Page 23 – Electronics-Lab.com

3 Phase Inverter for Low Power AC Motor

This is a 3-Phase Inverter Driver project designed for Low Power AC Motor. The FSB50450AS chip is the heart of the project. A compact inverter project is ideal for small power motors such as fans and pumps. An H-bridge configuration can also be created for low-current high-voltage brushed DC motors. The Inverter project supports loads up to 1.5A. The DC power supply voltage is up to 400V DC. The project operates with 6 x PWM signals. Parameters for bootstrap circuit elements are dependent on the PWM algorithm. For 15 kHz switching frequency, Bootstrap capacitors C1, C2, and C3 value is 1uF. Resistor R1, R2, R3, R4, R6, R7 and capacitor C4, C5, C6, C7, C8, C9 prevent improper signal due to surge-noise. Over-current or fault detection circuit is created using comparator CN5 – LMV7239 and shunt resistor R5. The output of LMV7239 is normally high and goes low when an over-current condition occurs. So, the CS pin goes low when the current goes above the threshold level. ATS pin is provided to monitor the temperature of the IPM chip. Heatsink is a must for this IPM module. Adhesive material with high thermal conductivity, such as Loctite 384, can be used to fix the heat sink on the top surface of the chip. Refer to the heatsink installation diagram.

Note: The project operates with lethal voltage, user must take care of safety and all necessary precautions before testing the project.

Operation Truth Table

HIN 0, LIN 0 = Output Z Both MOSFET Off
HIN 0, LIN 1 = Output 0 Low Side MOSFET On
HIN 1, LIN 0 = Output VDC High Side MOSFET On
HIN 1, LIN 1 = Forbidden Shoot Through

Features

Load Power Supply Up to 400V DC
Load Current Continues 1.5A
Logic Supply 15V DC
Power Supply for Over Current Circuit (Comparator CN5) 5V DC
One Board Power LED for Logic Supply
Shunt Based Over Current Output
V_TS Pin for Temperature Monitor (Temperature sensing built in HVIC)
HVIC for gate driving and under-voltage protection
Active-High interface, can work with 3.3V/5V logic (PWM Signals)
Optimized for low electromagnetic interference
Isolation voltage rating of 1500Vrms for 1min
PCB Dimensions 50.17 x 45.72 mm

Temperature Sensing Output (VTS) – CN1 Pin 4-VTS

Pin: VTS This indicates the temperature of the V-phase HVIC with analog voltage. HVIC itself creates some power loss, but mainly heat generated from the MOSFETs increases the temperature of the HVIC. This Pin provides voltage output when the temperature rises, this pin can be interfaced with the ADC of the MCU to monitor the temperature. Refer to the Temperature vs V_TS output diagram for more information.

Signal Input Pins

Pins: IN(UL), IN(VL), IN(WL), IN(UH), IN(VH), IN(WH)
These pins control the operation of the MOSFETs.
These pins are activated by voltage input signals. The terminals are internally connected to the Schmitt trigger circuit.
The signal logic of these pins is active HIGH; the MOSFET turns ON when sufficient logic voltage is applied to the associated input pin.
The wiring of each input needs to be short to protect the module against noise influences.
An RC filter is used to mitigate signal oscillations or any noise that traces of input signals may pick up.

Minimum Pulse Width (Signal Input)

There are input noise filters of 90 ns time constant inside the HVIC. It screens out pulses narrower than the filter time constant. Additional propagation delay in level-shifters and other circuits, together with gate charging time, prevent SPM 5 products from responding to an input pulse narrower than ~120 ns. Gate signal inputs are active-HIGH with 500 kΩ internal pull-down resistors.

FSB50450AS is an advanced Motion SPM® 5 series based on fast-recovery MOSFET (FRFET®) technology as a compact inverter solution for small power motor drive applications such as fans and pumps. FSB50450AS contains six FRFET MOSFETs, three half-bridge gate driver HVICs with temperature sensing, and three bootstrap diodes in a compact package fully isolated and optimized for thermal performance. FSB50450AS features low electromagnetic interference (EMI) characteristics through optimizing switching speed and reducing parasitic inductance. Since FSB50450AS employs MOSFETs as power switches, it provides much more ruggedness and larger safe operating area (SOA) than IGBT-based power modules. FSB50450AS is the right solution for compact and reliable inverter designs where the assembly space is constrained.

