Mustool MT13S is a relatively inexpensive 2-in-1 thermal imager and multimeter with a 2.8-inch touchscreen display and an IR camera with a 192×192 resolution.

Thermal cameras used to be quite expensive, but in recent years, we’ve seen cheaper models such as the HT-102 thermal camera for Android smartphones and M5Stack T-Lite Wi-Fi thermal camera, but those integrate fairly small 32×32 and 32×24 pixels IR arrays, The Mustool MT13S provides a higher resolution 192×192 thermal camera and doubles as a 10,000-count multimeter for about $166 plus shipping on Banggood or around $177 including shipping on Aliexpress.

Mustool MT13S specifications:

• Thermal imaging
  • Sensor – Uncooled focal plane
  • Image capture frequency – 20Hz
  • Thermal imaging resolution – 192 x 192
  • Display image resolution – 240 x 240
  • Field of view (FoV) – 50.0(H) × 50(V)/72.1(D)
  • Emissivity – 0.1-0.99 is tunable and 0.95 is the default
  • Temperature Range – -20°C to +550°C
  • Accuracy – 0.1°C/0.1F
  • Measurement error
    • > 0°C +/- 2°C or +/- 2%
    • <= 0°C +/- 5°C or +/-5 %
  • Mode – Automatic gain
  • Color palettes –  Iron Red, Rainbow, Fusion, White Heat, White Heat Highlights
• Multimeter
  • Input – DC up to 1000V, AC up to 750V
  • Resistance up to 99.99MΩ
  • Capacitance up to 99.99mF
  • Duty cycle measurement range – 0.1% ~ 99.9%
  • Diode measurement range – 0V ~ 3V
  • On-off test maximum resistance – 999.9Ω
  • 4-digit shown, updated about 3 times per second
• Data Storage – 3.5MB for BMP files
• Display – 2.8-inch resistive touchscreen display with 480 x 320 resolution
• USB – 1x USB Type-C port for charging and data transmission to the host
• Power Supply – 3.7V/850 mAh built-in lithium battery; support for auto screen off and auto power off
• Dimensions – 134 x 69 x 25 mm
• Weight – 130 grams
• Temperature Range – Operating: 0 to 50°C; storage: -20 to 60°C
• Humidity – < 85%RH (non-condensing)

The thermal imager and multimeter ships with a USB Type-C cable and a user manual. At first, I was confused since I could not see the leads and connectors, but those are also included in the package and connected to the bottom side of the device based on an unboxing from one customer on Banggood. I was unable to find a detailed review for the tool, but feedback from users looks genuine and they all seem to appreciate the thermal imager function. User photos also show the device is smaller than I would have expected.

The MT13S is at least the second generation model improving on the earlier MT12S model with higher thermal imaging resolution, wider temperature measurement range, and better multimeter function with 10,000 counts instead of 4,000 counts.

It’s unclear how the USB data transmission to the host works, as I could not find any download link for a PC program. Maybe the MT13S shows as a USB drive allowing the user to download BMP files and CSV data for further processing.

The Mustool MT13S can be purchased for about $165.99 plus shipping on Banggood and  $177.23 including shipping on Aliexpress.

The post $166 Mustool MT13S thermal imager doubles as multimeter appeared first on CNX Software - Embedded Systems News.

STMicro ST60A3H0 and ST60A3H1 are short-range 60 GHz transceiver ICs that tunnel eUSB2, I2C, SPI, UART, and GPIO signals and aim to replace USB and other cables in consumer devices such as digital cameras, wearables, portable hard drives, and small gaming terminals. They should also find their way into industrial applications such as rotating machinery where cable use may be challenging.

The smaller ST60A3H0 chip provides more flexibility and requires an external antenna, while the ST60A3H1 chip is a fully integrated solution with a built-in linear antenna. Both are capable of USB 2.0 speeds of up to 480 Mbps and support UART, GPIO, and/or I2C signals so they are not limited to USB cables and can be used in a range of applications.

ST60A3H0 and ST60A3H1 key features and specifications:

• 60 GHz V-Band transceiver for short-range contactless connectivity up to 480 Mbit/s
• eUSB2, UART, GPIO, or I2C RF tunneling
• Low power consumption (typical values with a single 1.8 V supply):
  • eUSB2 Rx/Tx – 110/130 mW
  • UART/GPIO/I2C – 90 mW
  • Standby – 23 μW
• Optimized BOM without external matching network and clock references. A reference clock may be used at one end of the RF link to comply with specific regional regulation
• ST60A3H0 features
  • Integrated full RF transceiver operating in Half-Duplex mode
  • 34 dB typical total link budget
  • 50 Ω single-ended nominal RF input/output impedance with recommended PCB transition
  • Single 1.8 V supply or dual supply 1.8 V (analog/RF) and 1.2 V (digital/GPIO)
  • Package – VFBGA 2.2 x 2.6 x 0.8 mm, 30 balls, 5×6 array, 0.4 mm
• ST60A3H1
  • Integrated full RF transceiver and linear polarization antenna, operating in Half-Duplex mode
  • 42 dB typical total link budget, up to 5 cm free-space propagation loss
  • Single 1.8 V supply
  • Package – VFBGA 2.9 x 4.1 x 0.8 mm, 23 balls, 0.4 mm pitch
• Temperature Range – -20 to +85°C

Detailed technical data and evaluation kits are available but as explained in the press release getting access to those requires signing a non-disclosure agreement, so there’s limited public information. As I understand it there should be one transceiver in the device and another one connected to the host and data transfer can occur over a few centimeters distance. Charging does not seem to be handled by the solution.

The ST60A3H0 and ST60A3H1 60 GHz transceivers look to be especially useful for waterproof devices since they remove the need for cabling while keeping USB 2.0 compatibility, but should also enable thinner devices and support for rotating devices as illustrated in the video demo below with the earlier ST60A2 industrial-grade transceiver supporting SLVS or GPIO tunneling.

STMicro says the ST60A3H0 and ST60A3H1 are in mass production with a life-cycle of at least 10 years. Samples are available now at $5.00 and up. A few more details may be found on the respective product pages.

The post STMicro ST60A3H0 and ST60A3H1 60 GHz transceiver ICs aim to replace USB cables appeared first on CNX Software - Embedded Systems News.

Arduino board clones have been around for many years, but I don’t think I have ever seen clones of the new Renesas-based Arduino boards so far. Waveshare changes that with the R7FA4 PLUS A that clones with Arduino UNO R4 Minima, and the R7FA4 PLUS B board duplicating the Arduino UNO R4 WiFi.

The Waveshare boards are not 100% clones with some small differences in the PCB layout, support for 5V and 3.3V shields, an additional 6-pin “power output header” with 5V, 3.3V, and GND signals, and a USB communication jumper to select between the Espressif ESP32-S3 and Renesas RA4M1 microcontrollers.

Waveshare R7FA4 PLUS A and B specifications:

• Microcontroller – Renesas RA4M1 Arm Cortex-M4F MCU @ 48 MHz with 32KB SRAM, 256KB flash
  
• Wireless (B model only) – ESP32-S3-MINI-1 module based on ESP32-S3 dual-core Xtensa LX7 microcontroller with 512KB SRAM, 384KB ROM, WiFi 4 and Bluetooth 5.0 connectivity, PCB antenna
• Display (B model only) –  12×8 LED matrix (red)
• USB – 1 x USB Type-C port for power and programming
• Expansions
  • Arduino UNO headers with Pins
    • 14x digital I/Os
    • 13x LED pins
    • Analog – 6x analog input pin, 2x 12-bit analog DAC
    • 6x PWM
    • 1x UART, 1x I2C, 1x SPI
    • CAN Bus support
    • I/O Voltage – 5V and 3.3V (Not available on Arduino boards, but works on Waveshare board – TBC)
    • I/O current – 8 mA
  • R7FA4 PLUS B only
    • Qwiic I2C connector for expansion modules.
    • 3-pin header with an “OFF” pin to turn off the board and a “VRTC” pin to keep the internal Real-Time Clock powered and running.
• Debugging and programming – 6-pin ICSP header; R7FA4 PLUS A only: 10-pin SWD header
• Misc – Reset button, Power LED, USB communication jumper
• Power Supply
  • Input voltage – 6 to 24V via a power barrel jack or Vin, 5V via USB-C port
  • Output header – 6-pin header with 5.5V and 3.3V (not found on the original Arduino UNO R4 boards)
• Dimensions – 68.6 x 53.4mm

Both the Waveshare R7FA4 Plus A and B boards are functionally equivalent to the Arduino UNO R4 Minima and WiFi boards, except for the additional support for 3.3V Arduino shields (on top of 5V ones) and the extra power output header. That means software for the official Arduino boards will run on the “clones”, but Waveshare still provides a wiki for each board with further hardware documentation, the PDF schematics, and a guide showing how to get started with the Arduino IDE.

Pricing for the Arduino UNO R4 “clones” is not that different from the original boards with the Waveshare R7FA4 Plus A going for $21.59 including shipping on Aliexpress and the Waveshare R7FA4 Plus B selling for $27.89 shipped. Both boards are also available on Amazon for respectively $21.99 (Model A) and $29.99 (Model B).

The post Waveshare R7FA4 PLUS A and B boards are clones of the Arduino UNO R4 Minima and WiFi appeared first on CNX Software - Embedded Systems News.

The 52Pi NVdigi is another PCIe expansion board for the Raspberry Pi 5 which integrates HiFiBerry Digi+ to provide high-quality S/PDIF output. It also features an M.2 PCIe x1 slot that supports NVMe 2242/2230 SSDs. Furthermore, it offers an optical output (TOSLink) and an RCA output for versatile audio connections.

The HiFiBerry Digi+ is a high-quality S/PDIF output for the Raspberry Pi. It uses the I2S sound port that connects directly to the CPU without the need for an additional USB conversion. It supports sample rates up to 192kHz/24bit.

52Pi NVdigi Expansion Board Specification:

• HiFiBerry Digi+ Integration – Provides high-quality S/PDIF output for Raspberry Pi 5.
• Direct I2S Connection – Connects directly to the CPU via the I2S sound port for optimal audio.
• High-Resolution Audio – Supports sample rates up to 192kHz and 24-bit depth for immersive audio.
• Multiple Audio Outputs – Features both optical (TOSLink) and electrical (RCA) outputs.
• M.2 PCIe x1 Slot – For NVMe 2242/2230 SSDs, with PCIe 3.0 support.
• Applications – Ideal for audio enthusiasts and users seeking expanded storage for projects.
• Weight and Dimensions – 20 grams, 65mm x 56mm.

Previously we’ve seen HATs for SSDs like the Mcuzone MPS2280 M.2 NVMe HAT, Mcuzone MP4GM 4G LTE HAT, and the Sixfab 5G Modem Kit along with HATs for Wi-Fi 7 and Google TPU. However, this is the first time we have seen a HAT that combines audio support with M.2 storage.

The company mentions that this device is perfect for those who love high-quality sound and need more storage for their Raspberry Pi 5 projects. Mostly, this little HAT targets those who want to build their own online media center. The company also provides a wiki page with further hardware details and instructions to enable it in Raspberry Pi Official OS.

The package includes one 52Pi NVDigi board, a 40-pin PC104 header, a 40mm PCIe FFC cable, four M2.5 x 15mm copper pillars, eight M2.5 x 4mm flat head screws, and an M2.5 screwdriver, providing all necessary components for setup. The 52Pi NVdigi Extension Adapter Board is priced at $39.99 and it’s available for pre-sale on the 52Pi online store.

The post 52Pi NVdigi Expansion Board for Raspberry Pi 5 combines HiFiBerry Digi+ audio output with M.2 PCIe x1 slot appeared first on CNX Software - Embedded Systems News.

Banana Pi BPI-M6 is a credit-card single board computer based on SenaryTech SN3680 SoC comprised of a quad-core Arm Cortex-A73 processor, an Arm Cortex-M3 real-time core, an Imagination GE9920 GPU, and an NPU delivering up to 6.75 TOPS.

The board ships with 4GB LPDDR4 RAM and 16GB eMMC flash. Its layout is fairly similar to the one of the Raspberry Pi 4 with four USB ports, Gigabit Ethernet, a 40-pin GPIO header, a USB Type-C port for power, and two micro HDMI ports. However, only one of those is for HDMI output, as the second is for HDMI input, and there’s also an M.2 Key-E socket for expansion.

Banana Pi BPI-M6 specifications:

• SoC – SenaryTech SN3680 (also known as Synaptics VS680) with
  • CPU – Quad-core Arm Cortex-A73 processor up to 2.1GHz
  • MCU – Arm Cortex-M3 real-time security core @ 250MHz
  • GPU – Imagination PowerVR Series9XE GE9920 GPU
  • VPU –
    • 4Kp60 H265, H264, VP9, ​​VP8, AV1, MPEG-2 video decoding
    • Dual 1080p60 H.264/VP8 video encoding
  • NPU – Up to 6.75 TOPS
  • Package – FCBGA, 17mm x 17mm
  • 12nm manufacturing process
• System Memory – 4 GB LPDDR4
• Storage
  • 16GB eMMC flash (option up to 64GB)
  • MicroSD card slot
  • SPI flash
• Video & Audio I/F
  • Micro HDMI 2.1 output up to 4Kp60 with HDR, CEC, EDID
  • MIPI DSI interface
  • Micro HDMI input
•  Networking
  • Gigabit Ethernet RJ45 port
  • Optional WiFi via USB dongle
• USB – 4x USB 3.0 ports
• Expansion
  • M.2 Key E socket (PCIe + MIPI CSI)
  • 40-pin header with up to 28x GPIO, UART, I2C, SPI, PWM, and power signals (+5V, +3.3V and GND)
• Misc
  • SPI BOOT, UBOOT, and Reset buttons
  • Power and Activity LEDs
• Power Supply – 5V/3A via USB Type-C port
• Dimensions –  92 x 60mm
• Weight – 48grams

Banana Pi provides Android and Ubuntu 20.04 images for the board which you’ll find in the wiki along with hardware documentation, a Linux SDK with Kernel 5.4 and Buildroot 2019.10, the config file for the Armbian build system, and instructions to use the SenarySocSystemTool flashing tool.

The VideoSmart VS680 is shown to score 29.90 points in the AI benchmark rankings for IoT processors which shows the NPU is a Vivante VIP9000. It has a higher score than the Amlogic A311D (21.9) and Rockchip RK3566 (14.1), but not quite as good as the better-supported Rockchip RK3588S (95.7) with a (slower in theory) 6 TOPS AI accelerator. For reference, the MediaTek Genio 1200 processor comes in third with 151 points in that list, only outperformed by Qualcomm Snapdragon SA8295P (161) and Mediatek MT8195 (Kompanio 1200) with 165 points.

Banana Pi shared some demo videos showing the AI capabilities of the board, but I was unable to find documentation and resources to make use of the Vivante VIP9000 NPU in the VS680 SoC. Synaptics mentions support for the Synap AI framework, but the company is not exactly known for releasing development tools publicly.  Note that the Vivante NPU in the Amlogic A331D SoC recently got Etnaviv open-source driver support, so the VS680 may end up being supported as well, although that will likely depend on the developer community interest rather than Banana Pi working on it…

The Banana Pi BPI-M6 is sold on Aliexpress for $74.74 plus shipping and on Amazon for $84.55 after ticking the box to get a 5% discount.

Updated: This post was initially published on November 22, 2022, when Banana Pi unveiled the board and updated following the availability of the BPI-M6 SBC on Amazon and Aliexpress

The post Banana Pi BPI-M6 SBC features SenaryTech SN3680 quad-core Cortex-A73 AI processor appeared first on CNX Software - Embedded Systems News.

We’ve already covered a range of ESP32-C6 boards, but none supporting Ethernet and PoE so far, and the ESP32-C6-Bug board brings that to the table thanks to the Esp32-Bug-Eth shield with a W5500 Ethernet chip, an RJ45 jack and a PoE power module.

Like other ESP32-C6 devices, the little board supports Wi-Fi 6, Bluetooth LE 5, as well as Thread and Zigbee through its 802.15.4 radio, but it also integrates some other interesting features such as castellated holes for easy soldering on a carrier board and support for LiPo batteries with built-in battery charging and protection circuits.

ESP32-C6-Bug board specifications:

• SoC – ESP32-C6FH4
  • MCU cores
    • 32-bit RISC-V core @ 160 MHz
    • 32-bit RISC-V core @ 20 MHz low-power coprocessor can run tasks even when the main system is in deep sleep state
  • Memory – 512 KB SRAM
  • Storage – 4 MB Flash
  • Wireless – WiFi 6, Bluetooth LE 5, and 802.15.4 radio (Zigbee, Thread, etc…)
• USB – 1x USB Type-C port for power, programming, and data
• I/Os – 2x 12-pin headers with through and castellated holes
  • Up to 19x GPIOs
  • SPI, UART, I2C, I2S, PWM, SDIO, Motor Control PWM, 12-bit ADC, etc…
• Misc
  • User-controlled LEDs
  • External 32.768 kHz RTC oscillator and 40 MHz oscillator
  • Reset and user-controlled buttons
  • Antenna – PCB antenna
• Power Supply
  • 5V via USB-C port
  • LiPo battery support with
    • Under-voltage and reverse-polarity protection
    • On-board battery charging and level measurement w/ indicator LED
    • 20 uA deep sleep power consumption (with timer wake-up)
  • 700 mA low-noise LDO
• Dimensions – Small (and breadboard compatible)

While the board can be used standalone, some users will want to combine it with the Esp32-Bug-Eth shield to add both Ethernet and PoE support to create a tiny IoT gateway with WiFi 6, BLE, Thread, Zigbee, and Ethernet.

