About Circuit Cellar Staff

Circuit Cellar's editorial team comprises professional engineers, technical editors, and digital media specialists. You can reach the Editorial Department at editorial@circuitcellar.com, @circuitcellar, and facebook.com/circuitcellar

Non-isolated Up Converters Support High-Performance GPUs

Vicor has announced a 12 V to 48 V non-isolated up converter to support 48 V high-performance GPUs in data centers that are still relying on legacy 12 V power distribution. The 2317 NBM converts 12 V to 48 V with over 98% peak efficiency, 750 W continuous and 1 kW peak power in a 23 mm x 17 mm x 7.4mm surface-mount SM-ChiP package. The NBM (NBM2317S14B5415T00) provides a complete solution with no external input filter or bulk capacitors required. By switching at 2 MHz with ZVS and ZCS, the NBM provides low output impedance and Megahertz-fast transient response to dynamic loads. The NBM incorporates hot-swap and inrush current limiting.

The NBM supports state-of-the-art 48 V input GPUs using Power-on-Package (“PoP”) Modular Current Multipliers (“MCMs”) driven from a 48 V node sourcing a small fraction (1/48th) of the GPU current. Current multiplication overcomes the power delivery boundaries imposed by traditional 12 V systems standing in the way of higher bandwidth and connectivity.

The Vicor Power-on-Package modules build upon Factorized Power Architecture (FPA) systems deployed in high-performance computers and large-scale data centers. FPA provides efficient power distribution and direct conversion from 48 V to 1 V for GPUs, CPUs and ASICs demanding up to 1,000 A. By deploying current multiplication in close proximity to high-current Artificial Intelligence (AI) processors, PoP MCMs enable higher performance and system efficiency.

Vicor | www.vicorpower.com

 

Displays Fit Diverse Embedded System Needs

Many Sizes and Solutions

The types of displays available for embedded applications are as diverse as embedded applications themselves. Whether your requirement is for small, smart, rugged or rain-proof, there’s probably a display solution that suits your system design needs.

By Jeff Child, Editor-in-Chief

Long gone are the days when the Graphics Processor Unit (GPU) market was filled with many semiconductor vendors jockeying for position. A combination of chip integration: graphics function moving inside microprocessors—and business consolidation: graphics chip vendors getting acquired, has narrowed the technology space down to mostly Intel, AMD and NVIDIA. And while these vendors tailor their products for high-volume markets, embedded applications must adapt those same GPUs to their needs.

With that in mind, makers of displays for embedded applications are constantly evolving their products to keep pace with the latest GPU technologies and both new and legacy display interface standards. Technologies range from small e-paper displays to rugged sunlight readable displays for the outdoors to complete Panel PC solutions that embed PC functionality as part of the display.

Mobile Dominates GPU Market

Although this article is focused on displays in embedded systems, it’s helpful to first understand the larger markets that are driving GPU technology. For its part, Jon Peddie Research (JPR), a market research and consulting firm focused on graphics and multimedia saw mobile devices as the dominate market when they did their annual review of GPU developments for 2017. In spite of the slow decline of the PC market overall, PC-based GPU sales (which include workstations) have been increasing, according to the review. In the mobile market, integrated GPUs have risen at the same rate as mobile devices and the SoCs in them. The same is true for the console market where integrated graphics are in every console and they too have increased in sales over the year.

Nearly 28% of the world’s population bought a GPU device in 2017, and that’s in addition to the systems already in use. And yet, probably less than half of them even know what the term GPU stands for, or what it does. To them the technology is invisible, and that means it’s working—they don’t have to know about it.

The market for, and use of, GPUs stretches from supercomputers and medical devices to gaming machines, mobile devices, automobiles and wearables. Just about everyone in the industrialized world has at least a half-dozen products with a GPU, and technophiles can easily count a dozen or more. The manufacturing of GPUs approaches science fiction with features that will move below 10 nm next year and have a glide-path to 3 nm—and some think even 1 nm.

Innovative Adaptations

Throughout 2017 JPR saw a few new, and some clever adaptations of GPUs that show the path for future developments and subsequent applications. 2017 was an amazing year for GPU development driven by games, eSports, AI, crypto currency mining and simulations. Autonomous vehicles started to become a reality, as did augmented reality. The over-hyped, consumer-based PC VR market explosion didn’t happen—and had little to no impact on GPU developments or sales. Most of the participants in VR already had a high-end system and the head-mounted display (HMD) was just another display to them.

Mobile GPUs, exemplified by products from Qualcomm, ARM and Imagination Technologies, are key to amazing devices with long battery life and screens at or approaching 4K. And in 2017 people started talking about and showing High dynamic range (HDR). JPR’s review says that many, if not all, the developments we will see in 2018 were started as early as 2015, and that three to four-year lead time will continue.