Connections

CN1: Pin 1 VDD 12V to 15V DC, Pin 2 GND, Pin 3 = VCC 5V DC, Pin 4 = VTS/Temperature NTC, Pin 5 = Over Current/Fault Output, Pin 6 = GND
CN2: Pin 1 = PWM U-High, Pin 2 = PWM U-Low, Pin 3 = PWM V-High, Pin 4 = PWM V-Low, Pin 5 = PWM W-High, Pin 6 = PWM W-Low, Pin 7 = GND
CN3: 3 Phase Motor, Pin 1 = Phase=W, Pin 2 = Phase-V, Pin 3 = Phase-U
CN4: Pin 1 = Motor Supply 400V DC, Pin 2 = GND
D1: Power LED for Logic Supply

Block Diagram

Typical Application

Schematic

Parts List

NO.	QNTY.	REF.	DESC.	MANUFACTURER	SUPPLIER PART NO
1	1	CN1	6 PIN MALE HEADER PITCH 2.54MM	WURTH	732-5319-ND
2	1	CN2	7 PIN MALE HEADER PITCH 2.54MM	WURTH	732-5320-ND
3	1	CN3	3 PIN SCREW TERMINAL PITCH 5.08MM	PHOENIX	277-1248-ND
4	1	CN4	2 PIN SCREW TERMINAL PITCH 5.08MM	PHOENIX	277-1247-ND
5	1	CN5	LMV7239 SOT23-5	TI	LMV7239M5X/NOPBCT-ND
6	3	C1,C2,C3	1uF/35V CERAMIC SMD SIZE 1206	YAGEO/MURATA
7	8	C4,C5,C6,C7,C8,C9,C16,C17	1nF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA
8	1	C10	10uF/450V ELECTROLYTIC 12.5MM DIA	PANASONIC	P13558-ND
9	1	C11	0.1uF/630V CERAMIC SMD SIZE 2220	TDK	445-13136-1-ND
10	1	C12	220uF/25V ELECTROLYTIC	PANASONIC	P10271-ND
11	3	C18,C19,C21	100nF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA
12	1	D1	LED RED SMD SIZE 0805	OSRAM	475-1278-1-ND
13	1	D2	1N4747 1W 20V ZENER	MICROCHIP	150-1N4747UR-1/TRCT-ND
14	1	L1	FERRITE BEAD 600OHM/1.5A SMD SIZE 0805	LAIRD
15	6	R1,R2,R3,R4,R6,R7	100E 5% SMD SIZE 0805	YAGEO/MURATA
16	1	R5	0.26E/2W 1% SMD SIZE 2512	YAGEO/MURATA	13-PT2512FK-7W0R26L-ND
17	1	R8	0E SMD SIZE 0805	YAGEO/MURATA
18	1	R9	2K 1% SMD SIZE 0805	YAGEO/MURATA
19	2	R10,R12	1K 1% SMD SIZE 0805	YAGEO/MURATA
20	1	R11	9.1K 1% SMD SIZE 0805	YAGEO/MURATA
21	1	R13	DNP
22	1	U1	FSB50450AS	ON SEMI	FSB50450ASCT-ND
23	1	C20	10uF/25V CERAMIC SMD SIZE 0805	YAGEO/MURATA
24	3	C13, C14,C15	1uf/25V CERAMIC SMD SIZE 0805	YAGEO/MURATA

Connections

Heatsink Installation

Microcontroller Interface

Gerber View

Photos

Video

FSB50450AS Datasheet

2.1 Channel Audio Processor with Subwoofer Output

This board is an Audio Processor built using PT2033 chip audio processor designed for versatile applications. It includes 3 stereo input selectors with internal gain and master volume control with low-frequency loudness compensation. It is capable of individual channel output volume adjustment and also tone control.