Esp32-Bug-Eth add-on board features:

• Wiznet W5500 Ethernet module
• USB – 1x USB-C port supporting both power and data
• Expansion – STEMMA-QT connector for connecting peripherals
• Power Supply
  • 5V via USB-C port
  • Isolated PoE support provided via SDAPO DP1435-5V module

The ESP32-C6-Bug can be programmed with the ESP-IDF framework or the Arduino IDE with various examples for the latter available on GitHub namely a blinky sample, an Ethernet sample to check wired connection when used in combination with the Esp32-Bug-Eth shield, an I2C OLED display sample, and a telegram bot pushing BMP280 sensor to Telegram over its Ethernet connection. At this time, Zigbee and Thread connectivity requires using the ESP-IDF, and it’s not implemented into the ESP32 Arduino core.

Hardware documentation including a datasheet, PDF schematics, the bill-of-materials (BoM), and 3D models can be found on a separate GitHub repository. Prokyber s.r.o also creates two 3D printable enclosures for the ESP32-C6-Bug board only and the combo with the Ethernet shield that you’ll find on Thingiverse.

Prokyber s.r.o has launched the ESP32-C6-Bug board on Crowd Supply with a $1,500 funding goal. Rewards start at $29 for the ESP32-C6-Bug board only, and the Esp32-Bug-Eth shield adds an extra $34, meaning a complete system would cost $63 before shipping which may make the solution a hard sell. Shipping adds $8 to the US, and $18 to the rest of the world, and backers should expect their perks to ship by August 2024 as long as there aren’t any unexpected issues.

The post ESP32-C6-Bug WiFi 6, Bluetooth LE, and 802.15.4 board takes a PoE Ethernet shield (Crowdfunding) appeared first on CNX Software - Embedded Systems News.

The ASUS Tinker board 3 was first unveiled in April 2023 before being renamed as Tinker Board 3N later that year, and the three variants of the Rockchip RK3568 single board computer (SBC) are now available.

The standard configuration is the Tinker Board 3N in the commercial temperature range, while the Tinker Board 3N Plus has the same features, except it can operate in the industrial temperature range (-40°C to 85°C). The Tinker Board 3N Lite is a cost-down version in the same form factor, but with a single gigabit Ethernet port without PoE support, no M.2 B-key socket for an NVMe SSD or 4G/5G cellular connectivity, no 16MB SPI flash, fewer serial interfaces, and no CAN Bus.

You’ll find a comparison of the specifications for the three variants in the table below.

Note the prices above are from Amazon with a 10% discount when applicable.

ASUS provides support for Debian 11 and Android 12 operating systems for the Tinker Board 3N with support for the ASUS IoT Cloud Console and firmware over-the-air (FOTA) updates support. Documentation such as the user manual, drawing & schematics, and Qualified Vendors List (QVL) – i.e. devices tested to be compatible – can be found on the Tinker board website. Some instructions to build the Linux image can also be found on GitHub.

All three boards can be purchased on Amazon starting at $169.99 for the Lite variant with 4GB RAM and 64GB eMMC flash, and $179.10 and $251.10 for the Standard and Plus models respectively after ticking for a 10% discount coupon. ASUS has an extensive distribution network, so it may pay to shop around, and for instance, Rutronik24’s pricing may be more attractive to European customers, although it’s still fairly pricey for a Rockchip RK3568 board considering you can have similarly featured alternatives such as Radxa Rock 3B or Hardkernel ODROID-M1 at significantly lower price.

Via Liliputing

The post Rockchip RK3568-powered ASUS Tinker Board 3N is now available in three variants appeared first on CNX Software - Embedded Systems News.

Limitbit Doly is a cute little autonomous robot with two continuous tracks, two small arms controlled by servos, two round color displays acting as the eyes, and various sensors, all controlled by a Raspberry Pi CM4 system-on-module.

The robot can be used for STEM (Science, Technology, Engineering, and Mathematics) education or as a developer platform. AI workloads can also run on the Raspberry Pi CM4 module taking sensors, camera, and microphone inputs, with the robot interacting with the user through the built-in stereo speaker and two eyes. In practice, that means Doly supports features such as face recognition and smart audio with the robot capable of recognizing its owner and responding to voice commands.

Doly specifications:

• System-on-Module – Raspberry Pi CM4 Lite model CM4101000 (1GB RAM, Wireless) by default, but also supports other CM4/CM Lite modules with wireless
• Storage – MicroSD card slot
• Display – 2x high-resolution color displays (the two eyes)
• Camera – 8 MP wide-angle camera
• Audio – 2x microphones, 2W stereo speakers
• USB – 1x USB Type-A port
• Motor control
  • 2x servo motors for arms
  • 2x metal gear motors for the continuous tracks
  • 2x hi-res encoders
•  Sensors
  • 2x touch sensors (so you can pat the robot like you would with a cat or dog)
  • 2x ToF sensors for ranging
  • 6-axis IMU sensor
  • 4x IR edge Sensors
• Expansion for add-ons such as an additional robot arm
  • 12x header with 6x GPIO pins, power signals
  • 6-pin header with 2x servo outputs
  • 1x Qwiic I2C connector
  • 1x UART connector
• Misc – 2x RGB LEDs
• Power Supply
  • Input – 5V DC & 3.3V DC
  • Battery – 2,600 mAh battery good for about two hours
• Dimensions – 110 x 114 x 68 mm

Doly is said to be “fully open-hardware, open-design and 3D-printable”, but the company is being honest as they also mention that “while our hardware is open for community use, it’s important to note that we don’t completely align with the Open Source Hardware Association (OSHW) philosophy.”, which I understand as “some bits and pieces will be closed-source”.

No actual software resources have been shared right now because it’s a crowdfunding campaign, but Limitbit does provide some information about the software for Doly

• The CM4 runs a Linux-based OS, unsurprisingly…
• It can be programmed with Blockly visual programming IDE, as well as C, C++, and Python.
• The Python-based Doly SDK will provide API to retrieve sensor data (camera, ToF, edge sensor…) and some AI data (face recognition, emotion,..), and to control the motors, servos, and displays (e.g. Eye customize controls).
• While the Raspberry Pi robot can be interacted with in autonomous mode, the company will also provide the Dolby mobile app for Android and iOS, as well as a Windows program to control the robot remotely as an FPV (first-person view) vehicle.

You’ll get an idea of what the Doly robot is capable of in the video embedded below.

Limitbit has recently launched the Doly robot on Kickstarter and easily surpassed its ~$6,000 funding target with over $130,000 raised so far thanks to pledges from over 400 backers. Rewards start at $269 for the “3D Maker Edition” with all the electronics minus the Raspberry Pi CM4 module and no plastic parts that can be 3D printed by the user.  I’d expect most people to go with the $299 reward for a fully assembled Doly robot including a Raspberry Pi CM4 Lite module with 1GB RAM and WiFi. Prices do not include shipping which will be calculated once Limitbit is ready to ship rewards to backers in August 2024 if everything goes according to plans…

Thanks to TLS for the tip.

The post Doly – A cute little autonomous AI-powered robot based on Raspberry Pi CM4 module (Crowdfunding) appeared first on CNX Software - Embedded Systems News.

We previously had a look at the hardware of the GEEKOM A7 with an unboxing and a teardown of the powerful AMD Ryzen 9 7940HS mini PC with 32GB DDR5, a 2TB NVMe SSD, four 4K-capable video outputs, and high-speed interfaces such as USB4 and 2.5GbE, as well as WiFi 6E and Bluetooth 5.3 wireless connectivity.

We’ve now had time to test it with Windows 11 Pro in detail, so in the second part of the GEEKOM A7 review, we’ll report our experience with the mini PC including a software overview, features testing, various benchmarks, networking and storage performance testing, fan noise, power consumption, and more.

Software overview and features testing

The System->About window confirms the GEEKOM A7 mini PC is powered by an AMD Ryzen 9 7840HS processor with Radeon 780M graphics with 32GB RAM and runs Windows 11 Pro 23H2 build 22631.2861. That also means we only had to apply a few minor updates to get a fully updated system.

HWiNFO64 provides additional details about the AMD Ryzen 9 7940HS 8-core/16-thread Zen4 processor, the motherboard, and the integrated AMD Radeon 780M GPU.

A few more details about the “Phoenix” GPU can also be found in GPU-Z.

The PL1 and PL2 power limits are set to 45W (PBP) and 60W (MTP) respectively with the processor having a configurable 35W to 54W TDP. So GEEKOM was not as conservative with power settings as in their other recent mini PCs.

HWiNFO64 reports two Crucial 16GB DDR5 SO-DIMM memory sticks based on Micron chips clocked at 2800 MHz (DDR5-5600) for a total capacity of 32 GB.

Windows Task Manager confirms that with 32GB (31.3GB) RAM at 5,600 MHz via two SODIMM modules.

Let’s now check the Network adapters in Device Manager to find more information about 2.5GbE, WiFi 6E, and Bluetooth

HWiNFO64 shows 2.5GbE networking is implemented through an RTL8125 2.5GbE controller.

In the teardown the WiFi 6E module was “Azurewave AW-XB591NF”, but at the time we could not find public information about the chipset used in that module. It turns out it’s the MediaTek MT7922 with a maximum link speed of 2402 Mbps. Both WiFi and Bluetooth happen to be supported in Linux too, so that’s good news, although that’s still something we will be testing in Ubuntu 22.04.

We can go back to the Device Manager to check the Bluetooth version in the Advanced tab of the MediaTek Bluetooth Adapter Properties.

LMP 12.xxx firmware version looks up to Bluetooth 5.3 as advertised, and I successfully tested it with a Bluetooth audio headset. However, trying to transfer files from an Android 14 smartphone to the mini PC failed. Pairing worked, but the phone and PC would not stay connected and all my attempts to transfer files failed. It’s not the first time it happens with this phone though, so that may not be an issue specific to the mini PC.

GEEKOM will properly mark all USB ports with speed and feature markings on its mini PCs. But that does not mean we can’t check those, and we’ll do that with ORICO M234C3-U4 M.2 NVMe SSD enclosure along with HWiNFO64 to check the USB version and speed, and CrystalDiskMark to confirm the file transfer speed.

When connected to the 40 Gbps USB4/Thunderbolt port, the drive shows as an NVMe 1.3 SSD connected over a PCIe x4 8 GT/s interface using Phison Electronics PS5013 controller.

2196 MB/s is around the maximum read speed of the Apacer AS2280P4 NVMe SSD and confirms the USB4 port delivers well over 10 Gbps.

Results for the USB ports on GEEKOM A7’s front panel (left to right) in Windows 11:

• USB-A #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 897.84 MB/s
• USB-A #2 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 895.53MB/s

Same tests for the rear panel (left to right):

• USB-C #1 – Thunderbolt/NVMe 8GT/s – 2196.30 MB/s
• USB-A #1 (Top) – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 965.54 MB/s
• USB-A #2 (Bottom) – “USB 3.0 (connected to a USB 2.0 port) – USB 2.0 High-Speed (480 Mbps) – 43.34 MB/s
• USB-C #2 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 965.35 MB/s

All USB ports are performing as advertised, but the front USB 3.2 ports are somewhat slower likely because they are behind a Genesys Logic USB 3.2 hub chip.

The GEEKOM A7 supports up to four 8K/4K displays. While I don’t have any 4K or higher displays for testing right now, I still tested a quad display setup through the two HDMI ports and two USB-C ports using various displays and adapters namely a TCL Full HD TV, CrowView laptop monitor (USB-C), “RPI-All-in-One display” using a Beelink Expand M USB-C dock with an HDMI port, and a VGA monitor connected through an HDMI to VGA adapter.

Everything worked fine.

GEEKOM A7 benchmarks on Windows 11

At this point, I would usually set the system to “Best Performance” or “High Performance”, but the only power plan available is “balanced”. So I went with that. It should be possible to enable more power plans with tweaks in the BIOS and/or Windows, but in our experience, the results are not that different…

The first benchmark I ran to evaluate the performance of the GEEKOM A7 with Windows 11 was PCMark 10.

The mini PC scored 7,516 points. That’s the highest score I got in the reviews I’ve personally done for CNX software (and better than 92% of results in PCMark 10), although the Chatreey AM08 Pro got 7,561 points with the same Ryzen 9 7940HS processor. You’ll find the full results on the 3DMark website.

We carried on with 3DMark’s Fire Strike 3D graphics benchmark where the GEEKOM Mini PC achieved 7,895 points. That’s the best score we’ve gotten so far, and still much higher than the 6,603 points reported in Vladislav Losev’s review of the Chatreey AM08 Pro.

The GEEKOM A7 mini PC achieved 8,058.2 points in PassMark’s Performance Test 11.0, which again is the best score we’ve ever seen for a mini PC. The 3D Graphics mark is not too bad with a 49th percentile ranking. For reference, the Intel Core i9-13900H powered GEEKOM Mini IT13 only got 5580 points in the same benchmarks.

PassMark shows the GEEKOM A7 is in the top 99th percentile for storage, and CrystalDiskMark confirms the excellent performance of the 2TB NVMe SSD with 4906.30 MB/s and 4710.80 MB/s sequential read and write speeds respectively. But it’s only the second best in our tests, as the Khadas Mind Premium’s SSD (WD PC SN740) is still ahead with 5.2 GB/s and 4.9 GB/s read/write speeds, and better random I/Os as well.

Cinbench R23 was used to test both single-core and multi-core performance and the GEEKOM A7 mini PC scored 15,231 points for the multi-core benchmark and 1,831 points for the single-core one with an 8.4x MP ratio which is pretty good compared to some other mini PCs I’ve recently reviewed.

I started testing the GPU with Unigine Heaven Benchmark 4.0 with the AMD Ryzen 9 7940HS mini PC achieving 80.7 fps on average and a score of 2,033 points at 1920×1080 resolution.

Next up was YouTube video playback at 4K and 8K resolution in the latest version of Firefox.

4Kp30 worked flawlessly with a smooth video and no frame dropped.

8K 30 fps video streaming was equally good with only 8 frames dropped at the very beginning.

No problem switching to 4K 30 fps either with no frames dropped, but 8K 60 fps was another story with the video being rather choppy, even those the stats for nerds overlay only reported 80 frames dropped out of 3614.  That’s not what my eyes were telling me. I used a USB-C display and switched to an HDMI TV just in case, but the result was similar.

So I switched to Chrome web browser, and the video was smoother but not perfect, and this time the “stats for nerds” overlay did show 275 frames dropped out of 5413 (or about 5% of frames).

I would have expected perfect video playback in both Firefox and Chrome in such a high-end mini PC up to 8K 60 fps, but that’s not the case for unclear reasons. I also took the opportunity to test audio through HDMI and the 3.5mm audio jack and both worked as expected.

GEEKOM A7 benchmarks comparison against other mini PCs running Windows 11 Pro

Let’s compare some of Windows 11 benchmark results for the GEEKOM A7 (AMD Ryzen 9 7940HS) mini PC against other high-0end mini PCs including the Chatreey AM087 Pro based on the same processor, the GEEKOM Mini IT13 (13th gen Core i9-13900H Raptor Lake), the Khadas Mind Premium (13th Gen Core i7-1360P Raptor Lake), and the GEEKOM AS 6 (AMD Ryzen 9 6900HX) in similar environmental conditions (28-30°C room temperature). But first a quick summary of the main features of the five mini PCs.

* The Chatreey AM08 Pro mini PC shipped with a 512GB (PCIe Gen 3) SSD, but was replaced by a 1TB Samsung 990 Pro NVMe (PCIe Gen4 x4) SSD for review.

And now the benchmark results.

The GEEKOM A7 has similar results compared to the Chatreey AM08 Pro, so its small size does not seem to affect its performance. It bears repeating that you should ignore the high “Disk Mark” score for the AM08 Pro because Vladislav replaced the included SSD with his own, much faster Samsung SSD. What’s surprising is the significantly higher score in 3DMark for the GEEKOM A7 (also confirmed with Unigine Heaven Benchmark 4.0), so it might be that AMD released new drivers to extract more performance and the 5600MHz (vs 4800 MHz) DDR5 memory may have helped too… The 3D graphics and multi-core performance of the AMD Ryzen 9 7940HS CPU are unmatched, but the Intel Core i9-13900H in the Mini IT13 still delivers a higher single-core performance which explains why it still tops some of the benchmarks.

Networking (WiFi 6 and 2.5GbE) benchmarks

Let’s now test the performance of the 2.5GbE interface (192.168.31.15) in the GEEKOM A7 using iperf3 and UP Xtreme i11 mini PC on the other side of the connection.

• Upload

C:\Users\jaufr\Downloads\iperf-3.1.3-win64\iperf-3.1.3-win64>iperf3 -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  4] local 192.168.31.128 port 54042 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec  2.75 GBytes  2.36 Gbits/sec
[  4]  10.00-20.00  sec  2.75 GBytes  2.37 Gbits/sec
[  4]  20.00-30.00  sec  2.75 GBytes  2.36 Gbits/sec
[  4]  30.00-40.00  sec  2.75 GBytes  2.37 Gbits/sec
[  4]  40.00-50.00  sec  2.75 GBytes  2.36 Gbits/sec
[  4]  50.00-60.00  sec  2.75 GBytes  2.36 Gbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-60.00  sec  16.5 GBytes  2.36 Gbits/sec                  sender
[  4]   0.00-60.00  sec  16.5 GBytes  2.36 Gbits/sec                  receiver

iperf Done.

• Download

C:\Users\jaufr\Downloads\iperf-3.1.3-win64\iperf-3.1.3-win64>iperf3 -t 60 -c 192.168.31.12 -i 10  -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  4] local 192.168.31.128 port 54077 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec  2.76 GBytes  2.37 Gbits/sec
[  4]  10.00-20.00  sec  2.76 GBytes  2.37 Gbits/sec
[  4]  20.00-30.00  sec  2.76 GBytes  2.37 Gbits/sec
[  4]  30.00-40.00  sec  2.76 GBytes  2.37 Gbits/sec
[  4]  40.00-50.00  sec  2.76 GBytes  2.37 Gbits/sec
[  4]  50.00-60.00  sec  2.76 GBytes  2.37 Gbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-60.00  sec  16.6 GBytes  2.37 Gbits/sec    0             sender
[  4]   0.00-60.00  sec  16.6 GBytes  2.37 Gbits/sec                  receiver

iperf Done

All good. Time for WiFi 6 testing through Xiaomi Mi AX6000 router.