Lead times could get longer as semiconductor engineers learn how to deal with chips constructed with billions of transistors manufactured at feature sizes smaller than X-rays. Ironically, buying cycles are also accelerating ensuring strong competition as players try to leap-frog each other in innovation. According to JPR, we’ll see considerable innovation in 2018, with AI being the leading application that will permeate every sector of our lives. The JPR GPU Developments in 2017 Report is free to all subscribers of JPR. Individual copies of the report can be purchased for $100.

Photo 1.
The Internet of Displays is a range of miniature displays that offer small color displays with integrated Wi-Fi and a microSD/HDC slot.

Internet of Displays

Focusing on the small side of the display spectrum, in November 4D Systems announced the latest addition to its Internet-of-Display module family with its smallest LCD display yet. At 0.9-inch and powered by the Wi-Fi enabled ESP8266, it is well suited for miniature IoT projects. The Internet of Displays is the company’s range of miniature feature rich displays that offer small color displays with integrated Wi-Fi and a microSD/HDC slot (Photo 1). …

Read the full article in the April 333 issue of Circuit Cellar

Don’t miss out on upcoming issues of Circuit Cellar. Subscribe today!
Note: We’ve made the October 2017 issue of Circuit Cellar available as a free sample issue. In it, you’ll find a rich variety of the kinds of articles and information that exemplify a typical issue of the current magazine.

ESC Boston Is a Must

 

Tools and templates you can use to show the value of this event and justify your attendance.

Convince Your Boss That
ESC Boston Is a Must

Ready to register for Embedded Systems Conference (ESC) Boston but need help getting the OK from your boss? We have got some handy tools you can use.

Whether you want to attend the free expo or need budget approval for the conference, our Justification Toolkit has everything you need to get that  Yes! €

  • Approval request sample letters
  • Projected expenses worksheet
  • Expert tips and advice

Use promo code KCK when you register to save 20% on an ESC Boston Conference pass! Get the OK, then get your pass!

Bonus! Save More When You Bring the Team
Join forces with your colleagues to get the green light. For a limited time, you can get up to 25% off current pricing when you register a group of three or more.
UBM
Facebook Twitter
LinkedIn
#ESCbostonESC Boston c/o UBM
2901 28th St. #100
Santa Monica, CA 90405

esc-boston.com
© UBM 2018. All Rights Reserved. Privacy Policy

DC-DC Converters Serve Renewable Energy Applications

CUI’s Power Group has announced a line of DC-DC converters ranging from 5 W to 40 W. The AE series, available in board mount, chassis mount, and DIN rail mount configurations, offers power ratings of 5 W, 10 W, 15 W and 40 W with input voltages up to 1,500 Vdc and input ratio ranges up to 10:1. Designed for reliable operation in renewable energy applications such as solar power equipment, wind turbines, and electric charging stations, the new models feature 5,600 Vdc input to output isolation and a wide operating temperature range up to -40 to +70°C at full load. The series is also rated to operate at up to 5000 meters, allowing it to support remote, high altitude installations.

These new isolated DC-DC converters are housed in fully encapsulated packages for use in harsh environments with measurements beginning at 2.75″ x 1.89″ x 0.93″ (70 mm x 48 mm x 23.50 mm). Output voltage options of 5-, 9-, 12-, 15-, and 24-VDC are available, depending on the series, including efficiency ratings up to 84%. Protections for over voltage and over current as well as continuous short circuit with automatic recovery come as standard.

The AE series also carries EN 62109 safety approvals and complies with CISPR22/EN55022 Class A limits for conducted and radiated emissions, while the AE-UW series is designed to meet an additional UL 1741 safety standard. The various models in the AE series are available immediately with prices starting at $42.28 per unit at 100 pieces through distribution

CUI | www.cui.com

SMARC Module Features Hexa-Core i.MX8 QuadMax

By Eric Brown

iWave has unveiled a rugged, wireless enabled SMARC module with 4 GB LPDDR4 and dual GbE controllers that runs Linux or Android on NXP’s i.MX8 QuadMax SoC with 2x Cortex-A72, 4x -A53, 2x -M4F and 2x GPU cores.

iW-RainboW-G27M (front)

iWave has posted specs for an 82 mm x 50 mm, industrial temperature “iW-RainboW-G27M” SMARC 2.0 module that builds on NXP’s i.MX8 QuadMax system-on-chip. The i.MX8 QuadMax was announced in Oct. 2016 as the higher end model of an automotive focused i.MX8 Quad family.