The PT2033 is 2.1 CH Audio Processor with I²C bus interface. The IC is also a sound processor that includes all of the functions required to process the audio signal for TV & Mini Compo, such as Tone control, Volume, Mute, Balance, Loudness, 3 Stereo inputs selector, 1 Stereo output, and 1 Subwoofer out are all built into a single chip to provide Audio system having the highest audio performance and reliability. If a system wants subwoofer signal output, a traditional solution needs a lot of discrete components and OPAMPs to consist of a low pass filter; the PT2033, only needs 2 external capacitors. The PT2033 combines the low pass filter inside the audio processor and the user may determine low pass corner frequency by the capacitance. A broader supply range emphasizes support to every application.

Note: Refer to Datasheet of PT2033 chip for I²C control signal information.

Features

Wide operation range (VDD from 4V to 10V)
Controlled by I²C interface
3 Stereo Inputs with gain selection.
Screw Terminals for Easy Inputs and Outputs connections
Header Connector for Power Supply Input
Input gain range, from 0dB to +11.25dB, 3.75dB per step
2 channels master volume: from 0 dB to -78.75dB, 1.25dB per step
Built In channel mixer and lowpass filter for subwoofer output
Subwoofer output volume: from 0dB to -37.5dB, 1.25dB per step
Tone control (Bass and Treble): from -14dB to +14dB, 2dB per step
Low harmonic distortion (0.002%, Vo=200mVrms)
Low noise and DC offset
On Board Power LED
4 x 3MM Mounting Holes
PCB Dimensions 65.41 x 43.18 mm

Connections

CN1: Pin 1 = Audio Input Left 1, Pin 2 = Audio Input Left 2, Pin 3 = Audio Input Left 3, Pin 4 = GND
CN2: Pin 1 = Audio Input Right 1, Pin 2 = Audio Input Right 2, Pin 3 = Audio Input Right 3, Pin 4 = GND
CN3: Pin 1 = Audio Output Left Channel, Pin 2 = GND
CN4: Pin 1 = Audio Output Right Channel, Pin 2 = GND
CN5: Pin 1 = Audio Output Subwoofer, Pin 2 = GND
CN6: Pin 1 NC, Pin 2 = SCL-I2C, Pin 3 = SDA-I2C, Pin 4 = GND
CN7: Pin 1,2 = VDD , Pin 3,4 = GND
D1: Power LED

Application Diagram

Block Diagram

Schematic

Parts List

NO	QNTY.	REF.	DESC.	MANUFACTURER	SUPPLIER	SUPPLIER PART NO
1	4	CN1,CN2	2 PIN SCREW TERMINAL PITCH 5.08MM	PHOENIX	DIGIKEY	277-1247-ND
2	3	CN3,CN4,CN5	2 PIN SCREW TERMINAL PITCH 5.08MM	PHOENIX	DIGIKEY	277-1247-ND
3	2	CN6,CN7	4 PIN MALE HEADER PITCH 2.54MM	WURTH	DIGIKEY	732-5317-ND
4	1	C1	10uF/25V CERAMIC SMD SIZE 1206	YAGEO/MURATA	DIGIKEY
5	8	C2,C4,C5,C6,C7,C8,C11,C12	1uF/25V CERAMIC SMD SIZE 1206	YAGEO/MURATA	DIGIKEY
6	3	C9,C10,C21	2.2uF/25V CERAMIC SMD SIZE 1206	YAGEO/MURATA	DIGIKEY
7	6	C13,C14,C15,C16,C25,C26	100nF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
8	2	C17,C18	8.2KPF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
9	3	C19,C20,C22	4.7uF/25V CERAMIC SMD SIZE 1206	YAGEO/MURATA	DIGIKEY
10	1	C23	33nF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
11	1	C24	68nF/50V CERAMIC SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
12	1	D1	LED SMD SIZE 0805	OSRAM	DIGIKEY	475-1278-1-ND
13	1	L1	HZ0805E601R-10	LAIRD	DIGIKEY	240-2399-1-ND
14	2	R1,R2	5.6K 5% SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
15	3	R3,R4,R5	47K 5% SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
16	1	R6	1K 5% SMD SIZE 0805	YAGEO/MURATA	DIGIKEY
17	1	U1	PT2033	PRINCETON	ALIEXPRESS
18	1	C3	100uF/16V	PANASONIC	DIGIKEY	PCE3783CT-ND

Connections

Gerber View

Photos

PT2033 Datasheet

2 Channel MOSFET Module

This is a low profile and small size two-channel MOSFET board that can drive loads up to 2 x 3A. The Power Supply is 5V to 14V DC. Inputs and input Enable are TTL-level signals. The Output is by default enabled and an active high signal disables the output. Parallel Operation of Dual Outputs for Larger Driver Output Current is possible. The Circuit can drive a high-speed PWM signal as well as ON/OFF signal.

