the company’s website onto either their PC or USB stick. Beyond that, VIPER includes an embedded, portable Python 3.0 engine to help make everything as easy as can be. Users can then launch the VIPER IDE and begin making. All that’s left from there is connecting its accompanying mobile app to serve as the UI for the project.

someshield

VIPER runs on an Arduino Due (SAM3X8E), and can even be ported onto the recently-revealed Atmel | SMART Cortex-M7 family. As its creators reveal, code developed on an Arduino Due can also be implemented onto these new MCUs in a matter of two clicks. Furthermore, the suite features an Arduino and Spark Proton-compatible, plug-and-play TOI shield. Simply attach either a Due or Photon to the shield and start playing with any of the VIPER examples found in its library. (This collection of modules includes CC3000 Wi-Fi for Spark Core and Adafruit Shield, Adafruit/Sparkfun Thermal Printer, Adafruit NeoPixel LED, RTTTL smart melody player, Streams library, as well as TCP and UDP network protocols.) Aux ports are even included, enabling the use of other sensors like Grove, ThinkerKit, Phidgets, and Adafruit NeoPixel LED strips.

Through its IDE, users can ‘viperize’ theirs boards by installing them using the VIPER Virtual Machine. Once completed, a board is no longer a simple Arduino Due, Spark Photon or UDOO; instead, it has a multi-threaded, real-time operating system running on it, and a virtual machine ready to execute compiled Python 3 scripts. Ready to design your next smart project? You can head over to its official website, or check out the team’s recent successfully-funded Kickstarter campaign here.

Turning on a lamp via the Internet the Big Bang Theory way


A team of Atmel Norway engineers decided to make their own rendition of the Big Bang Theory Internet-controlled lamp scene. (Yes, even Sheldon Cooper would approve of this one.) 


How many of you are fans of the CBS hit sitcom series, Big Bang Theory? Well, you’re in luck. If you recall an episode from the show’s first season, entitled “The Cooper-Hofstadter Polarization,” the team of Sheldon Cooper, Leonard Hofstadter, Howard Wolowitz and Raj Koothrappali successfully turned on a lamp via the Internet using an X-10 system.

To do so, the gang sent signals across the web and around the world from their apartment to connect not only their lights, but other electronics like their stereo and remote control cars as well.

“Gentlemen, I am now about to send a signal from this laptop through our local ISP racing down fiber optic cable at the of light to San Francisco bouncing off a satellite in geosynchronous orbit to Lisbon, Portugal, where the data packets will be handed off to submerged transatlantic cables terminating in Halifax, Nova Scotia and transferred across the continent via microwave relays back to our ISP and the external receiver attached to this…lamp,”  Wolowitz excitedly prefaced.

800px-X10_1

What’s funny is, the technology that the group of sitcom scientists was simulating could have just as well been done using a Wi-Fi network controller, like the WINC1500 module. However, at the time of airing back in March of 2008, open access for Internet users looking to control “things” around the house was seemingly something only engineers and super geeks thought possible.

In an effort to generate awareness around the upcoming IoT Secure Hello World training series, a team of Atmel Norway engineers decided to make their own rendition of the Big Bang Theory lamp scene using the ATWINC1500 IEEE 802.11b/g/n network controller and an Atmel | SMART SAM D21 Xplained Pro board, all secured by Atmel CryptoAuthentication devices.

After watching the Trondheim-based crew’s Cooper-Hofstadter IoT experiment above, be sure to check out a detailed description of the technology behind the project and learn more about the IoT Secure Hello World Tech on Tour seminar below.

Air traffic control system vulnerable to hackers, report finds


The United States’ system for guiding planes and other forms of aircraft is at an increased and unnecessary risk of being hacked.


A new Government Accountability Office report reveals that cybersecurity is “threatening the agency’s ability to ensure the safe and uninterrupted operation of the national airspace system.” The 42-page document entitled “Information Security: FAA Needs to Address Weaknesses in Air Traffic Control Systems” credits the FAA with taking steps to deter malicious hackers but concluded that significant security control weaknesses still remain.

Verkeerstoren_Brussels_Airport-1

One area of supcetibility, in particular, is the ability to prevent and detect unauthorized access to the vast network of computer and communication systems. These include controls for protecting system boundaries, identifying and authenticating users, authorizing users to access systems, encrypting sensitive data, and monitoring activity on the FAA’s systems, the report states.

The FAA relies on more than 100 of these air traffic systems to direct planes, with air traffic controllers responsible for an average of 2,850 flights at any given moment. As the Washington Post points out, 14,000 controllers work in three different types of facilities: 500 airport control towers that oversee landings and takeoffs; 160 facilities that direct planes to and from cruising altitudes; and 22 centers that supervise aircraft at cruising altitude.