• Upload

C:\Users\jaufr\Downloads\iperf-3.1.3-win64\iperf-3.1.3-win64>iperf3 -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  4] local 192.168.31.9 port 53999 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec   439 MBytes   368 Mbits/sec
[  4]  10.00-20.00  sec   433 MBytes   363 Mbits/sec
[  4]  20.00-30.00  sec   440 MBytes   369 Mbits/sec
[  4]  30.00-40.01  sec   444 MBytes   372 Mbits/sec
[  4]  40.01-50.00  sec   442 MBytes   371 Mbits/sec
[  4]  50.00-60.00  sec   442 MBytes   371 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-60.00  sec  2.58 GBytes   369 Mbits/sec                  sender
[  4]   0.00-60.00  sec  2.58 GBytes   369 Mbits/sec                  receiver

iperf Done.

• Download

C:\Users\jaufr\Downloads\iperf-3.1.3-win64\iperf-3.1.3-win64>iperf3 -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  4] local 192.168.31.9 port 54011 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec   745 MBytes   625 Mbits/sec
[  4]  10.00-20.00  sec   745 MBytes   625 Mbits/sec
[  4]  20.00-30.00  sec   755 MBytes   634 Mbits/sec
[  4]  30.00-40.00  sec   765 MBytes   641 Mbits/sec
[  4]  40.00-50.00  sec   762 MBytes   639 Mbits/sec
[  4]  50.00-60.00  sec   768 MBytes   644 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-60.00  sec  4.44 GBytes   635 Mbits/sec   12             sender
[  4]   0.00-60.00  sec  4.43 GBytes   635 Mbits/sec                  receiver

iperf Done.

The 635 Mbps download speed is similar to what I got with Khadas Mind Premium, but the 369 Mbps upload speed is quite slower (712 Mbps in the Mind Premium).  Nevertheless, it’s still pretty good.

Thermal performance

The GEEKOM A7 is arguably the fastest mini PC we’ve reviewed so far, and while our benchmarks did not indicate any CPU throttling for the AMD Ryzen 9 7940HS processor, we tested the thermal performance to check CPU Throttling in Windows 11 with HWiNFO64 and various utilities starting with 3D Mark Fire Strike benchmarks.

The CPU temperature topped around 95°C, but no Throttling was detected by HWiNFO64.

We then rebooted the mini PC and tried again with AIDA64’s stability test.

Again, the CPU temperature topped at 95.4°C, but no thermal throttling was reported. The CPU frequency stabilized at around 4,000 MHz for all eight cores during the stress test.

Fan noise

The fan is running all the time, but it’s barely audible during idle or light tasks. It does get noisier during heavier loads although not annoyingly so, but obviously, this depends on the sensibility of each user.  I used a sound level meter 5 cm from the top of the device to measure the fan noise:

• Idle – 41.8 to 45 dBa
• AIDA64 stability test – 48.1 – 48.6 dBa

The room background noise is 38 to 39 dBa.

GEEKOM A7 power consumption

Here are the power consumption numbers of the GEEKOM A7 mini PC running Windows 11 Pro using a wall power meter:

• Power Off – 1.4 to 1.5 Watts
• Idle – 4.3 – 5.4 Watts
• Video playback
  • YouTube 8K 60fps in Chrome – 24.2 – 30.9 Watts
  • YouTube 8K 60fps in Firefox – 29.2 – 35.7 Watts
• CPU stress tests
  • Cinebench R23 Multi-core
    • First few seconds – 79 to 81 Watts
    • Long run – 57 – 60.5
  • AIDA64 stability test
    • First few seconds – 83 to 87 Watts
    • Long run – 59.3 – 67 Watts

The mini PC was connected to WiFi 6, two RF dongles were plugged into USB ports for a wireless keyboard and mouse, and the CrowView display was connected to the mini PC through an HDMI cable and a separate USB-C power adapter.

Conclusion

The GEEKOM A7 is the fastest mini PC we’ve reviewed so far, and the AMD Ryzen 9 7940HS 8-core/16-thread processor especially shines when it comes to multi-core and 3D graphics performance which are respectively 14% (Cinebench R23 multi-core at 45W PL1) and 46% faster (3DMark Fire Strike) than the GEEKOM Mini IT13 based on an Intel Core i9-13900H 14-core/20-core processor. That means you should be able to run pretty much anything that you’d normally do on a larger PC, even play some AAA games, although the iGPU won’t quite match the performance of discrete graphics cards from NVIDIA or AMD.

Most features I tested worked fine including driving four independent displays through HDMI and USB ports, Thunderbolt support, and audio output from HDMI and the headphone jack. The performance of the NVMe SSD is very good, the 2.5GbE port and WiFi 6 delivered throughput that matched our expectations, and CPU throttling did not happen even under heavy loads. The mini PC’s fan is fairly quiet most of the time, and not too annoying when more demanding tasks are running. The only real issue I had was with 8K 60 fps video playback that was not perfectly smooth in Chrome (better) and Firefox (worse), but 4K 60 fps and 8K 30fps were fine.

I’d like to thank GEEKOM for sending the A7 mini PC for review. The model reviewed here with 32GB DDR5 and 2TB SSD can be purchased on Amazon for $829 with the coupon code CNXSW3A7 as well as on the GEEKOM store. with the discount code cnxsoftwarea7, which also works on the GEEKOM UK store.

The post GEEKOM A7 review – Part 2: Windows 11 Pro tested in a compact AMD Ryzen 9 7940HS mini PC appeared first on CNX Software - Embedded Systems News.

Waveshare has recently launched two new ESP32-S3 4G dev boards – the ESP32-S3-SIM7670G-4G and the ESP32-S3-A7670E-4G. These boards support 4G LTE Cat-1, Wi-Fi, Bluetooth, and GNSS, and come with an OV2640 camera, and a battery holder for a 18650 battery. The main difference between the two is that the A7670E module also supports 2G GSM/GPRS/EDGE at 900/1800MHz while the SIM7670G module does not.

The board has two rows of I/Os including GPIO, I2C, SPI, ADC, and USB 2.0. It also has a USB-C port for power and programming, a slot for a MicroSD card, and an option to connect an external speaker. There’s a USB switching IC and DIP switch for easily connecting the module to a PC for internet or debugging.

Waveshare ESP32-S3 4G Dev Boards Specification:

• Communication module
  • A7670E 4G – A7670E Cat-1 4G module supporting 4G Cat-1 + 2G networking, GNSS positioning, telephone calls, and SMS.
  • SIM7670G 4G – SIM7670G Cat-1 4G module supporting LTE Cat-1 networking, GNSS positioning (GPS, BeiDou, GLONASS, Galileo).
• Global coverage and band support
  • A7670E 4G – Focuses on LTE-FDD bands more suitable for Europe, Southeast Asia, West Asia, Africa, China, and South Korea with additional GSM/GPRS/EDGE support.
  • SIM7670G 4G – Offers global coverage with a broader range of LTE-FDD and LTE-TDD bands, making it applicable worldwide, but lacks GSM/GPRS/EDGE support.
• Wireless MCU – ESP32-S3 Xtensa 32-bit LX7 dual-core, up to 240MHz frequency, 512KB SRAM, 384KB ROM, 2.4 GHz WiFi 4 and Bluetooth 5.0
• Memory and storage
  • 2MB PSRAM
  • 16MB Flash storage
• Peripheral Interfaces
  • OV2640 camera
  • MicroSD card slot
  • USB Type-C port used to AT commands, GNSS positioning, firmware upgrading, program burning, etc…
  • 38-pin header
• Power Management
  • Lithium battery support
  • Solar charging (5V to 18V) via CN3791
  • USB charging through ETA6098
  • Battery voltage measurement using MAX17048G
• Security features
  • Hardware encryption
  • Random number generator
  • HMAC
  • Digital Signature
• Additional features
  • RGB LED
  • 18650 battery holder
  • GNSS high-precision ceramic antenna and
  • Antenna connector – IPEX 1
• Board Dimensions – 110 x 30.44 mm

The onboard OV2640 camera captures video at 1600×1200 resolution and 15 frames per second, offering decent quality. Additionally, the calling module features a microphone and a speaker to enable calling functionality.

The boards can be equipped with an 18650 lithium battery, which can be recharged using solar power. To safeguard the battery, the board incorporates an ETA6098 IC for battery charging, a CN3791 IC for solar charging, and a MAX17048G IC for measuring battery voltage.

Previously we have covered several other ESP32-based 4G modules like LILYGO T-SIMCAM ESP32-S3 and  LILYGO T-SIM7080G-S3 feel free to check those out if you are interested in the topic.

Waveshare offers code samples and instructions using VSCode (ESP-IDF) and the Arduino IDE, and details about the AT commands for the A7670E and SIM7670G modules on their respective wiki page, where you’ll also find more technical information, documents, and drivers.

You can purchase the Waveshare’s ESP32-S3 Boards from both Amazon and AliExpress. The ESP32-S3 SIM7670G 4G is available on Amazon for $56.29 and on AliExpress for $56.48. Meanwhile, the ESP32-S3 A7670E 4G is priced at $49.29 on Amazon and $34.00 on AliExpress. Both boards can also be found on the Waveshare store.

The post Waveshare ESP32-S3 4G dev boards feature LTE Cat-1, 2G, and camera support appeared first on CNX Software - Embedded Systems News.

Bojan Jurca’s “Esp32_oscilloscope” is an open-source Arduino sketch that can transform an ESP32 board into a web-based oscilloscope that works over WiFi.

We had also written about the Scoppy project to turn the Raspberry Pi Pico W into a 2-channel oscilloscope, but there’s no reason the more powerful ESP32-series microcontroller could not be used for the same purpose, and Bojan’s Esp32_oscilloscope project does just that and works with ESP32, ESP32-S2, ESP32-S3 and ESP32-C3 boards using the I2S interface for fast data sampling.

The project was initially designed to demonstrate the multitasking abilities of the ESP32 microcontroller with Arduino, but this evolved into an ESP32 oscilloscope firmware. It works both with output/PWM and input signals, digital (0 or 1) and analog (0 to 4095) signals, and the web interface shows up to 736 samples per screen although the sampling rate may not be completely constant all the time.

To install it on your board, you’ll need to load the project in the Arduino IDE and set your router credentials and the hostname:

// if these #definitions are missing STAtion will not be set up
    #define DEFAULT_STA_SSID                          "YOUR_STA_SSID"       // <- replace with your information
    #define DEFAULT_STA_PASSWORD                      "YOUR_STA_PASSWORD"   // <- replace with your information

    // define the name Esp32 will use as its host name 
    #define HOSTNAME                                   "MyEsp32Oscilloscope"  // <- replace with your information,  max 32 bytes

The program relies on the FAT file system, so you’ll need to select one of the FATFS partition schemas in Tools->Partition scheme-> ... unless the ESP32 board does not have external flash in which case you’d need to comment out the

#define FILE_SYSTEM FILE_SYSTEM_FAT

…in order for the firmware to use progmem to store the oscilloscope.html file. Now you can build the program and flash it to your board. You’ll also need to upload a few files to the /var/www/html directory (android-192-osc.png, apple-180-osc.png, oscilloscope.html) over FTP, and you should be able to access the oscilloscope with “http://your-esp32-ip/oscilloscope.html” or “http://esp32-hostname/oscilloscope.html” URL from your web browser.

If you are short on time and/or don’t happen to have an ESP32 board on hand, Bojan is also running a demo at http://jurca.dyn.ts.si/oscilloscope.html as shown in the screenshot above. The full details for the Esp32_Oscilloscope project can be found on GitHub. It’s not the first time we cover a project that shows the status of ESP32 GPIO pins in a web browser, as GPIOViewer does just that but serves a different use case.

Via gojimmypi on Twitter/X

The post Esp32_oscilloscope Arduino firmware turns your ESP32 board into a web-based oscilloscope appeared first on CNX Software - Embedded Systems News.

The SoC Discovery Kit is the latest addition to Microchip’s list of development kits for the PolarFire series. The series is the first SoC FPGA family powered by a deterministic, coherent RISC-V CPU cluster. They provide low power consumption, thermal efficiency, and defense-grade security for smart, networked systems. They also support a deterministic L2 memory system for Linux and real-time applications.

Microchip launched the Icicle Kit for the PolarFire SoC in 2020 and it was followed by the Video and Imaging Kit which was intended for mid-bandwidth imaging and video applications. Now, Microchip has announced the Discovery Kit which is billed as a low-cost alternative to the Icicle. The Discovery Kit retains the full range of features needed for testing concepts quickly, developing firmware applications, and programming/debugging user code.

According to Microchip, the kit will bring “a low-cost RISC-V and FPGA development for learning and rapid innovation” to new and experienced engineers as well as university students. Other recent products like the BeagleV-Fire single-board computer have implemented the PolarFire SoC FPGA.

Polarfire SoC Discovery Kit specifications:

• SoC FPGA – PolarFire SoC MPFS095T-1FCSG325E penta–core RISC-V CPU subsystem (1x RV64IMAC @ 625MHz, 4x RV64GC @ 625MHz) with
  • 95K LE non-volatile fabric
  • 292 18 × 18 math blocks
  • Secure boot
  • 4x 12.7 Gbps SERDES
• System Memory – 1GB LPDDR4 x16
• Storage – 1x microSD card slot
• Video Output – MIPI video interface
• Connectivity – 1x Gigabit Ethernet
• Expansion Ports
  • 40-pin Raspberry Pi compatible header with GPIO, I2C, SPI, UART
  • mikroBUS socket
  • 3x UART via USB Type-C port
• Debugging
  • 2 x push buttons
  • 8x debug LEDs
  • USB-C (UART)
  • 8x dip switches
• Misc – 3x power LEDs, 2x user buttons, a jumper for selecting a power supply, 7-segment display connector
• Power Supply – 5V @ 3A via USB Type-C or external power supply
• Dimensions – 4.1”  x 3.3”

The Polarfire Discovery Kit is smaller than the Icicle Kit and excludes features such as a PCIe slot, micro-USB ports, onboard storage, wireless connectivity, a power barrel jack, and the I2C power sensor. Also, it has less RAM, only one Ethernet port, and uses a single USB Type-C port for power and debugging. It retains the 40-pin Raspberry Pi-compatible header and the mikroBus socket for connecting Click Boards.

The board has an embedded FlashPro5 (FP5) programmer for programming and debugging the FPGA, and for developing firmware applications. The board is supported by Microchip’s Libero SoC software and buyers will be provided with a free Libero Silver license. You can find further information, schematics for the kit, a user guide, and other accompanying documentation on the product page.

The PolarFire SoC Discovery Kit is priced at $132 for a single board. However, a discount is available via Microchip’s Academic Program and members only have to pay $99. Production kits are expected to ship from April 2024. The package will contain the Discovery Kit board, a quickstart card, and a USB 2.0 Type C-to-C cable.

The post Microchip announces the PolarFire SoC Discovery Kit, a low-cost devkit for Linux and real-time applications appeared first on CNX Software - Embedded Systems News.

The SparkFun MicroMod Single Pair Ethernet Kit utilizes the ADIN1110 transceiver for 10BASE-T1L communication, enabling long-distance connectivity over a single twisted pair cable. Things get easier as SparkFun has designed the module to be compatible with the MicroMod ecosystem, which allows for testing connections between industrial devices over long distances, up to 1,700 meters.

The standout feature of this board is its ADIN1110 transceiver, which supports 10BASE-T1L Ethernet. This technology uses a single twisted pair for lightweight, long-distance data transmission, achieving speeds up to 10Mbps by the 802.3cg IEEE standard, all without delivering power over the cable.

SparkFun MicroMod Single Pair Ethernet Kit Specification:

• 2x SparkFun MicroMod Single Pair Ethernet Function boards
  • 10BASE-T1L IEEE Standard 802.3cg-2019 Compliant Transceiver
    • Single-pair Ethernet transmission at speeds up to 10Mbps
    • 1km transmission distance (1.7km max cable reach)
  • Supply Voltage of 1.8V or 3.3V
  • 2.4V transmission amplitude
  • Integrated MAC connects via SPI – Supports 16 MAC addresses
  • Supports both Generic and OPEN Alliance SPI protocols
• 2x SparkFun MicroMod Main boards
  • Power Supply Options – 5V via USB-C or 3.7V~4.2V via LiPo Battery.
  • Protection – Built-in resettable PTC fuse (5V/2A).
  • Voltage Regulators – AP7361C (3.3V/1A) and AP7347DQ (3.3V/500mA for Qwiic devices).
  • Charging Circuit – Integrated MCP73831 for single-cell LiPo, 500mA charging.
  • Connectivity
    • 1x USB-C
    • 1x 2-Pin JST for LiPo Battery
    • 2x MicroMod (Processor and Function boards)
    • 2x Qwiic I2C
    • 1x MicroSD card slot
    • 1x SWD 2×5 header
    • Built-in MUX for UART1
  • User Interface
    • Buttons: 1x Reset, 1x Boot
    • LEDs: VIN, 3.3V, Qwiic 3.3V, CHG
  • Dimensions – 3.40″ x 2.90″.

The SparkFun MicroMod Single Pair Ethernet Kit is open-sourced so Sparkfun provides all the necessary hardware-related documentation on their GitHub Hardware Repo.

The ADIN1110 transceiver used in SparkFun’s Kit isn’t new; we discussed it back in 2021. Meanwhile, Microchip released their new LAN8650 and LAN8651 10BASE-T1S MAC-PHY Ethernet controllers in 2023. But with the release of this new SparkFun Kit, it’s good to see that now it can be bought easily in a shop.