It is recommended to use an electrolytic capacitor near supply inputs if the full load current is used.

Features

+5V to +14V Single Power-Supply Range
Load up to 3A, each Channel
IN-A, IN-B, ENB-A, ENB-B, TTL Level Inputs
Dual Drivers with Enable Inputs
Low 12ns Propagation Delay
6ns Typical Rise and 5ns Typical Fall Times with 1nF Load
Matched Delays Between Channels
Parallel Operation of Dual Outputs for Larger Driver Output Current
TTL or HNM Logic-Level Inputs with Hysteresis for Noise Immunity
Low Input Capacitance: 10pF (typ)
Thermal Shutdown Protection
PCB Dimensions 23.50 x 18.73 mm

The MAX17600–MAX17605 devices are high-speed MOSFET drivers capable of sinking /sourcing 4A peak currents. The devices have various inverting and noninverting part options that provide greater flexibility in controlling the MOSFET. The devices have internal logic circuitry that prevents shoot-through during output-state changes. The logic inputs are protected against voltage spikes up to +14V, regardless of VDD voltage. Propagation delay time is minimized and matched between the dual channels. The devices have very fast switching time, combined with short propagation delays (12ns typ), making them ideal for high-frequency circuits. The devices operate from a +4V to +14V single power supply and typically consume 1mA of supply current. The MAX17600/MAX17601 have standard TTL input logic levels, while the MAX17603 /MAX17604/MAX17605 have CMOS-like high-noise margin (HNM) input logic levels. The MAX17600/MAX17603 are dual inverting input drivers, the MAX17601/MAX17604 are dual noninverting input drivers, and the MAX17602/MAX17605 devices have one noninverting and one inverting input. These devices are provided with enable pins (ENA, ENB) for better control of driver operation.

Connections

CN1: Pin 1 = VDD, Pin 2 = GND
CN2: Pin 1 = VDD, Pin 2 = Output A
CN3: Pin 1 = Enable B, Pin 2 Enable A, Pin 3 = Input A, Pin 4 = Input B, Input = GND
CN4: Pin 1 = VDD, Pin 2 = Output B
D1: Power LED

Typical Schematic

Schematic

Parts List

NO	QNTY.	REF.	DESC	MANUFACTURER	SUPPLIER	SUPPLIER PART NO
1	3	CN1,CN2,CN4	2 PIN MALE HEADER PITCH 2.54MM	WURTH	DIGIKEY	732-5315-ND
2	1	CN3	5 PIN MALE HEADER PITCH 2.54MM	WURTH	DIGIKEY	732-5318-ND
3	4	C1,C2,C3,C4	22uF/25V CERAMIC SMD SIZE 0805	WURTH	DIGIKEY
4	1	C5	100nF/50V CERAMIC SMD SIZE 0805	WURTH	DIGIKEY
5	1	D1	LED SMD RED SIZE 0805	OSRAM	DIGIKEY	475-1278-1-ND
6	1	R1	1.5K 5% SMD SIZE 0805	WURTH	DIGIKEY
7	2	U1,U3	IRF7862 MOSFET SOIC8	INFINEON	DIGIKEY	IRF7862TRPBFCT-ND
7	1	U2	MAX17600ASA SOIC8	ANALOG DEVICE	DIGIKEY	MAX17600ASA+-ND

Connections

Gerber View

Photos

Video

MAX17600 Datasheet

SAKURA-IIEdgeCortix Launches SAKURA-II Platform to Power the Next Wave of Generative AI at the Edge

Posted on 23 May, 202423 May, 2024 by mixos

The next-generation high performance, energy efficient Edge AI accelerator addresses the latest Generative AI solutions at the edge from vision to billions of parameters large language models

EdgeCortix® Inc., a leading fabless semiconductor company specializing in energy-efficient AI processing at the edge, today unveiled its next-generation SAKURA-II Edge AI accelerator.