According to the GAO’s latest findings, there are also a number of inadequate safeguards to prevent entry into the air traffic network from other, less-secure computer systems not directly involved in traffic operations. The report goes on to note that threats to the ATC system are on the rise, especially from terrorists, criminals and other foreign governments.

11943716-Air-traffic-controller-point-to-plane-on-radar-sonar-Air-Traffic-Control-Tower-Stock-Photo

Among the other notable vulnerabilities listed in the report include security weaknesses identified by the FAA weren’t always addressed in timely fashion, control assessments weren’t always comprehensive enough to find weaknesses, and shortcomings in monitoring for hacking incidents or unauthorized entries mean the FAA may not be able to contain, eradicate or recover from incidents.

“These shortcomings put (national airspace) systems at increased and unnecessary risk of unauthorized access, use, or modification that could disrupt air traffic control operations,” the report concludes.

Interested in reading more? Access the entire report here. With the number of cyber attacks on the rise and no apparent end in sight, how can you ensure that your network is indeed protected?

Report: Internet of Things expected to quadruple in size by 2020


Verizon reveals that while the IoT has expanded massively in the last couple of years, we’ve barely scratched the surface.


The Internet of Things has certainly transcended beyond its state of infancy and is well on its way of gaining momentum, according to Verizon at least. In its latest findings, the carrier revealed that more than a billion devices are already connected and running business-to-business IoT operations.

verizon-iot

In its “State of the Market” reportVerizon published that there were 1.2 billion various smart devices, and that the number is expected to rise to 5.4 billion by 2020 for an annual growth rate of 28%.

“It’s not hype. The Internet of Things is already having a massive impact on business. It offers organizations the opportunity to transform how they operate, and gives both new entrants and established players the ability to innovate and disrupt,” the company writes. “Adoption is growing rapidly, but IoT isn’t yet widespread. Whether you’re in the public sector or private; big or small — if you don’t have an IoT strategy, you should.”

Verizon experienced a 45% year-over-year revenue growth in its IoT business in 2014, with 4G LTE activations growing by 135%. Currently, the telecom manages more than 15 million IoT-enabled connections for a wide range of industries. To date, company experts estimate that just 10% of enterprises have deployed IoT technologies extensively, however research commissioned by Verizon from ABI Research forecasts massive growth ahead, with the number of business-to-business IoT connections more than quadrupling over the next five years.

Additionally, the global communications company also cites ABI Research in its revelations that organizations will introduce more than 13 million health and fitness tracking devices to the enterprise by 2018. In doing so, wearables can enhance wellness throughout the workplace, not to mention improve efficiency in hospitals and reduce the cost of healthcare.

b2b_iot_verizon_forecast

Among those who use Verizon’s services, manufacturing has seen the biggest increase in machine-to-machine operations, with a 204% increase year-over-year. It’s followed by finance and insurance with a 128% increase, and media and entertainment, which has experience an uptick of 120%. Home monitoring and hospitality weren’t too far beyond with 89% and 88% jumps, respectively. Verizon data also shows an 83% YoY growth in IoT in the transportation and distribution sector as well.

In fact, Verizon’s telematics experts note that 14 car manufacturers account for 80% of the worldwide automotive market, and all of them have a connected car strategy. The report predicts that by 2025, at least five countries will have set a “zero road fatalities” target, relying on intelligent connected cars and smart road infrastructure to avoid and mitigate accidents.

Verizon analysts add that in 10 years, smart cities capabilities will become a critical consideration for companies deciding where to invest and open facilities, due to their impact on operating costs and talent availability. Recent data already shows a 46% YoY growth in the number of IoT connections in the public sector.

According to the report, IoT growth is being fueled by a mix of technological, political and social factors which are driving more organizations to adopt IoT-enabled solutions. For example, use of social media and mobile technology has transformed consumer and citizen expectations, while the declining cost of sensors, connectivity, and data processing power is making the ROI equations for IoT projects look even more appealing.

The carrier went on to highlight ever-growing security concerns for a constantly-connected world as well by noting, “In a mature IoT world, there will be millions of intelligent endpoints, such as cars, pacemakers, and aircon units, each equipped with dozens of active sensors and millions of lines of code. Many of these endpoints will be accessible, often physically, to hackers. The network connections that these endpoints use to communicate may also be vulnerable, giving access to central applications and databases.”

Interested in reading the Verizon report in its entirety? Download it here.

ChipWhisperer-Lite is an educational board for embedded security


ChipWhisperer is the first open-source toolchain for embedded hardware security research including side-channel power analysis and glitching.