The board is all well and good but there are some limitations, as the kit requires separate power for each MicroMod Main Board as it supports data-only transmission, and uses a T1 Industrial AH IP20 Jack for the network cable, which SparkFun notes is outdated for 10BASE-T1L Ethernet.

The MicroMod Main Board has an M.2 connector in which the Single Pair Ethernet boards plugin. It features a USB-C for power and programming, a jumper to isolate the USB-C’s shield, reset and boot buttons, 2x 5-pin SWD pin breakouts, a 2A resettable fuse, and dual voltage regulators (3.3V/1A for general use and 3.3V/500mA for Qwiic devices). Additionally, it includes PTH jumpers for current measurement, a 2-pin JST for LiPo batteries with a charging IC, four status LEDs, a microSD card socket for data logging, and two Qwiic connectors for I2C expansion. more information about this board is available on MicroMod Main Board products page.

The kit includes two SparkFun MicroMod Single Pair Ethernet Function boards with ADIN1110, two single MicroMod Main boards, and one 0.5m shielded single-pair Ethernet cable. Available on the SparkFun store, the kit is priced at $89.95, with additional discounts for bulk purchases. More information about the product can also be found in the SparkFun Getting Started Guide.

The post SparkFun launches MicroMod Single Pair Ethernet kit for easier 10BASE-T1L experimentation appeared first on CNX Software - Embedded Systems News.

We’ve already listed the specs and done an unboxing and teardown for the Beelink SEi12 i7-12650H Mini PC powered by a 10-core Intel Core i7-12650H processor with 32GB DDR4 and a 500GB M.2 SSD preloaded with Windows 11 Pro.

In the second part of the review, we will test the Beelink SEi12 i7-12650H with the Windows 11 Pro operating system in detail with a software overview and feature testing, benchmarks, networking and storage testing, thermal efficiency, fan noise, and power consumption. Since we also reviewed the GEEKOM Mini IT12 with the same processor last month, we’ll try to compare both in this review and list the pros and cons for each system.

Software overview and features testing

The System->About menu confirms that we have a “SEi” Mini PC powered by a 12th Gen 1.5 GHz (base frequency) Intel Core i7-12650H processor and  32GB of RAM, running Windows 11 Pro operating system version 23H2 after we went through all the updates needed from the 21H2 version the system shipped with.

The HWiNFO64 program provides more details about the Intel Core i7-12650H 10-core (6E+4P) processor with 16 Threads (12P+4E), the AZW SEi motherboard, and the Intel UHD graphics found in the SoC.

GPU-Z program offers additional details about the 64EU Intel UHD Graphics found in the Intel Core i7-12650H SoC.

The PL1 and PL2 power limits are set to 35W and 55W while the Intel Core i7-12650H processor has a TDP of 45W, so Beelink looks to have opted for conservative settings here.

HWiNFO64 reports two 16 GB 1600 MHz Crucial DDR4-3200MHz SO-DIMMs.

Windows Task Manager confirms this by showing 32GB of RAM clocked at 3,200 MHz with two SODIMM memory sticks.

We can go to the Device Manager’s Network adapters section to check gigabit Ethernet, WiFi, and Bluetooth 5.2 support.

The Beelink SEi12 i7-12650H mini PC features a gigabit Ethernet port through a RealTek Semiconductor RTL8168/8111 PCIe Gigabit Ethernet controller.

WiFi 6 is implemented through an Intel AX200 module with a maximum link speed of 1729 Mbps.

We now need to go back to the Device Manager to double-check the Bluetooth version.

“LMP11” firmware looks up to Bluetooth 5.2 and we have successfully tested it with a Bluetooth headset and transferred some files from an Android smartphone headset.

We will now test the USB ports’ speed using HWiNFO64 and CrystalDiskMark programs and an ORICO M234C3-U4 M.2 NVMe SSD enclosure, except for the USB 2.0 port where another USB expansion drive will be used.

The results of all five USB ports are summarized as follows (from left to right):

• Front panel
  • USB-A #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 1,049 MB/s read speed
  • USB-A #2 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 1,047 MB/s read speed
  • USB-C #1 – USB 3.2 – USB 3.1 SuperSpeedPlus (10 Gbps) – 1,047 MB/s read speed
• Rear panel
  • USB-A #1 (top) – USB 2.0 – USB 2.0 Hight-Speed (480 Mbps) – 43.81 MB/s read speed
  • USB-A #2 (bottom) – USB 2.0 – USB 2.0 Hight-Speed (480 Mbps) – 43.72 MB/s read speed

We also installed a 2.5-inch SATA SSD drive…

… and tested the performance with CrystalDiskMark.

A read speed of 221 MB/s and a write speed of 152 MB/s are expected for this drive.

The Beelink SEi12 i7-12650H Mini PC supports up to two independent displays via HDMI 2.0 and DisplayPort 1.4 ports. We don’t own a monitor with DisplayPort input,  so we used a DisplayPort to HDMI cable that we previously tested successfully with the GEEKOM Mini Air12‘s mini DP port using an additional mini DP to DP adapter.

We connected the HDMI port to a VGA monitor through an adapter and the DisplayPort connector to the HDMI input of the 10.1-inch “RPI All-in-One” display. HDMI output worked fine, but DisplayPort output would not work and the monitor shows the output from the internal display interface connected to an Arm SBC instead. So we switched to an HD television instead, but the result was the same with the TV showing “No Signal”.

That means we were unable to drive two displays with the Beelink SEi12 i7-12650H mini PC likely due to some incompatibilities between our DisplayPort to HDMI cable and the DisplayPort video output in the mini PC

Beelink SEi12 i7-12650H benchmarks in Windows 11

We set the system’s power mode to “Best performance” before running benchmarks on the mini PC.  Note that the ambient temperature was 28 to 30°C during testing, and your own results may end up being different.

We started testing the performance of the Beelink SEi12 i7-12650H mini PC with the PCMark 10 benchmark.

The mini PC achieved 5,360 points in PCMark. You’ll find the full results on the 3DMark website.

Next up was 3DMark Fire Strike where the SEi12 got 3,618 points.

When we first ran PassMark PerformanceTest 11.0, we noted the 3D Mark test had no score because “GPU Compute” failed to run but without any specific error message. You’ll also notice the overall score is crazy high at 9,222 points, while for reference, the more powerful GEEKOM Mini IT13  (Core i9-13900H) got only 5580.4 points, and the GEEKOM Mini IT12 with the same Core i7-12650H processor only got 3,521 points. So it’s clear that one of those “benchmark gone wrong” results…

When trying the GPU Compute benchmark again, we noted it would only run one test and exit without any error messages or any score. We spent some time checking opencl.dll was indeed installed, updated the drivers, and tried to find a solution online. But nothing seemed to work. We eventually noticed a new version of the PerformanceTest benchmark was available (build 1009), and after the update, the CPU Compute benchmark could complete normally. So we ran the full PassMark benchmark again and got a believable score. Just make sure you avoid the 1008 build if you encounter a similar error.

We tested the 500 GB NVMe SSD (PCIe x4 16.0 GT/s @ x4 16.0 GT/s) included with the mini PC with CrystalDiskMark, and the results are OK with a sequential read speed of 4,836 MB/s and a sequential write speed of 1,906 MB/s.

The Cinebench R23 benchmark was used to evaluate single-core and multi-core performance.

The Beelink SEi12 i7-12650H mini PC achieved 8,494 points in the multi-core benchmark and 1,646 points in the single-core test with an MP Ratio of 5.16x which is quite better than the 5,273 points (2.96x MP radio) for the GEEKOM Mini IT12 (PL1 set to 35W) indicating a better cooling performance.

Unigine Heaven Benchmark 4.0 was used to further test 3D graphics acceleration, and the Core i7-12650H mini PC managed to render the demo at 39.3 FPS on average with a 989 score at 1920×1080 resolution. That one is a bit lower than the results on the GEEKOM Mini IT12  (41.9 FPS).

We then played some YouTube videos at 4K and 8K resolutions in Google Chrome.

YouTube 4Kp30 played smoothly closed to 8 minutes with no frames dropped at all.

4Kp60 was equally good with only 10 frames dropped out of 29,882.

The mini PC still performed nicely at 8K 30 FPS with only one frame dropped while playing the video for over 8 minutes.

One final test at 8K 60 FPS was all good too with only 10 frames dropped out of 38,221 while playing the video for a little 10 minutes. Those results should not be surprising as the GEEKOM Mini IT12 has the same results, so the Core i7-12650H is well-supported in Google Chrome when playing YouTube videos.

Beelink SEi12 i7-12650H benchmarks comparison against other Intel and AMD mini PCs

To better understand the weaknesses and strengths of the Beelink SEi12 i7-12650H in Windows 11, we’ll compare the benchmark results against other mini PCs, namely GEEKOM IT12 (Intel Core i7-12650H), GEEKOM IT13 (Intel Core i9-13900H), GEEKOM AS 6 (AMD Ryzen 9 6900HX), and Khadas Mind Premium (Intel Core i7-1360P). All systems were tested at an ambient temperature of around 28-30°C.

But before looking at the benchmarks, let’s list the main features of the five systems under test.

Benchmark results.

The Beelink SEi12 i7-12650H and GEEKOM IT12 are comparable with some minor differences except for the Cinebench R23 multi-core benchmark where the SEi12 mini PC is much better due to better cooling. The results from the three other mini PCs show you do get some extra performance by spending a few extra hundred dollars, but for many users, the cheaper models will be more than enough.

Networking performance (Gigabit Ethernet and WiFi 6)

We’ll use iperf3 to test the gigabit Ethernet port with UP Xtreme i11 mini PC (192.168.31.12) serving as the iperf server on the other side:

• Download

PS C:\Users\aey\downloads\iperf-3.1.3-win64\iperf-3.1.3-win64> .\iperf3.exe -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  4] local 192.168.31.74 port 57011 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec  1.11 GBytes   950 Mbits/sec
[  4]  10.00-20.00  sec  1.11 GBytes   949 Mbits/sec
[  4]  20.00-30.00  sec  1.11 GBytes   949 Mbits/sec
[  4]  30.00-40.00  sec  1.11 GBytes   949 Mbits/sec
[  4]  40.00-50.00  sec  1.11 GBytes   949 Mbits/sec
[  4]  50.00-60.00  sec  1.11 GBytes   949 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-60.00  sec  6.63 GBytes   950 Mbits/sec    0             sender
[  4]   0.00-60.00  sec  6.63 GBytes   949 Mbits/sec                  receiver

iperf Done.

• Upload

PS C:\Users\aey\downloads\iperf-3.1.3-win64\iperf-3.1.3-win64> .\iperf3.exe -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  4] local 192.168.31.74 port 57015 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec  1.03 GBytes   887 Mbits/sec
[  4]  10.00-20.00  sec  1.03 GBytes   889 Mbits/sec
[  4]  20.00-30.00  sec  1.04 GBytes   890 Mbits/sec
[  4]  30.00-40.00  sec  1.03 GBytes   886 Mbits/sec
[  4]  40.00-50.00  sec  1.04 GBytes   890 Mbits/sec
[  4]  50.00-60.00  sec  1.04 GBytes   891 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-60.00  sec  6.21 GBytes   889 Mbits/sec                  sender
[  4]   0.00-60.00  sec  6.21 GBytes   889 Mbits/sec                  receiver

iperf Done.

Download is great at 950 Mbps, and upload is OK at 889 Mbps. This is one aspect where the GEEKOM Mini IT12 is better thanks to its 2.5 GbE port.

We repeated the test with WiFi 6 (5GHz) adding a Xiaomi Mi AX6000 router to the mix.

• Download

PS C:\Users\aey\downloads\iperf-3.1.3-win64\iperf-3.1.3-win64> .\iperf3.exe -t 60 -c 192.168.31.12 -i 10 -R
Connecting to host 192.168.31.12, port 5201
Reverse mode, remote host 192.168.31.12 is sending
[  4] local 192.168.31.127 port 64755 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec   896 MBytes   752 Mbits/sec
[  4]  10.00-20.00  sec   902 MBytes   757 Mbits/sec
[  4]  20.00-30.00  sec   920 MBytes   771 Mbits/sec
[  4]  30.00-40.00  sec   928 MBytes   779 Mbits/sec
[  4]  40.00-50.00  sec   924 MBytes   775 Mbits/sec
[  4]  50.00-60.00  sec   921 MBytes   772 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth       Retr
[  4]   0.00-60.00  sec  5.36 GBytes   768 Mbits/sec    6             sender
[  4]   0.00-60.00  sec  5.36 GBytes   768 Mbits/sec                  receiver

iperf Done.

• Upload

PS C:\Users\aey\downloads\iperf-3.1.3-win64\iperf-3.1.3-win64> .\iperf3.exe -t 60 -c 192.168.31.12 -i 10
Connecting to host 192.168.31.12, port 5201
[  4] local 192.168.31.127 port 64763 connected to 192.168.31.12 port 5201
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-10.00  sec   954 MBytes   800 Mbits/sec
[  4]  10.00-20.00  sec   961 MBytes   806 Mbits/sec
[  4]  20.00-30.00  sec   930 MBytes   781 Mbits/sec
[  4]  30.00-40.00  sec   974 MBytes   817 Mbits/sec
[  4]  40.00-50.00  sec   938 MBytes   787 Mbits/sec
[  4]  50.00-60.01  sec   804 MBytes   674 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval           Transfer     Bandwidth
[  4]   0.00-60.01  sec  5.43 GBytes   778 Mbits/sec                  sender
[  4]   0.00-60.01  sec  5.43 GBytes   778 Mbits/sec                  receiver

iperf Done.

768 Mbps and 778 Mbps are excellent download and upload speeds in Windows and close to the 803 Mbps and 830 Mbps transfer rates achieved with the Mini IT12. So we basically have a draw here.

Thermal performance

We used HWiNFO64 and 3DMark Fire Strike benchmarks to monitor the maximum CPU temperature under a CPU+GPU load and the maximum temperature was 91°C with some CPU throttling detected.

The CPU temperature under those conditions is quite lower than with the GEEKOM Mini IT12 mini PC as the Core i7-12650H reached a maximum of 102°C. So cooling looks to be better on the Beelink SEi12 i7-12650H which will be important for multi-core workloads and demanding games, but will likely not impact tasks such as web browsing and video playback when hardware video decoding is used.

Fan noise

The mini PC comes with a fan that’s not annoying under light loads, but it becomes noisier under heavier loads. We measured the fan noise with a sound level meter placed around 5 cm from the top of the SEi12 mini PC:

• Idle – 45 – 47 dBA
• 3DMark Fire Strike – 50 – 57 dBA

The meter measures 38-39 dBA in a quiet room.

Beelink SEi12 i7-12650H power consumption

We measured power consumption with a wall power meter:

• Power off – 0.9 to 1.1 Watt
• Idle – 18 – 19 Watts
• Web browsing – 19 to 33 Watts
• 3DMark –  19 – 33 Watts (Fire Strike)
• Video playback – 23 – 27 Watts (Youtube 8K 60 fps in Chrome)

Note: During the measurements, the mini PC was connected to WiFi 6, one USB RF dongle for a USB keyboard and mouse combo, and a VGA monitor through an HDMI to VGA adapter.

Conclusion

The Beelink SEi12 i7-12650H mini PC performs well in Windows 11 Pro with its 12th Gen Intel Core i7-12650H 10-core Alder Lake processor, 32GB RAM, and a 500GB M.2 NVMe SSD. It can handle YouTube video playback up to 8Kp60 and performs tasks like office work, web browsing, and online learning without issues. The fan is fairly quiet and is only clearly audible under heavy loads, and even then it’s not too bad.

The thermal design looks quite better than on the GEEKOM Mini IT12 with the same processor thanks to a much better multi-core score in Cinebench R23 and a lower maximal CPU temperature under a load such as 3DMark Fire Strike. The SEi12 i7-12650H connectivity options are not quite as good as the ones for the GEEKOM mini PC with no USB4 ports and gigabit Ethernet only, while many mini PCs in this price range use 2.5GbE, although WiFi 6 is working well.  The mini PC supports up to two displays with HDMI and DisplayPort video outputs, but we were only able to use one, as the DisplayPort to HDMI cable we using for testing does not seem to be compatible with the SEi12 mini PC. For reference, we could connect four displays to the GEEKOM Mini IT12 via HDMI and USB-C port. So neither one is perfect, and getting one over the other will depend on your specific needs.

We’ll now install Ubuntu 22.04 on the Beelink SEi12 i7-12650H to find out how Linux performs on the mini PC.

We’d like to thank Shenzhen AZW Technology (aka Beelink) for sending a review sample of the Beelink SEi12 i7-12650H with 32GB DDR4 and a 500GB M.2 NVMe SSD. This model can be ordered for $439 on Amazon (after ticking on the $110 discount coupon), Aliexpress (some countries only), and on the company’s online store where you can get a $50 discount with the code 1265050 valid until February 29. The GEEKOM Mini IT12 (32GB/1TB) typically sells for a little under $520, so the SEi12 model we tested is a cheaper device albeit with a smaller 500GB SSD and fewer ports.

CNXSoft: This review is a translation – with a few additional insights – of the original article on CNX Software Thailand by Suthinee Kerdkaew.

The post Beelink SEi12 i7-12650H mini PC review – Part 2: Windows 11 Pro appeared first on CNX Software - Embedded Systems News.

Google has just released the first Android 15 Developer Preview with some improvements related to privacy and security, the addition of the partial screen sharing feature, camera and audio improvements, and some new performance optimization that developers can leverage when running games or other demanding applications.