This state-of-the-art platform, paired with EdgeCortix’s innovative second generation Dynamic Neural Accelerator (DNA) architecture, is engineered to tackle the most challenging Generative AI tasks in the industry. Designed for flexibility and power efficiency, SAKURA-II empowers users to seamlessly manage a wide range of complex tasks including Large Language Models (LLMs), Large Vision Models (LVMs), and multi-modal transformer-based applications, even within the stringent environmental constraints at the edge. Featuring low latency, best-in-class memory bandwidth, high accuracy, and compact form factors, SAKURA-II delivers unparalleled performance and cost-efficiency across the diverse spectrum of edge AI applications.

Well-suited for numerous use cases across the manufacturing, industry 4.0, security, robotics, aerospace, and telecommunications industries, SAKURA-II features EdgeCortix’s latest generation runtime reconfigurable neural processing engine, DNA-II. Leveraging this highly configurable intellectual property block, SAKURA-II delivers power efficiency and real-time processing capabilities while simultaneously executing multiple deep neural network models with low latency. SAKURA-II can deliver up to 60 trillion operations per second (TOPS) of effective 8-bit integer performance and 30 trillion 16-bit brain floating-point operations per second (TFLOPS), while also supporting built-in mixed precision for handling the rigorous demands of next-generation AI tasks.

The SAKURA-II platform, with its sophisticated MERA software suite, features a heterogeneous compiler platform, advanced quantization, and model calibration capabilities. This software suite includes native support for leading development frameworks such as PyTorch, TensorFlow Lite, and ONNX. MERA’s flexible host-to-accelerator unified runtime is adept at scaling across single, multi-chip, and multi-card systems at the edge, significantly streamlining AI inferencing and shortening deployment times for data scientists. Furthermore, the integration with the MERA Model Library, with seamless interface to Hugging Face Optimum, offers users access to an extensive range of the latest transformer models, ensuring a smooth transition from training to edge inference.

“SAKURA-II’s impressive 60 TOPS performance within 8W of typical power consumption, combined with its mixed-precision and built-in memory compression capabilities, positions it as a pivotal technology for the latest Generative AI solutions at the edge,” said Sakyasingha Dasgupta, CEO and Founder of EdgeCortix. “Whether running traditional AI models or the latest Llama 2/3, Stable-diffusion, Whisper or Vision-transformer models, SAKURA-II provides deployment flexibility at superior performance per watt and cost-efficiency. We are committed to ensuring we meet our customer’s varied needs and also to securing a technological foundation that remains robust and adaptable within the swiftly evolving AI sector.”

Key Benefits of SAKURA-II include:

Optimized for Generative AI: Tailored specifically for processing Generative AI workloads at the edge with minimal power consumption.
Complex Model Handling: Capable of managing multi-billion parameter models like Llama 2, Stable Diffusion, DETR, and ViT within a typical power envelope of 8W.
Seamless Software Integration: Fully compatible with EdgeCortix’s MERA software suite, facilitating seamless transitions from model training to deployment.
Enhanced Memory Bandwidth: Offers up to four times more DRAM bandwidth than competing AI accelerators, ensuring superior performance for LLM and LVM.
Real-Time Data Streaming: Optimized for low-latency operations under real-time data streaming conditions.
Advanced Precision: Provides software-enabled mixed-precision support for near FP32 accuracy.
Sparse Computation: Supports sparse computation to reduce memory footprint and optimize bandwidth.
Versatile Functionality: Supports arbitrary activation functions with hardware approximation for enhanced adaptability.
Efficient Data Handling: Includes a dedicated Reshaper engine to manage complex data permutations on-chip and minimize host CPU load.
Power Management: Features on-chip power-gating and power management capabilities to facilitate ultra-high efficiency modes.