Although the lower-end, quad-core, Cortex-A53 i.MX8M SoC was not fully announced until after the hexa-core Quad, we’ve seen far more embedded boards based on the
i.MX8M , including a recent Seco SM-C12

iW-RainboW-G27M (back)

SMARC module. The only other i.MX8 Quad based product we’ve seen is Toradex’s QuadMax driven Apalis iMX8 module. The Apalis iMX8 was announced a year ago, but is still listed as “coming soon.”

 

 

i.MX8 Quad block diagram (dashed lines indicate model-specific features) (click image to enlarge)

 

Like Rockchip’s RK3399, NXP’s i.MX8 QuadMax features dual high-end Cortex-A72 cores and four Cortex-A53 cores. NXP also offers a similar i.MX8 QuadPlus design with only one Cortex-A72 core.

The QuadMax clock rates are lower than on the RK3399, which clocks to 1.8 GHz (A72) and 1.2 GHz (A53). Toradex says the Apalis iMX8’s -A72 and -A53 cores will clock to 1.6 GHz and 1.2 GHz, respectively.

Close-up of i.MX8 QuadMax on iW-RainboW-G27M

Whereas the i.MX8M has one 266 MHz Cortex-M4F microcontroller, the Quad SoCs have two. A HIFI4 DSP is also onboard, along with a dual-core Vivante GC7000LiteXS/VX GPU, which is alternately referred to as being two GPUs in one or having a split GPU design.

iWave doesn’t specifically name these coprocessors except to list features including a “4K H.265 decode and 1080p H.264 enc/dec capable VPU, 16-Shader 3D (Vec4), and Enhanced Vision Capabilities (via GPU).” The SoC is also said to offer a “dual failover-ready display controller.” The CPUs, meanwhile, are touted for their “full chip hardware virtualization capabilities.”

Inside the iW-RainboW-G27M

Like iWave’s SMARC 2.0 form factor Snapdragon 820 SOM, the iW-RainboW-G27M supports Linux and Android, in this case running Android Nougat (7.0) or higher. (Toradex’s Apalis iMX8 supports Linux, and also supports FreeRTOS running on the Cortex-M4F MCUs.)

Like Toradex, iWave is not promoting the automotive angle that was originally pushed by NXP. iWave’s module is designed to “offer maximum performance with higher efficiency for complex embedded application of consumer, medical and industrial embedded computing applications,” says iWave.

Like the QuadMax based Apalis iMX8, as well as most of the i.MX8M products we’ve seen, the iW-RainboW-G27M supports up to 4 GB LPDDR4 RAM and up to 16 GB eMMC. iWave notes that the RAM and eMMC are “expandable,” but does not say to what capacities. There’s also a microSD slot and 256 MB of optional QSPI flash.

Whereas Apalis iMX8 has a single GbE controller, iWave’s COM has two. It similarly offers onboard 802.11ac Wi-Fi and Bluetooth (4.1). The Microchip ATWILC3000-MR110CA module, which juts out a bit on one side, is listed by Digi-Key as 802.11b/g/n, but iWave has it as 802.11ac.

Interfaces expressed via the SMARC edge connector include 2x GbE, 2x USB 3.0 host (4-port hub), 4x USB 2.0 host, and USB 2.0 OTG. Additional SMARC I/O includes 3x UART (2x with CTS & RTS), 2x CAN, 2x I2C, 12x GPIO, and single PCIe, SATA, debug UART, SD, SPI and QSPI

Media features include an HDMI/DP transmitter, dual-channel LVDS or MIPI-DSI, and an SSI/I2S audio interface. iWave also lists HDMI, 2x LVDS, SPDIF, and ESAI separately under “expansion connector interfaces.” Other expansion I/O is said to include MLB, CAN and GPIO.

The 5 V module supports -40 to 80°C temperatures. There is no mention of a carrier board.

Further information

No pricing or availability was listed for the iW-RainboW-G27M, but a form is available for requesting a quote. More information may be found on iWave’s iW-RainboW-G27M product page.

iWave | www.iwavesystems.com

This article originally appeared on LinuxGizmos.com on March 13.

Tuesday’s Newsletter: Microcontroller Watch

Coming to your inbox tomorrow: Circuit Cellar’s Microcontroller Watch newsletter. Tomorrow’s newsletter keeps you up-to-date on latest microcontroller news. In this section, we examine the microcontrollers along with their associated tools and support products.

Bonus: We’ve added Drawings for Free Stuff to our weekly newsletters. Make sure you’ve subscribed to the newsletter so you can participate.

Already a Circuit Cellar Newsletter subscriber? Great!
You’ll get your Microcontroller Watch newsletter issue tomorrow.