Side-channel power analysis refers to a method of breaking implementations of completely secure algorithms such as AES-256. Such capabilities have been known for a long time – the attack was first published in 1998. But even today many consider side-channel attacks exotic, and don’t take them seriously when designing secure systems. That is why Canadian startup NewAE Technology has launched a new project to help inform designers that they need to take these threats seriously, by teaching them how the attacks work!

photo-1024x768

Recently debuted on Kickstarter, the aptly named ChipWhisperer-Lite is essentially an educational tool, designed to introduce embedded enthusiasts to the area of side-channel power analysis. You may also recall the project from last year’s Hackaday Prize, where it garnered second place accolades.

Side-channel attacks aren’t magic; in fact, it is possible to design systems which are resistant to them. For instance, Atmel has a line of secure processors which would have encryption peripherals which cannot easily be attacked. Another example is the ATAES132 device – again this has resistance against side-channel attacks, so you could be more confident in the security of that device, compared to a generic microcontroller with an AES hardware peripheral (such as the AVR XMEGA). It’s all about managing the risk!

8a98245ccbc082ea5c6f1c36fe33147c_large

Additionally, the ChipWhisperer-Lite required a high-speed USB interface, and so, the NewAE Technology team turned to the Atmel | SMART SAM3U2C to accomplish this feat.

“While a number of systems are designed around generic interface chips, using a high-speed USB microcontroller gave me a lot more flexibility. In addition the cost of the microcontroller chip was cheaper than the stand-alone interface chip I would have used, so all these benefits came at no penalty to the BOM cost,” writes company co-founder Colin O’Flynn.

This shows the basic connections between the SAM3U2C and the FPGA. The external memory interface on the SAM3U2C is used to simplify data and control transfer to and from the FPGA.

This shows the basic connections between the SAM3U2C and the FPGA. The external memory interface on the SAM3U2C is used to simplify data and control transfer to and from the FPGA.

According to O’Flynn, the SAM3U family was selected based on set of criteria:

  • High-speed USB 2.0 interface
  • External memory interface with programmable timing parameters
  • TQFP Package (as he wanted people to be able to build this project themselves)
  • Lower cost than standalone interface chip (he had been looking for roughly $3-$4 in a quantity of 1,000)
  • ROM-resident USB bootloader (so that people building their own don’t need a programmer, and makes the board unbrickable)

“The external memory interface is actually critical to achieving a simple FPGA interface. This allows me to memory-map sections of the FPGA right into the SAM3U processor memory. If transferring data over USB to the FPGA, I can point the USB code from the Atmel Software Framework (ASF) to the location in the FPGA I want the data to go,” O’Flynn adds. “This means no need to copy the data multiple times between buffers, or use some specialized protocol to transfer data from the microcontroller to the FPGA.”

Beyond that, the SAM3U2C simplifies system management. Meeting USB sleep mode current limits (2.5 mA) means shutting off the FPGA and analog portions of the board. Standalone interface chips provide a ‘SUSPEND’ output which you can use, but having the microcontroller offered much more control, which ChipWhisperer-Lite’s creators were able to use for meeting inrush current limits.

The USB standard has limits on the inrush current; this current occurs when the USB device is plugged in and all the capacitors start charging. To avoid exceeding these currents most boards need a ‘soft-start,’ where power supplies are turned on after some delay (or after the USB device finishes enumerating).

“Putting this in the microcontroller gives me control over that delay if fine-tuning is needed, or even having the option of adding multiple switches or slower ramps using a PWM output,” says O’Flynn.

This shows the switch for the FPGA and analog power supplies. Depending on the total load, an RC filter can be added to slow the turn-on speed of the FETs.

This shows the switch for the FPGA and analog power supplies. Depending on the total load, an RC filter can be added to slow the turn-on speed of the FETs.

Using the SAM3U2C also provided a nice set of peripherals to use, too. The ChipWhisperer-Lite required a ‘target’ device that the user (i.e. student) programs with their algorithm of interest. For this case, the team selected an XMEGA MCU to serve as a programmable target for the student.

The XMEGA device can easily be programmed with only two wires (PDI), and this is generated by one of the SPI modules in the SAM3U. O’Flynn also used a USART module to communicate with the XMEGA, and finally another SPI module to download configuration data to the FPGA.

“While generic interface chips often have support for serial protocols (such as SPI or USARTs), the problem is they are normally limited in the number of channels offered, or I couldn’t use the serial-interface mode at the same time as high-speed parallel interface mode.”