User privacy and security in Android 15

Android 15 features the latest version of the Privacy Sandbox on Android to improve user privacy while enabling personalized advertising experiences for mobile apps, the Heatlth Connect by Android adds support for new data types related to fitness, nutrition, and more, and the File integrity manager implement new APIs making use of the fs-verity feature that was added to the Linux 5.4 kernel so that files can be protected by custom cryptographic signatures.

Partial screen sharing is a completely new feature in Android 15 that allows users to share or record an app window rather than the entire device screen. It was first enabled in Android 14 QPR2 Beta, but it will be fully part of the latest version of Android during the preview and at launch later this year.

Camera and audio improvements

Android 15 adds some new camera features with low-light enhancements to boost the brightness of the camera preview and advanced flash strength adjustments to control the flash intensity in both SINGLE and TORCH modes.

Android 13 added support for connecting to MIDI 2.0 devices via USB, and Android 15 builds upon the feature adding support for virtual MIDI 2.0 devices.

Performance optimization

The Android Dynamic Performance Framework (ADPF) is a set of APIs that allow performance-intensive apps – such as games – to interact more directly with power and thermal systems of Android devices and Android 15 will add the following capabilities on supported devices:

• A power-efficiency mode for long-running background workloads.
• GPU and CPU work durations can both be reported allowing the system to adjust CPU and GPU frequencies accordingly
• Thermal headroom thresholds to interpret possible thermal throttling status based on headroom prediction.

Developers can get more information about the new features and ADPF in general on the Android developer’s website.

Android 15 Preview images and release schedule

There’s nothing ground-breaking in the improvements and new features above, but maybe others will be revealed as more people test the new images. If you are an Android app developer or simply a curious user, you’ll find the Android 15 Preview images for Pixel 6 and 6 Pro, Pixel 6a, Pixel 7 and 7 Pro, Pixel 7a, Pixel Fold, Pixel Tablet, and Pixel 8 and 8 Pro along with instructions in the relevant webpage. It’s also possible to use the Android Emulator in Android Studio if you don’t own any of those devices.

One more developer preview is scheduled for March, followed by two or three beta releases, plus one or two platform stability releases where the APIs are frozen, before the final release is outed, likely sometime in September or October.

Additional information about Android 15 can be found on the developer’s website and the preview announcement.

The post Android 15 Developer Preview released – What’s New? appeared first on CNX Software - Embedded Systems News.

Avnet MSC C10M-ALN is a COM Express Type 10 module powered by the Alder Lake-N family of processors including the Intel Core i3, Intel Atom x7000E, and Intel Processor N-Series. The design allows for easy adaptation of applications between various Intel CPU models, ensuring compatibility across different performance and power needs.

The module supports up to 16GB LPDDR5 memory with optional In-Band Error Correcting Code(IBECC), eMMC 5.1 storage, and features an Intel i226 2.5GbE controller. It can handle up to two 4K displays through DDI and eDP video outputs, ten USB ports including USB 3.2 Gen 2, and four PCI Express Gen 3 x1 slots for expanded connectivity options.

Avnet MSC C10M-ALN Com Express Module Specification:

• Alder Lake-N SoC (one or the other)
  • Intel Core i3-N305, eight-core, 1.0GHz/1.8GHz, 32EU Intel UHD Graphics, 9/15W, PUC (PC Client Use Conditions)
  • Intel Atom x7425E, four-core, 1.5GHz, 24EU Intel UHD Graphics, TCC, 12W, EUC (Embedded Use Conditions)
  • Intel Atom x7213E, two-core, 1.7GHz, 16EU Intel UHD Graphics, TCC, 10W, EUC
  • Intel Atom x7211E, two-core, 1.0GHz, 16EU Intel UHD Graphics, TCC, 6W, EUC
  • Intel Processor N200, quad-core, 1.0GHz, 32EU Intel UHD Graphics, 6W, PUC
  • Intel Processor N97, quad-core, 2.0GHz, 24EU Intel UHD Graphics, 12W, PUC
  • Intel Processor N50, dual-core, 1.0GHZ, 16EU Intel UHD Graphics, 6W, PUC
• Memory – Up to 16GB LPDDR5 SDRAM, up to 4800MT/s (soldered since LPDDR does not work on SO-DIMM) IBECC (only Atom SKUs)
• Storage
  • Up to 256GB eMMC flash
  • 2x SATA 6Gb/s interfaces
• Display Controller – Integrated  (Gen12) with up to 32 execution units (EU)
• Display Interfaces
  • 1x Digital Display Interface (DP, HDMI)
  • 1x Embedded DisplayPort (on variants without LVDS only); LVDS 24bit, single-channel (not on all variants); max. 1920 x 1200 @ 60Hz
  • Supports two independent displays
• Audio Interface – High Definition Audio (HD-A)
• Network Interface – 2.5GbE based on Intel i226 controller
• USB Ports
  • 2x USB 3.2 Gen 2 (10Gb/s)
  • 8x USB 2.0
• Serial Interfaces – 2x UARTs
• Security Device – TPM 2.0
• Miscellaneous –
  • Watchdog Timer – Initiates system reset
  • RTC battery – external
  • System Monitoring – voltage, temperature, system fan
  • GPIO, GSPI (COM.0 Rev. 3.1)
• Feature Highlights – Type 10 pin-out
• Firmware – AMI Aptio UEFI
• Power Requirement – Wide range input +5 … +20V, 5V Stby optional, 8W (typ.) up to 17W (typ.)
• Environment
  • Temperature Range – Commercial: 0° to +60°C (operating) -20° to +85°C (storage)
  • Humidity – 5 … 95% (operating, non-condensing); 5 … 95% (storage, non-condensing)
• Dimensions – 84 x 55mm
• Certificates – UL / CE
• Cooling Options
  • Passive heat sink
  • Heat spreader with threaded or non-threaded standoffs
• Carrier – Small carrier board (Mini-ITX) as part of a starter kit.

This module uses the new LPDDR5 instead of LPDDR4 delivering higher bandwidth and lower latency. previously we have seen products like AAEON PICO-ADN4 Pico-ITX SBC, and BOXER-6406-ADN have switched to using LPDDR5 memory from DDR4. The company says this board will support Windows 10 IoT Enterprise 2021 LTSC, and Yocto Project (LTS kernel 2021). This makes it great for creating point-of-sales terminals, digital signage controllers, HMI solutions, and medical equipment.

The product page mentions a small carrier board (Mini-ITX) in a starter kit, likely referring to the MSC C10-MB-EV Mini-ITX Evaluation Board. The board works with different COM Express Type 10 modules and has many connectors for easy use. T

The product is not available for purchase at the time of writing but the company indicates that the first samples will be available in Q1 of 2024. more information can be found on the press release page and a few extra tidbits of information may also be found on the products page.

The post Avnet MSC C10M-ALN – A COM Express Type 10 Mini Module with Intel Alder Lake-N CPU, LPDDR5 memory appeared first on CNX Software - Embedded Systems News.

Digi IX40 is a 5G edge computing industrial IoT cellular router solution designed for Industry 4.0 use cases such as advanced robotics, predictive maintenance, asset monitoring, industrial automation, and smart manufacturing.

The IIoT gateway is based on an NXP i.MX 8M Plus Arm processor running a custom Linux distribution, and besides 5G and 4G LTE cellular connectivity, offers gigabit Ethernet networking with 6 RJ45 and SFP ports, GNSS for geolocation and time, as well as digital and analog I/Os and an RS232/RS422/RS485 serial interface supporting Modbus.

Digi IX40 specifications:

• SoC – NXP i.MX 8M Plus Arm Cortex-A53 processor @ 1.6 GHz with 2.3 TOPS NPU
• System Memory – 1GB RAM
• Storage – 8GB eMMC flash
• Wireless
  • Cellular
    •  IX40-05
      • 5G NSA, 5G SA, 4G LTE-ADVANCED PRO CAT 19
      • 5G NR bands – n1, n2, n3, n5, n7, n8, n12, n13, n14, n18, n20, n25, n26, n28, n29, n30, n38, n40, n41, n48, n66, n71, n75, n76, n77, n78, n79;
      • 4G LTE-Advanced Pro bands –  B1, B2, B3, B4, B5, B7, B8, B12, B13, B14 FirstNet, B17, B20, B25, B26, B28, B29, B30, B32, B38, B40, B41, B42, B43, B46, B48, B66, B71;
      • 3G bands – B1, B2, B3, B4, B5, B6, B8, B19
    • IX40-04:
      • 4G LTE-ADVANCED PRO CAT 12, 3G HSPA+
      • 4G LTE-Advanced Pro bands – B1, B2, B3, B4, B5, B7, B8, B12, B13, B14, B17, B18, B19, B20, B25, B26, B28, B29, B30, B38, B39, B40, B41, B42, B43, B48, B66, B71;
      • 3G bands – B1, B2, B4, B5, B6, B8, B9, B19
    • Antenna – 2x or 4x 50 Ω SMA antenna connectors
    • 2x Nano-SIM (4FF) holders with cover plate
    • Digi SureLink, VRRP+, and dual SIMs for resilient cellular connectivity
  • GNSS
    • GPS Gen 9 with band L1 LNA for passive antenna use
    • Upper L-band: GPS L1; Glonass G1; Beidou B1; Galileo E1
    • Antenna
      • IX40-04 – supports passive GNSS antennas rated at 1.5 to 3 dBi only
      • IX40-05 – supports passive GNSS antennas rated at 1.5 to 3 dBi or active antennas with 7.5 to 26 dB gain
      • 50 Ω SMA antenna connector
    • Protocol – NMEA 4.11 and TAIP
• Wired networking
  • 4x Gigabit Ethernet RJ-45 port configurable as WAN or LAN
  • 2x Gigabit capable SFP sockets
• Serial – RS-232/422/485 DB-9 male connector (Modbus support)
• I/O – 2x Digital input/outputs; 2x analog inputs (0 – 10 V; 4 – 20 mA; 12-bit resolution)
• Security – FIPS 140-2 validation for encryption of sensitive data
• Misc – Power, WWAN, Ethernet (link, activity) LED indicators
• Power Supply – 12 to 30 VDC via screw-down terminal block
• Power Consumption – 4.5 Watts (light traffic), 8 Watts (peak data), 3.5 Watts (idle traffic)
• Dimensions – 234 x 154 x 41 mm (rugged enclosure with DIN rail and shelf mounting options)
• Weight – 1.14 kg
• Temperature Range – Operating: −30°C to 70°C; storage: −40°C to 85°C
• Relatively Humidity – 0% to 95% (non-condensing) at 25°C
• IP rating – IP30
• Approvals
  • Cellular – PTCRB, AT&T, FirstNet Capable, T-Mobile, Verizon (others may be added in the future)
  • Safety – IEC 62368-1, CB scheme
  • Emissions – FCC Part 15 Subpart B, ISED, CE, RED, UKCA

It’s quite a different beast than the earlier, more compact Digi IX30 cellular industrial router, but the Digi IX40 runs the same Digi Accelerated Linux operating system (DAL OS) with license-free enterprise software for VPN, firewall, logging, and authentication.

The gateways can be managed locally through a webUI or client (SSH, serial) or remotely using the company’s cloud-based Digi Remote Manager (aka Digi RM), SNMP v1/v2c/v3, or SMS. Digi International also lists some management tools such as SFTP, SCP, a protocol analyzer with PCAP for Wireshark, and event logging with Syslog, which typically ships with most Linux distributions…

Some of the use cases include industrial automation and control, predictive maintenance, 5G cellular and fiber failover, real-time monitoring and control for utilities, Distributed Energy Resource Management System (DERMS), Supervisory Control and Data Acquisition (SCADA), oil field drilling monitoring, and digital signage and traffic management for smart cities.

The Digi IX40-05 gateway with 5G cellular connectivity is available now, and the Digi IX40-04 with 4G LTE only is shown as coming soon. As one would expect, pricing for this type of hardware is not made public. Further information may be found on the product page and in the press release.

Thanks to TLS for the tip.

The post Digi IX40 5G edge computing industrial IoT cellular gateway is designed for Industry 4.0 use cases appeared first on CNX Software - Embedded Systems News.

If you’ve ever wondered which wireless standard may deliver the smallest lag (latency) when transmitting small packets, we’ve now gotten an answer thanks to Scott at Electric UI who benchmarked various wireless links in common MCU development boards.

More specifically the following hardware and wireless standards were tested:

• SiliconLabs 10×0-GM RF+8051 microcontroller with 240–960 MHz EZRadioPRO transceiver running SiK firmware
• HopeRF RFM95W LoRa module (on an Adafruit Breakout board) connected to an STM32F429 MCU
• Nordic Semi nRF24L01 2.4GHz transceiver module
• ESP32 board for ESP-NOW and WiFi testing is shown as ESP32 WS (WebSockets) or ESP32 TCP in the chart below. Raspberry Pi boards were also used for comparison
• ESP32-C6 board for 802.15.4 transfers (Thread)
• ESP32 and HC-05 modules for Bluetooth SPP (Serial Port Profile)
• ESP32 board with NimBLE and Bluedroid stacks and nRF52 for Bluetooth LE testing

Here are the results for 12 bytes, 128 bytes, and 1024 bytes data transfers.

nRF24 offers the lowest lag for 12-byte and 128-byte payloads with only 0.26 ms (rounded up to 300 microseconds in the main body of the article?) and 1.9 ms. That’s another story with the larger 1024-byte payload since nRF24 breaks it into multiple 32-byte packets, and instead, ESP32 TCP (WiFi) gets the upper hand here. Unsurprisingly, LoRa and SIK have really long latencies since those protocols are optimized for long-range low-power connectivity rather than fast transfers. We also learned that the Bluedroid stack has a lower latency than NimBLE on ESP32.

It’s quite tricky to measure latency between different wireless platforms since many factors have to be taken into account. For example, bare metal code will be faster than an Arduino sketch, and compilation flags may also impact performance with, for instance, the -Os flag (optimization for size) delivering better results than the -O3 flag (optimization for speed). The method to measure the lag and validate the results needs to be carefully selected and Scott goes over all those in great detail in his blog post (that’s a long read).

Via Hackaday (and somewhere on LinkedIn)

The post Comparing the latency of various wireless standards appeared first on CNX Software - Embedded Systems News.

The M1 is a multitool device that bundles several hacking and penetration tools in a package that looks like a retro-gaming console and could be viewed as a Flipper Zero alternative with a more powerful STMicro STM32H5 Cortex-M33 high-performance MCU featuring Arm TrustZone hardware-based security for additional protection for sensitive data.

The M1 multitool device features transceivers for infrared, sub-1 GHz, Bluetooth, NFC, RFID, and Wi-Fi. This means that the M1 can replace most of your remotes as well as your RFID and NFC-based items (membership cards, access fobs, business cards, credit cards, etc.) It also has twelve 3.3V (5V tolerant) GPIO pins that can be used to add extra functionality to the device.

M1 specifications:

• MCU – STM32H5-series microcontroller, with a 32-bit ARM Cortex-M33 core, 1MB RAM
• Storage – MicroSD card slot
• Display – 1.54-inch display, 128 x 64 resolution
• Connectivity
  • Bluetooth 4.2 BR/EDR BLE Sensitivity -96dBm
  • Infrared – 800nmTX/RX range, 38kHz
  • RFID –  125kHz antenna, with support for EM-4100 and HID Prox cards
  • Wi-Fi 4 – 802.11 b/g/n, 2.4Ghz
  • Sub-GHz – 142-1050Mhz, up to +20 dbm
  • NFC – 13.56MHz
• I/O
  • USB Type-C
  • 12x GPIO pins, 5V tolerant
• Buttons – 5-button directional pad, exit button
• Battery – 2,100mAh battery good for 14 days of operational time

The Flipper Zero is a portable, multipurpose hacking tool that launched on Kickstarter in 2020 and achieved massive success, raising almost 5 million dollars with a single campaign. It has since gained an app store and a video game module powered by the RP2040.

Going off the specs alone, the M1 looks like a clone of the Flipper Zero, and its tagline on Kickstarter is similar to the Flipper Zero’s. It is described as a “compact multitool for technophiles and hackers” and the Flipper is a “multitool device for hackers”. The M1’s Kickstarter campaign has been quite successful so far raising over $500,000 with all stretch goals reached for the $119 device, and still 20 days to go.

The Flipper Zero is hard to buy at the moment, driving up the demand for a product in the same category. Also, the M1 offers an upgrade to the Flipper Zero in certain areas, such as the processor, battery, wireless connectivity, and security. It remains to be seen whether the M1 can match the level of support and community offered by the Flipper Zero.

According to the makers, the project is open-source, and the firmware and blueprints for the M1 will be available in a public GitHub repository after launch. You can find more information about the M1 device on Kickstarter and the company’s website.  If you have been looking for a Flipper Zero alternative, the M1 multi-tool might be right for you. However, the makers have been reluctant to admit any similarities with the Flipper Zero when probed and we think that is quite sketchy.

The post The M1 device is a Flipper Zero alternative with a faster STM32H5 microcontroller and Wi-Fi connectivity (Crowdfunding) appeared first on CNX Software - Embedded Systems News.

Luxonis has announced its first thermal camera with the OAK Thermal (OAK-T) based on the company’s OAK-SoM Pro AI module featuring an Intel Movidius Myriad X, and two waterproof ports with an M12 PoE/Ethernet connector and an M8 auxiliary connector.

Luxonis has been making AI cameras based on Myriad X AI accelerator and its Depth AI solution at least since 2019, and its module is also found in third-party cameras as we’ve recently found out with the Arducam PiNSIGHT AI camera. But they had never made a thermal model, and following customers’ requests to fuse thermal and RGB data, they’ve now developed the OAK Thermal, or OAK-T for shorts, that is suitable for detecting leaks and fires or more accurately detect humans & animals than traditional vision-only based cameras.