Not a Circuit Cellar Newsletter subscriber?
Don’t be left out! Sign up now:

Our weekly Circuit Cellar Newsletter will switch its theme each week, so look for these in upcoming weeks:

IoT Technology Focus. (4/17) Covers what’s happening with Internet-of-Things (IoT) technology–-from devices to gateway networks to cloud architectures. This newsletter tackles news and trends about the products and technologies needed to build IoT implementations and devices.

Embedded Boards.(4/24) The focus here is on both standard and non-standard embedded computer boards that ease prototyping efforts and let you smoothly scale up to production volumes.

Analog & Power. (4/1) This newsletter content zeros in on the latest developments in analog and power technologies including DC-DC converters, AD-DC converters, power supplies, op amps, batteries and more.

Eval Board for Half-Bridge SiC Power Modules

Analog Devices, in collaboration with Microsemi, has introduced a high power evaluation board for half-bridge SiC power modules with up to 1,200 V and 50 A at 200 kHz switching frequency. The isolated board is engineered to improve design reliability while also reducing the need to create additional prototypes—saving time, lowering costs, and decreasing time to market for power conversion and energy storage system designers.

The board can be used as the building block of more complex topologies, such as full-bridge or multi-level converters, for complete bench debugging of customer solutions. It can also function as a final evaluation platform or in converter-like configuration for full test and evaluation of Analog Devices’ ADuM4135 isolated gate driver with iCoupler digital isolation technology and LT3999 DC-DC driver in a high-power system.

The high-power evaluation board enables Microsemi’s SiC power modules to provide benefits such as a common test bench, higher power density for reduced size and cost, and isolated and conductive substrate and minimum parasitic capacitance for higher efficiency, performance, and excellent thermal management. These attributes make the board suitable for applications including electric vehicle (EV) charging, hybrid EV (HEV)/EV onboard charging, DC-DC converters, switched mode power supply, high-power motor control and aviation actuation systems, plasma/semi cap equipment, lasers and welding, MRIs and X-rays. The EV-MS4135PL1Z-UI Evaluation Board is priced at $495.

Analog Devices | www.analog.com

Microsemi | www.microsemi.com

Chassis Mounted 250-V EMC Filters Deliver 6 A to 30 A

TDK has announced the introduction of the TDK-Lambda RSEV series of EMC filters, with a voltage of 250 VAC/VDC and a choice of 6-, 10-, 16-, 20- and 30-A currents ratings. The RSEV has an integrated touch-resistant terminal block with recessed, captive screws. Hardware cannot be misplaced or dropped inside the system enclosure. The filters are suitable for general purpose use in a wide range of industrial and communications equipment.

The RSEV series is chassis mountable, or, with an optional kit, can be attached to the industry standard DIN rails. The maximum weight is 150 g for the 30 A model.  All models share a common package size of 85 mm x 39 mm x 30 mm (L x W x H), a 37% reduction in size compared to previous models (RSEN 30A). The filters have an operating ambient temperature rating of -25 to +85°C with derating above 55°C.

The series has a maximum earth leakage current of 1 mA at 250 VAC 60 Hz, is safety certified to the UL 1283, CSA C22.2 No.8 and EN 60939-3 standards with compliancy to the RoHS2 Directive.  A five-year warranty is standard.

Main features and benefits:

  • Captive Connection Screws
  • Integrated Terminal Block
  • Compact Size
  • DIN Rail Mounting Kit

TDK-Lambda | www.tdk-lambda.com

Component Tolerance

Accuracy Unmasked

We take for granted sometimes that the tolerances of our electronic components fit the needs of our designs. In this article, Robert takes a deep look into the subject of tolerances, using the simple resistor as an example. He goes through the math to help you better understand accuracy and drift along with other factors.

By Robert Lacoste

One of the last projects I worked on with my colleagues was a kind of high-precision current meter. It turned out to be far more difficult than anticipated, even with our combined experience totaling almost 100 years. Maybe this has happened with your projects too: You discover that, even when you’re not looking for top performance out of your electronic components, the accuracy and stability of those components can be pernicious. My topic this month is examining component tolerances. And, for simplicity, I will focus on the simplest possible electronic device: a resistor.

FIGURE 1 A very simple voltage divider. With these values, Uout will be 1 V with Uin=100 V

Let’s start with a basic application. Imagine that you have to design a voltage divider with a ratio of 1/100 (Figure 1). I will assume that the source impedance is very low and that the load connected on the output draws no current at all. With those parameters the calculations are very easy. You just need to know Ohm’s Law. Because the resistors are in series, the current circulating through the two resistors is:

Similarly, the output voltage is:

Given that the current I is the same in both equations, we get:

This circuit is indeed a voltage divider, with a ratio of R2/(R1+R2). We want a ratio of 1/100, so one resistor could be fixed arbitrarily and the second easily calculated. For example: R1=9,900 Ω and R2=100 Ω will do the job as:

Of course, you can easily simulate such a circuit with any SPICE-based circuit simulator if you wish. I personally used Proteus from Labcenter to draw and simulate the small schematic provided on Figure 1, and the output voltage is 1 V with 100 V applied on the input, as expected. As usual, I encourage you to reproduce these small examples with your preferred simulator: for example the free LT-Spice.