In addition the details of the protocol (such as the low-level PDI programming protocol for the XMEGA) go into the firmware on the SAM3U2C, simplifying the higher-layer USB interface.

c0213e143d5dc2cf1c1fb4dde421ea6b_large

“I find it easier to develop those low-level protocols on an embedded system from within Atmel Studio 6.2, compared to trying to send timing-specific information across the USB bus to be processed by the interface chip! Anytime you can avoid USB debugging is time well spent in my books,” O’Flynn emphasizes. “Using an ASF application example as a starting point for the whole application let me rocket through development, with satisfyingly few moments of pounding my head against the desk figuring out why things weren’t working!”

A final nicety of the design was the ability to use the unique ID programmed into the SAM3U2C as part of the USB device serial number. In other words, the NewAE Technology crew could generate unique serial numbers for each device without requiring any special manufacturing step – every device is loaded with the same binary firmware yet still has a unique serial number. As an end-user, having unique USB serial numbers improves the experience since otherwise Windows will reload the driver when you change the USB port the device is plugged into.

“We’re eliminating the problem for good by making the tools open-source. Because this whole area is an active research area, the tools need to be open-source. This isn’t a case of attempting to seem sexy by adding the word ‘open-source’, but placing something of commercial value into the open-source domain, in the hope it spurs a larger community. This includes hours of tutorials on this area, more than just a few board files and some source code.”

Indeed, this project was devised as a fairly advanced piece of test equipment for well-seasoned Makers, embedded developers and computer engineers. That being said, it is important to note that it is not Arduino-compatible, nor does it work with Raspberry Pi or BeagleBone. However, O’Flynn does reveal that an Arduino-compatible, ATmega328P based target board is in the works. Impressively, ChipWhisperer-Lite also enables users to snap off the ‘target board,’ giving them both a main measurement tool and a target device.

e1dd963d975ca506d1570c18a6023fae_large-1

Interested in learning more? You can head over to its official Kickstarter page, where the team is well on its way to achieving its $50,000 goal. Pending all goes to plan, shipment is slated for August 2015.

EMAC releases a SODIMM-based take on the Atmel | SMART SAMA5D3


As its name implies, the ‘A5D36 is a SoM based on the Atmel | SMART ARM Cortex-A5 SAMA5D36.


EMAC has launched the SoM-A5D36, a system-on-module that runs Linux on a Cortex-A5-based Atmel SAMA5D3 processor, and offers up to 4GB of Flash, industrial temperature, and an optional carrier board. The wide temperature, fanless ARM 536 MHz SoM features 10/100/1000 BaseT Ethernet along with up to six serial ports. Beyond that, the board boasts up to 16MB of serial data flash and up to 512MB of LP DDR2 RAM, as well as supports LCD and resistive touch interfaces.

emac_som5d36

This isn’t the first time the Atmel | SMART SAMA5D3 has experienced some love from computer-on-module (COM) vendors. In fact, the Cortex-A5-based SoC has shown up on the ShiraTech AT-501, the Acme Systems Acqua A5, and the MYIR MYC-SAMA5D3X — all of which except the Acqua A5 use SODIMM connectors.

Using the same small 200-pin SODIMM form factor utilized by other EMAC SoM modules, the SoM-A5D36 packs all the core ARM processor functionality, including Flash, memory, serial ports, Ethernet, I2S audio, PWMs, timer/counters, A/D, digital I/O lines, clock/calendar, and more. The SoM-A5D36 is also designed to plug into a carrier board that contains all the connectors and any custom I/O required for an application. This approach enables a user to design a custom carrier board that meets for I/O, dimensional and connector requirements without having to worry about the processor, memory and standard I/O functionality.

Key specs of the SoM-A5D36 include:

  • Processor: Atmel | SMART SAMA5D36
  • Memory:
    • Up to 512MB LP DDR2 SDRAM
    • 16MB serial data flash
    • Up to 4MB eMMC flash
  • Networking: Gigabit Ethernet PHY; optional 2x GbE
  • Primary I/O:
    • 3x USB 2.0 host (1x includes device support)
    • 6x serial (4x defined) — 1x full, 3x RTS/CTS, 2x no handshake
    • 2x I2C
    • 2x CAN
    • I2S audio
    • 2x SD
    • 22x GPIO
  • Secondary I/O:
    • 24-bit LCD controller
    • 12-bit, 4-wire, analog resistive touch controller
    • 2x CAN
    • 2x SPI (“3 SPI CS”)
    • 2x I2C
    • 4x PWM
    • 6x ADC (12-bit)
    • 5x timer/counters, 3x prog. clock, RTC, reset, LEDs
    • External address data bus
  • Operating temperature: -40 to 85°C
  • Power: +3.3V
  • Dimensions: 67 x 60mm (200-pin SODIMM)
  • Operating system: EMAC OE Linux

Interested in learning more? Head over to EMAC’s official page here.