OAK Thermal camera specifications:

• System-on-Module – Luxonis OAK-SoM Pro with
  • AI accelerator – Intel Movidius Myriad X AI vision processing unit (VPU) with up to 4 TOPS
  • Storage – 16 GB eMMC 5.1 flash, 32 MB QSPI NOR flash, 32 kb I2C EEPROM
• Cameras
  • Thermal
    • Sensor type – Uncooled VOx detector
    • Thermal sensor resolution: 256×192
    • Measurement range – -15°C – 150°C
    • Thermal sensitivity – 50mK @ 25°C and 25Hz
    • Thermal Accuracy – ±2°C or ±2% (whichever is larger)
    • Spectral range – 8-14µm
    • Focus –
  • Color – 2.1MP (1920×1080) IMX462 sensor up to 30 fps; focus from 30cm to infinity
  • Night vision LED
• Connectivity – M12 Ethernet/PoE connector and M8 auxiliary connector (FSYNC, I2S, GND, GPIO)
• Sensor – Bosch SensorTech BMI270 6-axis IMU sensor with 16-bit tri-axial gyroscope and a 16-bit tri-axial accelerometer
• Power consumption
  • Base consumption + PoE circuitry + camera streaming – 3W to 3.5W
  • Thermal sensor – ≤1W
  • AI subsystem consumption – ≤1W
  • Stereo depth pipeline subsystem – ≤0.5W
  • Video Encoder subsystem – ≤0.5W
  • Maximum – Under 6W
• Dimensions – 80 x 52 x 46mm
• Weight – 297 grams (with metal enclosure)
• Operating Temperature Range –  -20°C – 50°C
• IP Rating – IP67

The hardware documentation provides more details about the specifications and wiring of the M12 and M8 connectors. I could not find any software information specific to the new thermal camera, but the Depth AI SDK and Python API documentation will hopefully be updated before launch, and code samples should eventually become available on GitHub.

Luxonis expects the fusion of thermal frames, color frames, and AI models to be especially useful for the following fields and applications:

• Smart agriculture – Health crops, detect irrigation issues, detect pests, livestock health monitoring, etc.
• Industrial monitoring –  Detect overheating, detect leaks, detect fires, etc…
• Security – Detect intruders (night vision), detect fires, etc…
• Smart cities/transportation – Detect pedestrians, vehicles, and bicycles for analytics and traffic optimization.

The video below shows the output from the OAK-T camera with a kettle and two cups.

Luxonis OAK Thermal PoE smart camera is up for pre-order for $599 and will start shipping in April 2024 with a 2-meter M12 plug to RJ45 cable, a 2-meter M8 plug male-to-male cable, and a cleaning cloth. More details may also be found in the announcement.

The post Luxonis OAK Thermal – A PoE thermal camera with Myriad X AI accelerator, waterproof M12 and M8 connectors appeared first on CNX Software - Embedded Systems News.

Waveshare RP2040-GEEK is a development board that looks like a USB flash drive but is based on a Raspberry Pi RP2040 microcontroller with a 1.14-inch 65K color LCD and some expansion ports all housed in a white plastic case.

The device comes with a 4MB flash to store the firmware, a microSD card slot for data storage, a BOOT button to enter bootloader mode, two 3-pin connectors for UART and SWD debug, and a 4-pin I2C port.

Waveshare RP2040-GEEK specifications:

• MCU – Raspberry Pi RP2040 dual-core Arm Cortex-M0+ microcontroller clocked up to 133 MHz with 264 kB SRAM
• Storage – 4MB flash (W25Q32JVSSIQ) and microSD card slot
• Display – 1.14-inch 240×135 pixel 65K color IPS LCD display
• USB – 1x USB Type-A female port for power and programming
• Debugging – 3-pin SWD port for connecting a target board; the standard CMSIS-DAP interface can be used to debug most Arm-based microcontrollers; Works with OpenOCD and other tools supporting CMSIS-DAP.
• Expansion – 4-pin I2C/ADC port, 3-pin UART port
• Misc – BOOT button
• Power Supply – 5V via USB-A port
• Dimensions – Around 61 x 25 x 9 mm (plastic case)
• Weight – About 15 grams

Like the Raspberry Pi Pico, the RP2040-GEEK USB development board supports the official MicroPython and C SDKs, as well as Arduino programming with documentation available on Waveshare Wiki. Example programs include displaying an image on the LCD display, running PicoProbe firmware like in the official Raspberry Pi Debug Probe, and mounting a microSD card formatted with FAT and reading its content through the no-OS-FatFS-SD-SDIO-SPI-RPi-Pico project.

Waveshare is selling the RP2040-GEEK USB development board for $16.19 including shipping on Aliexpress, $15.99 on Amazon, or $9.99 plus shipping on its own website.  It should ship with four cables as shown below, so you don’t need to source or make your own.

Thanks to ReNa for the tip.

The post Waveshare RP2040-GEEK USB development board features RP2040 MCU, 1.14-inch color display, UART/I2C/SWD ports appeared first on CNX Software - Embedded Systems News.

Launched by Yehuo Electronic EmbedFire LubanCat 4 card computer or LubanCat 4 in short, is a Rockchip RK3588S SBC that packs quite a lot of features in an 85x56mm form factor with Ethernet, USB, mini PCIe, HDMI 2.1, SIM & microSD card holder, and more.

The board comes with up to 16GB of RAM and 128GB of eMMC flash. It comes with a Gigabit Ethernet port, five USB ports (including one USB-C), a built-in microphone, multiple audio inputs and outputs, a 40-pin Raspberry Pi compatible expansion header, and supports HDMI input through an adapter connected to a MIPI CSI port.

EmbedFire LubanCat 4 card computer specifications:

• SoC – Rockchip RK3588S
  • CPU – Octa-core processor with 4x Cortex-A76 cores @ up to 2.2-2.4 GHz, 4x Cortex-A55 cores @ up to 1.8 GHz
  • GPU –  Arm Mali-G610 GPU with OpenGL ES 3.2,  OpenCL 2.2, and Vulkan 1.2 support
  • VPU – 8Kp60 video decoder for H.265/AVS2/VP9/H.264/AV1 codecs, 8Kp30 H.265/H.264 video encoder
  • AI accelerator – 6 TOPS NPU
• System Memory – 4/8/16GB, LPDDR4X (other memory requirements can be customized)
• Storage
  • Options include No eMMC/32/64/128GB, eMMC (other storage requirements can be customized)
  • MicroSD card socket
• Video Output
  • 1x Mini HDMI 2.1 video output
  • 2x MIPI DSI connectors for wildfire MIPI screen, supports multi-screen configurations
  • 1x USB Type-C interface – supports DP protocol
  • Maximum simultaneous output:
    • Three screens configuration – 8K @ 60fps + 4K @ 60fps + 2K @ 60fps
    • Four screens configuration – 3x 4K @ 60fps + 2K @ 60fps
• Camera Input – MIPI CSI connectors
• Audio
  • 1x MIC IN for condenser microphone
  • 1x Headphone output + microphone input 2-in-1 interface
• Networking
  • Gigabit Ethernet RJ45 port
  • Optional WiFi or cellular mini PCIe module (See Expansion section)
• USB Ports:
  • 3x USB 2.0 Type-A interfaces (HOST)
  • 1x USB 3.0 Type-A interface (HOST)
  • 1x USB 3.0 Type-C interface (OTG) for firmware burning interface with DP protocol supports
• Expansion
  • Mini PCIe interface (full-height or half-height WiFi network card, 4G module)
  • 40-pin Raspberry Pi-compatible GPIO header with PWM, I2C, SPI, and UART
• Misc
  • RTC battery connection socket
  • 5V fan header for cooling
• Power Supply – 5V/4A DC input via USB Type-C (Power Interface) that has no data transmission capability
• Dimensions – 85 x 56 mm

The company provides Android 13, Debian, Ubuntu, and ROS all of which can be found in their GitHub repository. However, after looking for a while, I couldn’t find much documentation or a Wiki page for the board, so the status of the documentation is about the same as when we wrote about Rockchip RK3566/RK3568-powered LucanCat SBCs last year. Another concern about the board is its lack of a heatsink or fan; so far, the company has not provided any updates on this issue.

Previously, we’ve talked about similar SBCs like Youyeetoo R1, Indiedroid Nova, and 9Tripod Pico Pi V2.0, all of which use an SD card for storage. But this is the first time we’ve seen a Raspberry Pi-sized board powered by Rockchip RK3588S that includes a SIM card slot.

You can buy the EmbedFire LubanCat 4 card computer  on several shops on Aliexpress where the “4GB RAM + NO EMMC” version starts at $104.27, and the 16GB + 128GB version goes for $173.22 and up.

The post EmbedFire LubanCat 4 card computer – A Rockchip RK3588S dev board with a mini PCIe socket for WiFi or 4G LTE appeared first on CNX Software - Embedded Systems News.

Flipper Zero hardware & wireless hacking tool can now be used as a proper game console thanks to a Raspberry Pi RP2040-powered video game module that mirrors the display of the device on a larger monitor or TV via DVI/HDMI video output, and also adds a 6-axis motion tracking sensor.

The Flipper Zero has been in the news in recent days, notably with Canada’s government banning the device due to car theft (although it only seems feasible on older cars), and today the company has announced the launch of a video game module developed in collaboration with Raspberry Pi Ltd.

Video game module specifications:

• MCU – Raspberry Pi RP2040 dual-core Arm Cortex-M0+ microcontroller clocked up to 133 MHz with 264 kB SRAM
• Video Output – DVI-D at 640х480 with 60 Hz refresh rate. It also supports HDMI.
• USB – USB Type-C port connected to the microcontroller. Acts as a USB device or host (with the limitation that USB power delivery is not supported).
• Sensor – TDK ICM-42688-P 6-axis MEMS motion-tracking sensor (IMU) with gyroscope & accelerometer connected to RP2040 and Flipper Zero over SPI (only one can be used as the master).
• GPIO breakout – 11x GPIO pins connected to the RP2040 microcontroller, two ground pins, and one 3.3 V power pin.
• Misc – Boot button (activates bootloader mode) and Reset button

The Video game module can be used as a standalone Raspberry Pi RP2040 board running open-source Raspberry Pi Pico projects without connecting to the Flipper Zero. One example is the Scoppy open-source digital oscilloscope which we’ve previously seen being used on the FHDM TECH DSO-500K 2-channel oscilloscope board.

Before being able to use the video module with the Flipper Zero, you’ll need to update the firmware and insert it into the device with the method depending on whether a silicon case is being used or not. The quick start guide has the details. Several apps have been ported to the Flipper Zero + Video Game Module combo including the Air Arkanoid game and an Air Mouse app leveraging the 6-axis sensor in the video game module. You’ll find both in the Flipper Mobile app.

People who would like to develop games for the Flipper Zero may want to have a look at the Flipper Zero Game Engine and the source code for a demo game to get started. The firmware for the Video Game Module itself is open-source source, and you’ll also find some hardware documentation with a pinout diagram and PDF schematics on the documentation website.

The Video Game Module can be purchased for $49 on the company’s online store as well as on Lab401 for 49 Euros ex. VAT.

Via Liliputing

The post Flipper Zero gets a Raspberry Pi RP2040-powered video game module appeared first on CNX Software - Embedded Systems News.

UUGear’s Vivid Unit is a low-profile SBC with an integrated 5.5-inch 1280×720 touchscreen display powered by the older Rockchip RK3399 hexa-core Cortex-A72/A53 SoC coupled with 4GB RAM and a 32GB eMMC flash.

The board also comes with gigabit Ethernet and WiFi 4 connectivity, supports M.2 NVMe storage, offers HDMI output and a MIPI CSI camera input,  integrates a speaker and a stereo microphone, and allows for expansion through a 40-pin GPIO header and other headers for ADC and USB.

Vivid Unit specifications:

• SoC – Rockchip RK3399
  • CPU – Hexa-core big.LITTLE processor with 2x Arm Cortex-A72 cores up to 1.8GHz, 4x Arm Cortex-A53 cores up to 1.4GHz
  • GPU – Arm Mali-T860MP4 GPU
  • AI accelerator – 6 TOPS NPU
• System Memory – 4GB LPDDR4
• Storage
  • 32GB eMMC flash
  • M.2 socket for NVMe SSD
• Display – 5.5-inch touchscreen display with 1280×720 resolution
• Video Output – HDMI port
• Camera Input – MIPI CSI camera connector
• Audio
  • Built-in speaker and stereo microphone
  • 3.5mm headphone jack
• Networking
  • Low-profile gigabit Ethernet RJ45 connector
  • 2.4GHz WiFi 4 (802.11b/g/n) and Bluetooth 4.1 wireless connectivity
• USB – 2x USB 3.1 Type-A ports, 2x USB 2.0 interfaces via 4-pin headers (blue), 1x USB Type-C for power and flashing the OS
• Expansion
  • 40-pin color-coded GPIO header with SPI, I2C, UART, SDIO, and ADC interfaces
  • 4-pin ADC header (yellow)
• Misc
  • Power, Volume +/-, and MASKROM buttons
  • Power button connector, Volume +/- button connectors
  • Power LED (red), Act LED (green)
  • RTC battery connector
• Power Supply – 5V DC via USB-C port or 48V via PoE
• Dimensions – 146 x 78.2 x 19mm
• Weight – 175 grams

The Vivid Unit SBC ships with a touchscreen display, two braces, and M3 screws and standoffs. Few new products ship with the Rockchip RK3399 as it has been displaced by the newer Rockchip RK3568 or RK3588 processors, but it does benefit from a mature software stack which I initially assumed may be why UUGear selected the SoC for their new product.

They offer two OS images for the Vivid Unit: Debian 11 and Debian 11 with RetroPie preinstalled, so it’s not the latest Debian 12. The low-level software has been uploaded on GitHub, and what we have there is Linux 5.10.110 and u-boot 2017.09, so not exactly mainline stuff… Two more repositories are available for the Vivid Unit for the VGP GPIO CLI/GUI utility and the VSA screen assistant. You’ll find some more documentation on the product page.

The video demo above shows the Vivid Unit used as a voice assistant relying on Google Speech Recognition, Google Text-to-Speech, and ChatGPT-4. Other applications for the system include Smart Home automation, robotics, industrial control systems, portable computing, and retro gaming.

Dun Cat B.V (UUGear) sells the Vivid Unit for 99 Euros on its own e-shop.

The post Vivid Unit is a low-profile Rockchip RK3399 SBC with an integrated touchscreen display appeared first on CNX Software - Embedded Systems News.

ElectronicsV2 is a small development board based on the NXP S32K144 Arm Cortex-M4F microcontroller designed for automotive enthusiasts and tech hobbyists who may be interested in DIY projects such as an electric immobilizer, a CAN-based data logger, or experiment with vehicle-to-vehicle (V2V) communication

Gettobyte aims to make the ElectronicsV2 the “Arduino of the Automotive World” with affordable pricing, plenty of I/Os, a 2-pole terminal block connected to a CAN transceiver, LED and buttons, and easy-to-follow documentation and tutorials.

ElectroncisV2 specifications:

• MCU – NXP 32K144 32-bit Arm Cortex-M4F microcontroller @ up to 112 MHz with 512KB flash, 64KB SRAM, 3x FlexCAN interfaces in LFQP100 package
• CAN Bus – 2-pole terminal block connected to TJA105 CAN transceiver
• USB – 1x USB-C port for power and serial console
• Expansion – 2x 40-pin GPIO headers with CAN Bus, LPUART, ADC, LPSPI, FlexIO, EWM, LPI2C, TRGMUX
• Debugging
  • Onboard UART for debugging via USB Type-C cable
  • 10pin JTAG-SWD connector for JLink/SWD debugger (JLink V9)
• Misc
  • Onboard RGB LED
  • TX-RX LED indicator pins for onboard UART and CAN communication
  • Power LEDs
  • Two push-buttons
  • Variable potentiometer connected to ADC
  • Onboard 8 MHz crystal oscillator
  • Reset configuration DIP switch (JTAG/External/USER)
• Power Supply
  • Input – 5V via USB-C port or JTAG
  • Output – 3.3V and 5V DC output on GPIO headers
• Dimensions – TBD

Gettobyte explains the ElectronicsV2 board is programmable in C using NXP Semiconductor’s S32 Design Studio based on the Eclipse IDE, and highlights support for the Autosaur (AUTomotive Open System ARchitecture) standard through the Autosar MCAL SDK from NXP.

The company/developer provides a getting started guide and a user manual on the product page, and has also uploaded some video tutorials on YouTube.

The ElectronicsV2 board looks like a cost-down version of the official S32K144EVB development kit ($105). It is sold for $50 on the Electronics Infra store and shipped from India since it is where the board has been designed and is being manufactured. For reference, the S32K144 was introduced in 2017, and while it’s an inexpensive way to get started, some projects may benefit from the new and more powerful NXP S32K344 Arm Cortex-M7 automotive general-purpose microcontroller that features more CAN FD interfaces (6x) and a 100BaseT1 single-pair Ethernet interface.

The post ElectronicsV2 – An NXP S32K144 development board for DIY automotive projects appeared first on CNX Software - Embedded Systems News.

SMLIGHT has recently released Zigbee Ethernet/WiFi/USB coordinators and USB dongles based on either Texas Instruments CC2652P7 or CC2674P10 wireless microcontrollers that update on the company’s SLZB-06 Zigbee 3.0 to PoE Ethernet, USB, and WiFi adapter and Silicon Labs EFR32MG21-based SLZB-07 Zigbee 3.0 USB adapter.

That’s a total of four new devices with the SLZB-06p7 and SLZB-07p7 based on CC2652P7 and designed to work with vendor-agnostic software such as Zigbee2MQTT and Home Assistant ZHA, and the similar SLZB-06p10 and SLZB-06p10 based on the CC2674P10 whose Zigbee firmware is still under development according to SMLIGHT.