Now let’s talk about accuracy. You want your divider to be as precise as possible and therefore you want to buy reasonably accurate resistors. But what if your budget is constrained? Will you use a high accuracy resistor for R1 (9,900 Ω)? Or for R2 (100 Ω)? Or for both? The good answer is both. In that case, a 1% error on either R1 or R2 gives close to a 1% error of the output voltage, as shown in Figure 2. Even if R1 has a stranger value than R2—9,900 Ω vs. 100 Ω—their accuracy is just as critical.

Figure 2
A 1% error either on the top or bottom resistors will induce a roughly 1% error on the output. That would not be the case for other division ratios.

Maybe you think this is too obvious? In that case I will give you another exercise: What happens with a divide-by-2 circuit using two resistors of the same value? Do the calculation or simulate it and you will find that both resistors have still the same impact on accuracy. But now a 1% error on one of the resistors has only a 0.5% impact on the output voltage. That means you could buy slightly less expensive resistors for the same overall precision! In fact, the higher the division ratio, the higher is the impact of each resistor on the overall accuracy.

E Series Resistors

Let’s go back to the 1/100 divider example. If you want to build it and look for a
9,900-Ω resistor, you will have some difficulties because nobody sells them.. …

Read the full article in the April 333 issue of Circuit Cellar

Don’t miss out on upcoming issues of Circuit Cellar. Subscribe today!
Note: We’ve made the October 2017 issue of Circuit Cellar available as a free sample issue. In it, you’ll find a rich variety of the kinds of articles and information that exemplify a typical issue of the current magazine.

Tiny i.MX8M Module Focuses on Streaming Media

By Eric Brown

Innocomm announced a 50 mm x 50 mm “WB10” module with an NXP i.MX8M Quad SoC, 8 GB eMMC, Wi-Fi-ac, BT 4.2, GbE, HDMI 2.0 with 4K HDR and audio I/O including SAI, SPDIF and DSD512.Among the many embedded products announced in recent weeks that run NXP’s 1.5 GHz, Cortex-A53-based i.MX8M SoC, Innocomm’s 50 mm x 500 mm WB10 is one of the smallest. The top prize goes to Variscite’s SODIMM-style, 55 mm x 30 mm DART-MX8M. Like Emcraft’s 80 mm x 60mm i.MX 8M SOM, the home entertainment focused WB10 supports only the quad-core i.MX8M instead of the dual-core model. Other i.MX8M modules include Compulab’s 68 mm x 42mm CL-SOM-iMX8.

WB10 (above) and NXP i.MX8M block diagram (below)
(click images to enlarge)
No OS support was listed, but all the other i.MX8M products we’ve seen have either run Linux or Linux and Android. The i.MX8M SoC incorporates a Vivante GC7000Lite GPU and VPU, enabling 4K HEVC/H265, H264, and VP9 video decoding with HDR. It also provides a 266MHz Cortex-M4 core for real-time tasks, as well as a security subsystem.

The WB10 module offers only 2 GB LPDDR4 instead of 4 GB for the other i.MX8M modules, and is also limited to 8GB eMMC. You do, however, get a GbE controller and onboard 802.11 a/b/g/n/ac with MIMO 2×2 and Bluetooth 4.2.

The WB10 is designed for Internet audio, home entertainment, and smart speaker applications, and offers more than the usual audio interfaces. Media I/O expressed via its three 80-pin connectors include HDMI 2.0a with 4K and HDR support, as well as MIPI-DSI, 2x MIPI-CSI, SPDIF Rx/Tx, 4x SAI and the high-end DSD512 audio interface.

WB10 block diagram (above) and WB10 mounted on optional carrier board (below)
(click images to enlarge)

You also get USB 3.0 host, USB 2.0 device, 2x I2C, 3x UART and single GPIO, PWM, SPI, and PCIe interfaces. No power or temperature range details were provided. The WB10 is also available with an optional, unnamed carrier board that is only slightly larger than the module itself. No more details were available. Further information

No pricing or availability information was provided for the WB10. More information may be found on Innocomm’s WB10 product page.

Innocomm | www.innocomm.com

This article originally appeared on LinuxGizmos.com on March 6.

Advantech Joins Amazon’s AWS Partner Network

Advantech has joined the Amazon Web Services (AWS) Partner Network (APN) as Standard Technology Partner. As an APN Standard Technology Partner, Advantech provides a comprehensive range of wireless sensors and edge intelligence computers with complete IoT software solutions on AWS. Embedded developers can connect devices to a range of services offered on AWS in order to build scalable, global, and secure IoT applications, bringing computing capabilities to edge devices to several domain-focused vertical markets such as smart city, smart manufacturing and smart energy markets.

Advantech’s WISE-1520 Wireless Sensor Node (shown) is on Amazon FreeRTOS so that customers can easily and securely connect small devices and sensors directly to AWS or to powerful edge devices running AWS Greengrass, thus allowing them to collect data for their IoT applications. As the first wireless sensor node for the M2.COM family, the WISE-1520 comes with an Arm Cortex-M4 processor and low-power Wi-Fi connectivity, providing full compatibility with existing Wi-Fi infrastructure.

Advantech also offers the EIS-D210 Edge Intelligence Server, which is equipped with an Intel Celeron Processor N3350 and is compatible with AWS Greengrass core, thus ensuring that IoT devices can respond quickly to local events, interact with local resources, operate with intermittent connections, and minimize the cost of transmitting IoT data to the cloud. In addition to supporting field protocols(MQTT/OPC/Modbus) for sensor/device data acquisition, the EIS-D210 can be used with the Advantech IoT SDK for wireless sensor (Wi-Fi, LoRa, Zigbee) data integration. Furthermore, the EIS-D210 comes pre-integrated with Advantech’s WISE-PaaS/EdgeSense software solution, allowing users to incorporate sensor data aggregation, edge analytics, and cloud applications for fast and easy real-time operational intelligence. This EIS provides a range of connectivity options with excellent data handling and networking connection capabilities for various IoT applications.

Advantech’s EPC-R4760 IoT gateway, powered by the Qualcomm Arm Cortex-A53 APQ8016 platform, provides a unique combination of power and performance. The system also integrates abundant wireless solutions including Wi-Fi, BT, GPS, and extended 3G/LTE connectivity. For OS support, the EPC-R4760 can run Debian Linux, Yocto Linux, Ubuntu Linux, Android, and Windows 10 IoT Core, and it also supports AWS Greengrass, which gives users tremendous flexibility by allowing them to create AWS Lambda functions that can be validated on AWS and then be easily deployed to devices.

Advantech’s UTX-3117 IoT gateway is compatible with AWS Greengrass and Wind River Pulsar and, in addition to having a small footprint, it offers real-time security and supports various protocols that are needed to run IoT applications seamlessly across both AWS and on local devices or sensor nodes. In addition, by equipping it with a LoRa solution, the UTX-3117 offers a wide range of wireless connection options for controlling and collecting data from devices and sensor nodes. With these solutions, the UTX-3117 IoT gateway is ideal for smart energy applications. For example, it can collect solar panel and solar radiation data in real time via LoRa, and with AWS Greengrass built in, it can analyze the data and adjust the angle of solar panels to follow the sun and thereby maximize the effectiveness of the solar panels. AWS Greengrass can also be employed to analyze weather data so that the panels can be adjusted to prevent damage from elements such as strong wind or hail.

Advantech | www.advantech.com

 

Tuesday’s Newsletter: Analog & Power

Coming to your inbox tomorrow: Circuit Cellar’s Analog & Power newsletter. Tomorrow’s newsletter content zeros in on the latest developments in analog and power technologies including ADCs, DACs, DC-DC converters, AD-DC converters, power supplies, op amps, batteries and more.

Bonus: We’ve added Drawings for Free Stuff to our weekly newsletters. Make sure you’ve subscribed to the newsletter so you can participate.

Already a Circuit Cellar Newsletter subscriber? Great!
You’ll get your Analog & Power newsletter issue tomorrow.

Not a Circuit Cellar Newsletter subscriber?
Don’t be left out! Sign up now:

Our weekly Circuit Cellar Newsletter will switch its theme each week, so look for these in upcoming weeks:

Microcontroller Watch. (4/10) This newsletter keeps you up-to-date on latest microcontroller news. In this section, we examine the microcontrollers along with their associated tools and support products.

IoT Technology Focus. (4/17) Covers what’s happening with Internet-of-Things (IoT) technology–-from devices to gateway networks to cloud architectures. This newsletter tackles news and trends about the products and technologies needed to build IoT implementations and devices.

Embedded Boards.(4/24) The focus here is on both standard and non-standard embedded computer boards that ease prototyping efforts and let you smoothly scale up to production volumes.