SLZB-06p7/SLZB-06p10 Zigbee to Ethernet/WiFi/USB coordinator

SLZB-06p7/SLZB-06p10 specifications:

• Wireless SoCs
  • SLZB-06p7 – Texas Instruments CC2652P7 Arm Cortex-M4F microcontroller @ 48-MHz with 704KB flash, 256KB ROM for protocol and library functions, 8KB of SRAM, integrated +20 dB power amplifier, Bluetooth 5.2 Low Energy, Matter, Thread, Zigbee 3.0
  • SLZB-06p10 – Texas Instruments CC2674P10 Arm Cortex-M33 microcontroller @ 48 MHz with TrustZone, FPU and DSP extension, 1024 KB flash, 8 KB of cache SRAM, 256KB of ultra-low leakage SRAM, Bluetooth 5.3 Low Energy, IEEE 802.15.4 PHY and MAC (Zigbee 3..0/Thread/Matter)
  • Common – Espressif Systems ESP32-DOWDQ5-V3 dual-core processor @ 240MHz with 448 KB ROM, 520 KB SRAM, 16 KB SRAM in RTC, WiFi and BLE connectivity
• Connectivity
  • Ethernet RJ45 port with PoE support (IEEE 802.3af) implemented through Microchip LAN8720 100M Ethernet controller
  • 2.4 GHz WiFi up to 150 Mbps
  • Zigbee 3.0 support, including Zigbee Green Power.
  • Thread/Matter support
  • +5dB SMA antenna included
• USB – USB Type-C port for data and power; USB/UART via CP2102N chip
• Misc – 4x user LEDs, 1x button
• Power Supply
  • Input – 5V/150mA via USB-C or PoE
  • Overvoltage protection, optoelectronic isolation
• Dimensions – 160 x 26 x 22mm

The SLZB-06p7 and SLZB-06p10 Zigbee/Ethernet/USB/WiFi adapter ship with a +5dB antenna. The devices are said to work with Home Assistant and Zigbee2MQTT out of the box thanks to the ESP32 firmware pre-loaded on the board with the following features:

• Update Zigbee and Core firmware OTA
• VPN support
• Localization in 15 languages
• Change Ethernet/USB/Wi-Fi adapter mode through firmware or by button
• Secure login with username and password
• DHCP or static IP address for Ethernet connection
• Customizable LEDs behavior
• Responsive web interface based on Bootstrap 5.2

The company still points to the older user manual for the original SLZB-06 launched in 2022.

SLZB-07p7 and SLZB-07p10 Zigbee 3.0 USB dongle

The SLZB-07p7 and SLZB-07p10 Zigbee 3.0 USB dongles are also based on TI CC2652P7 and CC2652P10 wireless microcontrollers and come with a +3Db antenna. They also feature the same CP2102N USB UART chip and work with ZHA Integration for Home Assistant and Zigbee2MQTT. Besides Zigbee connectivity, Thread/Matter experimental support is also available through Home Assistant. The user manual has some more detail, but as far as I can see, all information is still for the SiLabs model instead of the new ones.

SMLIGHT sells the SLZB-06p7 for $34.99, the SLZB-07p7 for $9.99, and the p10 variants can be had for $5 extra. You’ll find them on the SMLight shop and Aliexpress, Note that the p10 models are only suitable for developers right now since the Zigbee firmware is still under development.

The post SMLIGHT launches Zigbee Ethernet/WiFi coordinators and USB adapters based on TI CC2652P7 or CC2674P10 wireless chips appeared first on CNX Software - Embedded Systems News.

Duo 256M is a small board powered by SOPHGO SG2002 multi-architecture Arm/RISC-V/8051 SoC with 256MB of on-chip RAM and a 1 TOPS NPU, a microSD card for storage, a camera connector, a USB-C port for power and programming, and two headers for GPIO expansion.

We covered the SOPHGO SG2002 (and SG2000)  Arm+RISC-V+8051 AI SoC earlier this month saying a couple of boards were expected very soon. We’ve already covered Sipeed LicheeRV Nano with optional Ethernet or WiFi 6, and now we’ll look at the Duo 256M designed by Milk-V Technology in more detail since it’s available now.

Duo 256M specifications:

• SoC – SOPHGO SG2002
  • Main core – 1GHz 64-bit RISC-V C906 or Arm Cortex-A53 core (selectable)
  • Minor core – 700MHz 64-bit RISC-V C906 core
  • Low-power core – 25 to 300MHz 8051 MCU core
  • NPU – 1 TOPS INT8, supports BF16
  • Integrated 256MB DDR3 (SiP)
• Storage
  • MicroSD card slot
  • 32Gbit NAND flash (CSNP32GCR01)
• Camera Interface – 4-lane MIPI CSI input connector
• Audio – Built-in microphone
• Networking – Support for 100Mbps Ethernet via adapter connected to 5x I/O pins
• USB – 1x USB Type-C port for power and programming
• Expansion – 2x 20-pin 2.54 pitch headers with access to GPIO, PWM, I2C, SPI, UART, JTAG, ADC, Audio, Arm/RISC-V switch signal
• Misc – Power LED
• Power Supply – 5V via USB-C port
• Dimensions – TBD

You’ll find hardware documentation – including schematics, PCB layout, and datasheets – on GitHub. Instructions to getting started with buildroot (Linux 5.10 + FreeRTOS), the TDL SDK leveraging the NPU for computer vision workloads (e.g. face detection, face tracking, license plate recognition, gesture detection, etc…), and various code samples for GPIO and other features on the documentation website.

The Duo 256M board can be purchased for $7.99 plus shipping on Arace Tech. If you are interested in building your own board based on SG2000 or SG2002, samples of the chips can be purchased for respectively $6 and $5 each in packs of 5 pieces ($30 and $25).

The post Duo 256M is a compact SBC based on SG2002 multi-architecture SoC appeared first on CNX Software - Embedded Systems News.

Some of the newer AMD Ryzen processors come with an AI Engine (also called NPU or IPU) that works in Windows 11 including the Ryzen 9 7940HS, Ryzen 7 7840HS, Ryzen 5 7640HS, Ryzen 7 7840U, and Ryzen 5 7640U. I’ve just completed the review of the GEEKOM A7 mini PC powered by an AMD Ryzen 9 7940HS CPU with Windows 11 – but need to wait before publishing it – so I decided to try the AI engine on the Ryzen 9 7940HS processor.

The AI Engine relies on AMD XDNA architecture, and AMD provides instructions to get started with examples, demos, and developer resources. So I decided to try some examples, but for some reasons I’ll explain below, I eventually had to settle for some demos.

Ryzen AI software installation (for the demos)

The examples hosted on GitHub require us to install dependencies for the Ryzen AI software stack.

The first step is the installation of the IPU driver which can be done by downloading the “ipu_stack_rel_silicon_prod.zip” after registering an account on the AMD website.

Note that this requires signing a “Beta Software End User License Agreement”, and I initially assumed that was not needed here because the AMD IPU Device driver is already installed in Windows 11 Pro running on the GEEKOM A7 mini PC…

But I was wrong. I had to install the new one (and uninstall the old one) for this part to work.

We are then asked to install a few programs:

• Visual Studio 2019 (I’d recommend downloading it from a third-party website like TechSpot, because Microsoft requires login to download previous versions of VS).
• CMake version >= 3.26
• Python version >= 3.9
• Anaconda3 or (Miniconda3)

I made sure to select the option add to PATH when installing CMake, Python 3.12, and Anaconda3. Note that Visual Studio 2019 may also be used to install CMake and other tools.

The next step is to download the Ryzen AI Software installation package (ryzen-ai-sw-1.0.1.zip), but this time around, I could not avoid signing the Beta Software End User License Agreement, so I did. The agreement is confidential (although anybody who registers for an AMD account can read it) and also prevents me from reporting any results without a written agreement from AMD. So I’ll just write public information for that part…

We’ll need to run the installation script and accept another EULA for RyzenAI:

PS C:\Users\jaufr\Downloads\ryzen-ai-sw-1.0.1\ryzen-ai-sw-1.0.1> .\install.bat -env cnxsoft-ryzenai
Windows 11: OK
Visual Studio 2019: OK
Python: OK
CONDA Available: OK
CMake: OK
IPU driver Available: OK
All deps are available. Proceeding to Conda env creation...
Do you accept EULA for RyzenAI? [y/n]: y
Proceeding further ...
Creating conda env: ryzenai-1.0-20240211-173028 ...
Collecting package metadata (repodata.json): done
...

If there are any errors (look for CRITICAL string) or the script does not complete, you’ll want to correct that by installing the relevant programs and making sure they are in the PATH.

The next step is to activate the conda environment…

conda activate cnxsoft-ryzenai

… and run the test:

cd .\quicktest\
curl https://www.cs.toronto.edu/~kriz/cifar-10-sample/bird6.png -o image_0.png
python -m pip install -r requirements.txt
python quicktest.py --ep ipu

Once successful, the output should look like that as per the documentation:

I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:336] Vitis AI EP Load ONNX Model Success
I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:337] Graph Input Node Name/Shape (1)
I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:341]   input : [-1x3x32x32]
I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:347] Graph Output Node Name/Shape (1)
I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:351]   output : [-1x10]
I20231127 16:29:15.010130 13892 vitisai_compile_model.cpp:226] use cache key quickstart_modelcachekey
[Vitis AI EP] No. of Operators :   CPU     2    IPU   400  99.50%
[Vitis AI EP] No. of Subgraphs :   CPU     1    IPU     1 Actually running on IPU     1
....

It’s a bit longer than that, but you get the idea.

Yolov8 example with Ryzen AI software

Once the installation is complete, we can try some of the examples, and I thought to check out the Yolov8 example.  I spent a few hours on it, but eventually failed, and I did not want to spend more time on it since I couldn’t write about it to the Beta software agreement. But AMD published a video about the similar Yolov8_e2e tutorial from the same GitHub repository about two months ago. Some of the versions have changed, but the procedure is similar and requires quite a few more steps than the example.

Ryzen AI demos

I then switched to the demos directory in the same GitHub repository, and I was not asked to sign any legal documents… I started with the Ryzen AI “cloud-to-client” demo that showcases searching and sorting images on AMD Ryzen AI-based PCs using Yolov5 and Retinaface AI models. I used this with the older IPU driver that shipped with Windows 11 Pro in the GEEKOM A7 mini PC.

The first step was to run the setup batch script:

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\cloud-to-client> .\setup.bat

It ended with a “critical” file missing, but I was still able to run the one of the programs.

2024-02-11 18:16:24,015 - INFO - copying C:\Windows\System32\AMD\xrt_core.dll to C:\Users\jaufr\anaconda3\envs\ms-build-demo\lib\site-packages\onnxruntime\capi
2024-02-11 18:16:24,015 - INFO - copying C:\Windows\System32\AMD\xrt_coreutil.dll to C:\Users\jaufr\anaconda3\envs\ms-build-demo\lib\site-packages\onnxruntime\capi
2024-02-11 18:16:24,015 - INFO - copying C:\Windows\System32\AMD\xrt_phxcore.dll to C:\Users\jaufr\anaconda3\envs\ms-build-demo\lib\site-packages\onnxruntime\capi
2024-02-11 18:16:24,037 - CRITICAL - C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\cloud-to-client\voe-0.1.0-cp39-cp39-win_amd64\onnxruntime.dll does not exist.
PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\cloud-to-client>

You’ll find the log on pastebin.

I then run the program Qt demo:

run_pyapp.bat

I selected a folder with screenshots (I reckon it may not be the best for this type of test but that is what I had on the test PC), and the AMD Image Categorizer scanned my images, and I could see some NPU usage in HWiNFO64. See full output while scanning for reference.

Each image is now searchable and although many tags are not super relevant because of the source images used, it could find a bird in a screenshot I had done with YouTube.

The cloud-to-client demo is also available as a web server, but it did not work for me, even after installing some extra Python modules.

I’d kind of expect demos to work out of the box without having to spend hours getting the right module or DLL version although my recent experience with the Orbbec Femto Mega depth camera in Windows 11 seems to imply this is business as usual in Windows…

There’s also another demo called “multi-model-exec” with the following models:

• MobileNet_v2
• ResNet50
• Retinaface
• Segmentation
• Yolox

Sadly all failed me with the same error. I had to change some of the commands from the instructions or I would have had to stop earlier…

I first had to create a conda environment:

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec> conda env create --name cnxsoft --file=env.yaml
Collecting package metadata (repodata.json): done
Solving environment: done

Full output:

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec> conda env create --name cnxsoft --file=env.yaml
Collecting package metadata (repodata.json): done
Solving environment: done


==> WARNING: A newer version of conda exists. <==
  current version: 23.7.4
  latest version: 24.1.0

Please update conda by running

    $ conda update -n base -c defaults conda

Or to minimize the number of packages updated during conda update use

     conda install conda=24.1.0



Downloading and Extracting Packages

Preparing transaction: done
Verifying transaction: done
Executing transaction: done
Installing pip dependencies: \ Ran pip subprocess with arguments:
['C:\\Users\\jaufr\\anaconda3\\envs\\cnxsoft\\python.exe', '-m', 'pip', 'install', '-U', '-r', 'C:\\Users\\jaufr\\Downloads\\RyzenAI-SW-main\\RyzenAI-SW-main\\demo\\multi-model-exec\\condaenv.q1_003sq.requirements.txt', '--exists-action=b']
Pip subprocess output:
Processing c:\users\jaufr\downloads\ryzenai-sw-main\ryzenai-sw-main\demo\multi-model-exec\voe-0.1.0-cp39-cp39-win_amd64.whl (from -r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 2))
Collecting onnxruntime (from -r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Obtaining dependency information for onnxruntime from https://files.pythonhosted.org/packages/6d/22/f84599edb744a06ba86920f51a2f9d5317db2dc496876eb32831f7923196/onnxruntime-1.17.0-cp39-cp39-win_amd64.whl.metadata
  Using cached onnxruntime-1.17.0-cp39-cp39-win_amd64.whl.metadata (4.3 kB)
Collecting coloredlogs (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Using cached coloredlogs-15.0.1-py2.py3-none-any.whl (46 kB)
Collecting flatbuffers (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Obtaining dependency information for flatbuffers from https://files.pythonhosted.org/packages/6f/12/d5c79ee252793ffe845d58a913197bfa02ae9a0b5c9bc3dc4b58d477b9e7/flatbuffers-23.5.26-py2.py3-none-any.whl.metadata
  Using cached flatbuffers-23.5.26-py2.py3-none-any.whl.metadata (850 bytes)
Collecting numpy>=1.21.6 (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Obtaining dependency information for numpy>=1.21.6 from https://files.pythonhosted.org/packages/b5/42/054082bd8220bbf6f297f982f0a8f5479fcbc55c8b511d928df07b965869/numpy-1.26.4-cp39-cp39-win_amd64.whl.metadata
  Using cached numpy-1.26.4-cp39-cp39-win_amd64.whl.metadata (61 kB)
Collecting packaging (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Obtaining dependency information for packaging from https://files.pythonhosted.org/packages/ec/1a/610693ac4ee14fcdf2d9bf3c493370e4f2ef7ae2e19217d7a237ff42367d/packaging-23.2-py3-none-any.whl.metadata
  Using cached packaging-23.2-py3-none-any.whl.metadata (3.2 kB)
Collecting protobuf (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Obtaining dependency information for protobuf from https://files.pythonhosted.org/packages/f3/7c/9e78d866916fb07e193a53352453fdc44a9a47d5c30866c40231a03eb3a6/protobuf-4.25.2-cp39-cp39-win_amd64.whl.metadata
  Using cached protobuf-4.25.2-cp39-cp39-win_amd64.whl.metadata (541 bytes)
Collecting sympy (from onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Using cached sympy-1.12-py3-none-any.whl (5.7 MB)
Collecting glog==0.3.1 (from voe==0.1.0->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 2))
  Using cached glog-0.3.1-py2.py3-none-any.whl (7.8 kB)
Collecting python-gflags>=3.1 (from glog==0.3.1->voe==0.1.0->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 2))
  Using cached python_gflags-3.1.2-py3-none-any.whl
Collecting six (from glog==0.3.1->voe==0.1.0->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 2))
  Using cached six-1.16.0-py2.py3-none-any.whl (11 kB)
Collecting humanfriendly>=9.1 (from coloredlogs->onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Using cached humanfriendly-10.0-py2.py3-none-any.whl (86 kB)
Collecting mpmath>=0.19 (from sympy->onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Using cached mpmath-1.3.0-py3-none-any.whl (536 kB)
Collecting pyreadline3 (from humanfriendly>=9.1->coloredlogs->onnxruntime->-r C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\condaenv.q1_003sq.requirements.txt (line 1))
  Using cached pyreadline3-3.4.1-py3-none-any.whl (95 kB)
Using cached onnxruntime-1.17.0-cp39-cp39-win_amd64.whl (5.6 MB)
Using cached numpy-1.26.4-cp39-cp39-win_amd64.whl (15.8 MB)
Using cached flatbuffers-23.5.26-py2.py3-none-any.whl (26 kB)
Using cached packaging-23.2-py3-none-any.whl (53 kB)
Using cached protobuf-4.25.2-cp39-cp39-win_amd64.whl (413 kB)
Installing collected packages: python-gflags, pyreadline3, mpmath, flatbuffers, sympy, six, protobuf, packaging, numpy, humanfriendly, glog, coloredlogs, voe, onnxruntime
Successfully installed coloredlogs-15.0.1 flatbuffers-23.5.26 glog-0.3.1 humanfriendly-10.0 mpmath-1.3.0 numpy-1.26.4 onnxruntime-1.17.0 packaging-23.2 protobuf-4.25.2 pyreadline3-3.4.1 python-gflags-3.1.2 six-1.16.0 sympy-1.12 voe-0.1.0

done
#
# To activate this environment, use
#
#     $ conda activate cnxsoft
#
# To deactivate an active environment, use
#
#     $ conda deactivate