 

Commell Launches its First ARM-Based Pico-ITX

By Eric Brown

Commell has announced the LP-150, a Rockchip RK3128 based Pico-ITX SBC that appears to be its first ARM-based embedded board of any kind. The 100 mm x 72 mm LP-150 is the only ARM-based SBC out of the many dozens of mostly Intel-based boards listed on Commell’s SBC page.

Shipping with Android 4.4.4, but also supporting Linux, the LP-150 is intended primarily for imaging, machine vision and digital signage applications. Other Commell Pico-ITX SBCs include its Intel Braswell based LP-176.

Commell LP-150

Rockchip’s quad-core, Cortex-A7 RK3128 hasn’t seen as much uptake in the embedded world as the quad -A17 RK3288, which is found on hacker boards such as the Firefly-RK3288 Reload or the high-end, hexa-core RK3399, which has appeared on numerous recent products such as OpenEmbed’s em3399 module or Aaeon’s RICO-3399 PICO-ITX SBC. The only RK3128-based SBC we can recall is the open spec Firefly-FirePrime S.

The LP-150 SBC has a fairly modest feature set, with only 512 MB DDR3. Yet, it offers a few features you don’t typically find on x86 Pico-ITX SBCs like eMMC storage (8GB) and built-in Wi-Fi. Media features include an HDMI 1.4 port limited to HD resolution and an LVDS interface with capacitive touchscreen support. You also get CVBS inputs and outputs and audio I/O headers.

The LP-150 is further equipped with a GbE port and USB 2.0 host and OTG ports. There is also an RTC with battery, and a smattering of RS-232, UART, and GPIO interfaces.

Specifications listed for the Commell LP-150 include:

  • Processor — Rockchip RK3128 (4x Cortex-A7 @ 1.3 GHz); Mali-400 MP2 GPU with OpenGL ES1.1 and 2.0, OpenVG1.1
  • Memory — 512 MB DDR3
  • Storage — 8 GB eMMC; microSD slot
  • Display:
    • HDMI 1.4 port for up to 1080p
    • Single-channel 18/24-bit LVDS for up to 1280 x 720 displays or up to 1024 x 600 cap. touchscreens
    • LCD/LVDS panel and inverter connectors
    • CVBS in/out
  • Wireless — Wi-Fi with SMA antenna
  • Networking — Gigabit Ethernet port (RTL8211E)
  • Other I/O:
    • USB 2.0 host port with support for 4-port hub
    • USB 2.0 OTG port
    • 2x RS232 interfaces
    • 3x UART
    • Audio line-out, mic-in headers (Rockchip codec)
    • GPIO header
  • Other features — Power, recovery, reset buttons; RTC with lithium battery; LED
  • Operating temperature — 0 to 70°C
  • Power — DC input 5 V
  • Dimensions — 100 mm x 72 mm (Pico-ITX)
  • Operating system — Android 4.4.4; Linux also supported

Further information

No pricing or availability information was provided for the LP-150. More information may be found on Commell’s LP-150 product page.

Commell |  www.commell.com.tw

This article originally appeared on LinuxGizmos.com on March 6.

Movidius AI Acceleration Technology Comes to a Mini-PCIe Card

By Eric Brown

UP AI Core (front)

As promised by Intel when it announced an Intel AI: In Production program for its USB stick form factor Movidius Neural Compute Stick, Aaeon has launched a mini-PCIe version of the device called the UP AI Core. It similarly integrates Intel’s AI-infused Myriad 2 Vision Processing Unit (VPU). The mini-PCIe connection should provide faster response times for neural networking and machine vision compared to connecting to a cloud-based service.

UP AI Core (back)

The module, which is available for pre-order at $69 for delivery in April, is designed to “enhance industrial IoT edge devices with hardware accelerated deep learning and enhanced machine vision functionality,” says Aaeon. It can also enable “object recognition in products such as drones, high-end virtual reality headsets, robotics, smart home devices, smart cameras and video surveillance solutions.”

 

 

UP Squared

The UP AI Core is optimized for Aaeon’s Ubuntu-supported UP Squared hacker board, which runs on Intel’s Apollo Lake SoCs. However, it should work with any 64-bit x86 computer or SBC equipped with a mini-PCIe slot that runs Ubuntu 16.04. Host systems also require 1GB RAM and 4GB free storage. That presents plenty of options for PCs and embedded computers, although the UP Squared is currently the only x86-based community backed SBC equipped with a Mini-PCIe slot.

Myriad 2 architecture

Aaeon had few technical details about the module, except to say it ships with 512MB of DDR RAM, and offers ultra-low power consumption. The UP AI Core’s mini-PCIe interface likely provides a faster response time than the USB link used by Intel’s $79 Movidius Neural Compute Stick. Aaeon makes no claims to that effect, however, perhaps to avoid

Intel’s Movidius
Neural Compute Stick

disparaging Intel’s Neural Compute Stick or other USB-based products that might emerge from the Intel AI: In Production program.

It’s also possible the performance difference between the two products is negligible, especially compared with the difference between either local processing solutions vs. an Internet connection. Cloud-based connections for accessing neural networking services suffer from reduced latency, network bandwidth, reliability, and security, says Aaeon. The company recommends using the Linux-based SDK to “create and train your neural network in the cloud and then run it locally on AI Core.”

Performance issues aside, because a mini-PCIe module is usually embedded within computers, it provides more security than a USB stck. On the other hand, that same trait hinders ease of mobility. Unlike the UP AI Core, the Neural Compute Stick can run on an ARM-based Raspberry Pi, but only with the help of the Stretch desktop or an Ubuntu 16.04 VirtualBox instance.

In 2016, before it was acquired by Intel, Movidius launched its first local-processing version of the Myriad 2 VPU technology, called the Fathom. This Ubuntu-driven USB stick, which miniaturized the technology in the earlier Myriad 2 reference board, is essentially the same technology that re-emerged as Intel’s Movidius Neural Compute Stick.

UP AI Core, front and back

Neural network processors can significantly outperform traditional computing approaches in tasks like language comprehension, image recognition, and pattern detection. The vast majority of such processors — which are often repurposed GPUs — are designed to run on cloud servers.

AIY Vision Kit

The Myriad 2 technology can translate deep learning frameworks like Caffe and TensorFlow into its own format for rapid prototyping. This is one reason why Google adopted the Myriad 2 technology for its recent AIY Vision Kit for the Raspberry Pi Zero W. The kit’s VisionBonnet pHAT board uses the same Movidius MA2450 chip that powers the UP AI Core. On the VisionBonnet, the processor runs Google’s open source TensorFlow machine intelligence library for neural networking, enabling visual perception processing at up to 30 frames per second.

Intel and Google aren’t alone in their desire to bring AI acceleration to the edge. Huawei released a Kirin 970 SoC for its Mate 10 Pro phone that provides a neural processing coprocessor, and Qualcomm followed up with a Snapdragon 845 SoC with its own neural accelerator. The Snapdragon 845 will soon appear on the Samsung Galaxy S9, among other phones, and will also be heading for some high-end embedded devices.

Last month, Arm unveiled two new Project Trillium AI chip designs intended for use as mobile and embedded coprocessors. Available now is Arm’s second-gen Object Detection (OD) Processor for optimizing visual processing and people/object detection. Due this summer is a Machine Learning (ML) Processor, which will accelerate AI applications including machine translation and face recognition.

Further information

The UP AI Core is available for pre-order at $69 for delivery in late April. More information may be found at Aaeon’s UP AI Core announcement and its UP Community UP AI Edge page for the UP AI Core.

Aaeon | www.aaeon.com

This article originally appeared on LinuxGizmos.com on March 6.

STMicroelectonics Sensors Achieve Validation for Alibaba IoT OS

STMicroelectronics has announced the validation of its LSM6DSL 6-axis inertial sensor and LPS22HB pressure sensor for Alibaba IoT’s ecosystem, which enables users to create complete IoT nodes and gateway solutions with better time to market.

Announced last year, AliOS Things is a light-weight embedded operating system for IoT, developed by Alibaba. The company recently announced the release of AliOS Things v1.2, which includes a sensor-based component called uData. The ST sensors that have passed the AliOS validation have been integrated in uData. The two companies are cooperating on the development of IoT systems that aim to improve end-user experiences.

The LSM6DSL (shown) is a system-in-package featuring a 3D digital accelerometer and a 3D digital gyroscope that operates at 0.65 mA in high-performance mode and enables always-on low-power features for an optimal motion experience for the consumer. High robustness to mechanical shock makes the LSM6DSL well suited for the creation and manufacturing of reliable products. The LSM6DSL supports main OS requirements, offering real, virtual and batch sensors with 4 KB for dynamic data batching.

STMicroelectronics has announced the validation of its LSM6DSL 6-axis inertial sensor and LPS22HB pressure sensor for Alibaba IoT’s ecosystem, which enables users to create complete IoT (Internet-of-Things) nodes and gateway solutions with better time to market.

The LPS22HB is an ultra-compact piezoresistive absolute pressure sensor that functions as a digital output barometer. Dust-free and water-resistant by design, the sensor enables high accuracy and low-power operation. It is available in full-mold package with silicon cap and six 20µm holes guaranteeing sensor moisture resistance, relative accuracy of pressure measurement 0.1 mbar, and very low power consumption (12 µA in low-noise mode).

STMicroelectronics | www.st.com