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec>

We can now enable activate the environment:

conda activate cnxsoft

I download the ONNX models and test image/video package (resource_multi_model_demo.zip), and unzip it under demo/multi-model-exec/ipu_modelsx4_demo/. After that, I could run the command to generate the scripts:

It looks good so far:

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo> ls


    Directory: C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo


Mode                 LastWriteTime         Length Name
----                 -------------         ------ ----
d-----         2/11/2024   4:46 PM                config
d-----         2/11/2024   8:45 PM                resource
-a----         2/11/2024   4:46 PM           7386 generate_script.py
-a----         2/11/2024   8:45 PM            741 run_mobile_net_v2.bat
-a----         2/11/2024   8:45 PM            735 run_modelx4.bat
-a----         2/11/2024   8:45 PM            736 run_resnet50.bat
-a----         2/11/2024   8:45 PM            738 run_retinaface.bat
-a----         2/11/2024   8:45 PM            740 run_segmentation.bat
-a----         2/11/2024   8:45 PM            734 run_yolovx.bat

But trying any of the scripts fail with the same error:

PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo> .\run_mobile_net_v2.bat

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_VART_FIRMWARE=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\1x4.xclbin

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set PATH=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\bin;C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\python;C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..;C:\Windows\System32\AMD;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0\;C:\WINDOWS\System32\OpenSSH\;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\Program Files\PuTTY\;C:\Program Files\CMake\bin;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Scripts\;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\;C:\Users\jaufr\anaconda3;C:\Users\jaufr\anaconda3\Library\mingw-w64\bin;C:\Users\jaufr\anaconda3\Library\usr\bin;C:\Users\jaufr\anaconda3\Library\bin;C:\Users\jaufr\anaconda3\Scripts;C:\Users\jaufr\AppData\Local\Programs\Python\Launcher\;C:\Users\jaufr\AppData\Local\Microsoft\WindowsApps

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set PYTHONPATH=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\python312.zip;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\DLLs;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Lib;C:\Users\jaufr\AppData\Local\Programs\Python\Python312;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Lib\site-packages;set DEBUG_ONNX_TASK=0

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set DEBUG_DEMO=0

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set NUM_OF_DPU_RUNNERS=4

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_ENABLE_GRAPH_ENGINE_PAD=1

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_ENABLE_GRAPH_ENGINE_DEPAD=1

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\bin\ipu_multi_models.exe C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\config\mobile_net_v2.json
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20240211 20:47:53.260922 21700 ipu_multi_models.cpp:136]  config not set using_onnx_ep, using default: false
I20240211 20:47:53.260922 21700 ipu_multi_models.cpp:376] config mobile_net_v2 -> model_filter_id:5 thread_num:4 confidence_threshold:0.3 onnx_model_path:C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\resource\mobilenetv2_1.4_int.onnx video_file_path:C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\resource\detection.avi onnx_x:1 onnx_y:1 onnx_disable_spinning:0 onnx_disable_spinning_between_run:0 intra_op_thread_affinities: using_onnx_ep:0
I20240211 20:47:53.262941 21700 ipu_multi_models.cpp:332] g_show_width: 1024g_show_height: 640matrix_split_num: 1
I20240211 20:47:53.262941 21700 ipu_multi_models.cpp:340] use global gui thread
I20240211 20:47:53.486811 21700 onnx_task.hpp:113] using VitisAI
I20240211 20:47:56.181272 21700 ipu_multi_models.cpp:382] C:\Users\xbuild\Desktop\xj3\VAI_RT_WIN_ONNX_EP_ALL\onnxruntime\onnxruntime\core\providers\vitisai\imp\global_api.cc:56 OrtVitisAIEpAPI::Ensure [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\bin\onnxruntime_vitisai_ep.dll"
PS C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo> .\run_resnet50.bat

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_VART_FIRMWARE=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\1x4.xclbin

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set PATH=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\bin;C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\python;C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..;C:\Windows\System32\AMD;C:\WINDOWS\system32;C:\WINDOWS;C:\WINDOWS\System32\Wbem;C:\WINDOWS\System32\WindowsPowerShell\v1.0\;C:\WINDOWS\System32\OpenSSH\;C:\Users\Administrator\AppData\Local\Microsoft\WindowsApps;C:\Program Files\PuTTY\;C:\Program Files\CMake\bin;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Scripts\;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\;C:\Users\jaufr\anaconda3;C:\Users\jaufr\anaconda3\Library\mingw-w64\bin;C:\Users\jaufr\anaconda3\Library\usr\bin;C:\Users\jaufr\anaconda3\Library\bin;C:\Users\jaufr\anaconda3\Scripts;C:\Users\jaufr\AppData\Local\Programs\Python\Launcher\;C:\Users\jaufr\AppData\Local\Microsoft\WindowsApps

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set PYTHONPATH=C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\python312.zip;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\DLLs;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Lib;C:\Users\jaufr\AppData\Local\Programs\Python\Python312;C:\Users\jaufr\AppData\Local\Programs\Python\Python312\Lib\site-packages;set DEBUG_ONNX_TASK=0

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set DEBUG_DEMO=0

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set NUM_OF_DPU_RUNNERS=4

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_ENABLE_GRAPH_ENGINE_PAD=1

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>set XLNX_ENABLE_GRAPH_ENGINE_DEPAD=1

C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo>C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\..\bin\ipu_multi_models.exe C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\config\resnet50.json
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20240211 20:52:41.000353 20180 ipu_multi_models.cpp:83] resnet50 config not set confidence_threshold, using default: 0.3
I20240211 20:52:41.010800 20180 ipu_multi_models.cpp:136]  config not set using_onnx_ep, using default: false
I20240211 20:52:41.010800 20180 ipu_multi_models.cpp:376] config resnet50 -> model_filter_id:2 thread_num:4 confidence_threshold:0.3 onnx_model_path:C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\resource\resnet50_pt.onnx video_file_path:C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\ipu_modelsx4_demo\resource\detection.avi onnx_x:1 onnx_y:1 onnx_disable_spinning:0 onnx_disable_spinning_between_run:0 intra_op_thread_affinities: using_onnx_ep:0
I20240211 20:52:41.010800 20180 ipu_multi_models.cpp:332] g_show_width: 1024g_show_height: 640matrix_split_num: 1
I20240211 20:52:41.010800 20180 ipu_multi_models.cpp:340] use global gui thread
I20240211 20:52:41.210458 20180 onnx_task.hpp:113] using VitisAI
I20240211 20:52:41.910542 20180 ipu_multi_models.cpp:382] C:\Users\xbuild\Desktop\xj3\VAI_RT_WIN_ONNX_EP_ALL\onnxruntime\onnxruntime\core\providers\vitisai\imp\global_api.cc:56 OrtVitisAIEpAPI::Ensure [ONNXRuntimeError] : 1 : FAIL : LoadLibrary failed with error 126 "" when trying to load "C:\Users\jaufr\Downloads\RyzenAI-SW-main\RyzenAI-SW-main\demo\multi-model-exec\bin\onnxruntime_vitisai_ep.dll"

Windows programs often have cryptic error messages and don’t know what to do with:

FAIL : LoadLibrary failed with error 126 "" when trying to load "...\multi-model-exec\bin\onnxruntime_vitisai_ep.dll"

The file is there but the system complain about something for which a web search did not help. I’m suspecting some mismatch between versions… It’s probably a matter of spending a few hours or days to fix the issue, and I’m not going to do that.

So the NPU/IPU in the Ryzen 9 7940HS can be used in Windows 11, but that’s not exactly a straightforward task right now, and I’m not sure any programs for end-users have been released just yet… Let me know in the comments if you know any.

The post Testing AMD Ryzen 9 7940HS’ AI Engine in Windows 11 appeared first on CNX Software - Embedded Systems News.

SparkFun has launched yet another ESP32-C6 board with the “Thing Plus – ESP32-C6”  based on the ESP32-C6-WROOM-1-N16 module with 16MB flash and a PCB antenna and  range of I/Os and power options.

The board features 28 through holes with up to 23 multi-function GPIOs and a Qwicc connector for expansion, and supports 5V or LiPo battery power through respectively a USB-C port a 2-pin JST connector combined with a charging chip, and a fuel gauge.

SparkFun Thing Plus – ESP32-C6 specifications:

• Wireless module – ESP32-C6-WROOM-1-N16
  • MCU – ESP32-C6 32-bit single-core RISC-V microcontroller with 2.4 GHz WiFI 6, Bluetooth 5 LE, and 802.15.4 radio (Zigbee and Thread); Matter-compatible
  • Storage – 16 MB flash
  • PCB Antenna
• Storage – MicroSD card slot
• USB – 1x USB Type-C port for power and programming
• Expansion
  • 12-pin + 16-pin headers with
    • 23x multifunctional GPIOs
    • Up to 7x 12-bit ADC channels
    • Up to 2x UART channels (with flow control)
    • 1x Low Power UART, 1x I2C, 1x Low Power I2C
    • LED PWM
    • 1x I2S Channel
  • 4-pin Qwiic connector
• Misc
  • LEDs for Power (Red), charging status (Yellow), and system status (Blue)
  • 1x WS2812 RGB LED
  • BOOT and RESET buttons
• Power Supply
  • 5V via USB-C port
  • 2-pin JST connector for LiPo battery
  • MCP73831 battery charger – Charge Rate: 213mA @ 3.3V
  • MAX17048 fuel gauge – I2C Address: 0x36
• Dimensions – 64.8 x 22.9 mm; 2x mounting holes

The ESP32-C6 board is open-source hardware with Sparkfun providing the hardware design files (EAGLE and Gerber files),  some datasheets, and documentation on GitHub. I’m not sure who thought it was a good idea but the documentation is distributed as MD files and provides instructions to get started with Arduino programming.

The market for ESP32-C board is starting to get crowded starting from the cheap WeAct ESP32-C6 to DFRobot FireBeetle 2 also supporting LiPo batteries, or even Olimex ESP32-C6-EVB with more features for home automation, and many others. The “Sparkfun Things Plus – ESP32-C6” board adds another option with a larger flash (16MB) than most competitors and compatibility with Qwiic modules and the Things Plus form factor.

The company is selling the board for $19.95 plus shipping on its store.

Thanks to TLS for the tip.

The post SparkFun Thing Plus – ESP32-C6 board comes with 16MB flash, LiPo battery support appeared first on CNX Software - Embedded Systems News.

Intel decided to pull the plug on its NUC business in 2023, but that does not mean those are not available for sale anymore, and the Intel NUC 11 (NUC11PAHi7) powered by an Intel Core i7-1165G7 SoC with 16GB RAM and a 512GB SSD is now offered on Amazon with a 5% discount when using the coupon code A5QFKRCQ.

Introduced in 2021, the Intel NUC11PAHi7 supports up to four 4K displays through HDMI, Mini DP, and Thunderbolt interfaces, offers 2.5GbE and WiFi 6 connectivity, three USB 3.2 Gen 2 ports, and even a CIR (Consumer InfraRed) port.

Intel NUC11PAHi7 specifications:

• SoC – Intel Core i7-1165G7 quad-core/octa-thread Tiger Lake UP3 processor @ up to 2.8 GHz / 4.7 GHz (Turbo) with 96EU Iris Xe Graphics; 15W TDP (configurable between 12 and 28W)
• System Memory – 16GB DDR4-2666 via SO-DIMM sockets expandable up to 64GB RAM
• Storage
  • 512GB M.2 NVMe (PCIe Gen4) SSD
  • SD card slot with UHS-II support
  • 2.5-inch SATA slot for (up to 7mm  thick HDD)
• Video Output
  • HDMI 2.0b
  • Mini DisplayPort 1.4 up to 4Kp60
  • DisplayPort output via Thunderbolt 3 ports
  • 4x independent displays support
• Audio
  • 3.5mm headphone jack supporting 7.1 multi-channel
  • Far-field quad-mic array with Alexa support
• Networking
  • 2.5 Gbps Ethernet RJ45 jack via Intel Ethernet Controller i225-V
  • WiFi 6 & Bluetooth 5 module
• USB
  • 3x USB 3.2 Gen 2 ports
  • 2x Thunderbolt 3 USB-C ports
• Misc
  • Power button
  • Kensington security slot
  • Consumer infrared (CIR)
• Power Supply – 19V via DC jack
• Dimensions – 137x150x60 mm

The Intel NUC 11 comes pre-loaded with Windows 11 Pro and ships with a US power adapter, a mounting bracket, and a user manual.

Intel scrapped the Intel NUC models from its website but is still selling the NUCs on the Intel Store on Amazon through official distributors. The Intel NUC 11 covered here is sold by “GEEK+ Computer Mall (Intel Authorized Reseller)” which offers a 3-year global warranty in collaboration with Intel.

The post 5% discount for Intel NUC 11 with Core i7-1165G7 CPU, 16GB RAM, 512GB SSD (Sponsored) appeared first on CNX Software - Embedded Systems News.

STMicroelectronics has recently introduced the STM32WL5MOC system in package (SiP) module with a dual-core STM32 microcontroller, sub-1 GHz RF radio, power supply, and passive components into a 10×10 mm LGA package. According to ST, the new chip uses the STM32WL module which we have seen used in Arduino MKR-inspired MKR Windy board, smart building, and many other LoRa devices.

STMicroelectronics’ STM32WL, an Arm Cortex-M0+ microcontroller, operates in sub-GHz ISM bands (413-479MHz, 826-958MHz, and 169MHz later in 2024) for protocols like wireless M-Bus (mode N) and Wize. It supports multi-protocol and multi-modulation (4-(G)FSK, 2-(G)FSK, (G)MSK, DBPSK, DSSS, OOK, ASK) for various wireless standards (Sigfox, KNX, WiSun, mioty, M-Bus, etc.) and introduces power-saving features for up to 15 years of battery life.

STM32WL5MOC SiP module specifications:

• Core Specifications:
  •  STM32WL55JC SoC with 32-bit Arm Cortex-M4 and Cortex-M0+ CPUs, up to 48 MHz.
  • Adaptive real-time accelerator (ART Accelerator) for efficient flash memory execution.
  • DSP instructions support and memory protection unit (MPU) included.
• Memory:
  • 256-Kbyte flash memory and 64-Kbyte RAM.
  • 20x 32-bit backup register and bootloader supporting USART and SPI interfaces
  • Sector protection against read/write operations
• Radio and Connectivity:
  • Frequency range – 150 MHz to 960 MHz, supporting LoRa, (G)FSK, (G)MSK, and BPSK modulations.
  • RX sensitivity: –123 dBm for 2-FSK, –148 dBm for LoRa.
  • High and low transmitter output power, programmable up to +22 dBm and +15 dBm respectively.
  • Integrated passive devices (IPD) for optimized RF matching, filtering, and balun.
  • Compliance with ETSI, FCC, and ARIB STD radio frequency regulations.
  • Supports both standardized and proprietary protocols like LoRaWAN, Sigfox, and W-MBus.
• Security and Identification:
  • AES 256-bit hardware encryption, true random number generator (RNG).
  • Sector protection (PCROP, RDP, WRP) and CRC calculation unit.
  • Unique device identifier (64-bit UID) and 96-bit unique die identifier.
  • Hardware public key accelerator (PKA) and secure sub-GHz MAC layer.
  • Secure firmware update (SFU) and install (SFI) capabilities.
• Clock Sources:
  • Multiple internal and external clock sources including a 32 MHz crystal oscillator, TCXO support, and a range of RC oscillators.
  • PLL for CPU, ADC, and audio clocks.
• Ultra-Low-Power Platform:
  • The power supply ranges from 1.8 V to 3.6 V.
  • Ultra-low power consumption modes include shutdown (31 nA), standby (+ RTC) mode (360 nA), and Stop2 (+ RTC) mode (1.07 µA).
  • Active-mode MCU consumption below 72 µA/MHz, RX mode at 4.82 mA, and TX mode ranging from 15 mA to 87 mA.
• Analog and System Peripherals:
  • 12-bit ADC with hardware oversampling, 12-bit DAC, and ultra-low-power comparators.
  • Mailbox and semaphores for inter-core communication.
• Controllers and I/Os:
  • Comprehensive controller support including DMA, USART, LPUART, SPI, and I2C.
  • Multiple timers for general-purpose, motor control, and ultra-low-power applications.
  • Up to 43 I/Os, most of which are 5 V-tolerant.
• Operating temperature from –40 °C to +105 °C.

B-WL5M-SUBG1 development board specification:

• Embeds STM32WL5MOC module for immediate prototyping
• User Interface – Three user LEDs, user and reset push-buttons for interaction
• Connectivity & Expansion – MIPI debug connector, STMod+ connector for modular expansion, stubby antenna for robust wireless connectivity
• Power Management – USB Type-C for power through an add-on STMod+ adapter board, supports external power sources or USB VBUS
• Development Support – Offers comprehensive software libraries and examples with the STM32CubeWL MCU Package, supports major IDEs including IAR Embedded Workbench, MDK-ARM, and STM32CubeIDE for versatile development options

The B-WL5M-SUBG1 development board includes a temperature sensor, 3-axis magnetometer, 3D accelerometer, 3D gyroscope, piezoresistive absolute pressure sensor, CMOS serial flash, and serial I2C bus EEPROM. Additionally, the 20-pin STMod+ interface allows adding small-form-factor daughterboards.

Software support is available for the new SiP through the STM32CubeWL MCU package, featuring low-level APIs, a hardware abstraction layer (HAL), RTOS, and software stacks for LoRaWAN and Sigfox. Additionally, third-party IDEs such as IAR Embedded Workbench will also support this new MCU.

The chip ($9 to $10) and development board ($52.50) can currently be purchased through STMicroelectronics’ online store or from various other distributors like DigiKey Mouser. For more information, check out their press release and the product pages for the module and evaluation kit.

The post STMicro STM32WL5MOC SiP Module is pre-certified for LoRaWAN & Sigfox networks appeared first on CNX Software - Embedded Systems News.