SAKURA-II will be offered as a stand-alone device, two different M.2 modules with varying DRAM capacity, single and dual-device low-profile PCIe cards. Customers can reserve M.2 modules and PCIe cards today for delivery in the second half of 2024.

Reserve SAKURA-II accelerators today by registering here.

TerraMaster Unveils Its Highest-Performing 2-Bay NAS F2-424, Powered by Intel’s Latest Quad-Core Processor for Ultimate Performance

Posted on 23 May, 20243 July, 2024 by mixos

TerraMaster, a professional brand focused on providing innovative storage products for homes and businesses, recently introduced the F2-424, a new upgrade of F2-423. The brand-new F2-424 adopts the Intel Celeron N95 4-core 4-thread processor, 8GB DDR5 memory, two 2.5GB network ports, two M.2 NVMe sockets, and adds a USB Type-C host interface and a USB Type-A interface, achieving a 40% improvement in overall performance compared to the previous generation.

The F2-424 will come with the new generation TOS 5.1 operating system and can be smoothly upgraded to the TerraMaster TOS 6 operating system in the nearly future. With its robust features and state-of-the-art technology, the F2-424 NAS redefines data storage and management for both personal and business users.

F2-424 Key Features

Faster Application Response

Featuring a more powerful CPU and DDR5 memory, the F2-424 represents a performance improvement of more than 140% on the previous generation. Application load times have increased by 90%, file and photo retrieval speed by 40%, database response speed by 50%, and the PHP response for web pages has seen a remarkable 60% improvement. Such enhanced performance enables seamless handling of high-load workflows.

Higher Transmission Speed

Configured with two 2.5 GbE interfaces, F2-424 supports a 2.5 GbE high-speed network bandwidth. The linear data transmission speed can reach 283 MB/s. Link Aggregation offers a network bandwidth up to 5 Gb, thereby providing a cost-effective solution for multi-user and high-concurrent file access.

Latest TOS System and Virtualized Applications

F2-424 runs on the latest TOS 5.1 and the coming TOS 6 system. The F2-424 satisfies virtualized application needs and provides additional functionality within one device with the use of professional virtualized apps, including VirtualBox and Docker, and by cooperating with Docker-compose and Portainer. This lowers the cost of corporate investment in IT.

Rich Backup Solutions

Centralized Backup, Duple Backup, Snapshot, CloudSync, TFM Backup, and more enhance the security of your data with multiple backup applications. And the F2-424 supports various RAID types, including RAID 0/1/JBOD/Single, and TRAID.

Powerful Multimedia Service

With 4K video decoding capability, the F2-424 is compatible with uPnP/LNA protocol and can stream videos to various multimedia devices, including computers, smart phones, media players and televisions, by using TerraMasters proprietary application Multimedia Server or the one from another third-party, to deliver users constant, reliable entertainment experiences.

Easy M.2 SSD Installation

To facilitate the installation of M.2 SSDs, TerraMaster has specially designed a side sliding cover for the F2-424 chassis, which is simply pushed aside to install M.2 SSDs. TerraMaster also provides hand-tightened screws, making it easy to install an SSD within 5 seconds.

Less Noise

The F2-424 adopts TerraMaster new structure achieved a noise level in standby mode is only 19dB(A), which is 50% lower than the previous generations.

In addition to the F2-424, the 424 series includes another 2 models: F4-424 (4-bay) and F4-424 Pro (4-bay), which are available worldwide catering to the demands of home users, SOHO, and small businesses for high-performance NAS.

For more details, please visit

TerraMaster F2-424: https://www.terra-master.com/global/products/homesoho-nas/f2-690.html

Amazon Links

Raspberry Pi Introduces M.2 HAT+ for High-Speed Peripheral Connectivity

Posted on 21 May, 202421 May, 2024 by Aditi Ambadkar

The Raspberry Pi Foundation has unveiled the M.2 HAT+, an innovative accessory designed to enhance the capabilities of the Raspberry Pi 5. This new add-on allows users to connect M.2 peripherals, such as NVMe drives and AI accelerators, to the Raspberry Pi 5’s PCIe 2.0 interface, enabling data transfer speeds of up to 500 MB/s.

Key Features and Design

The M.2 HAT+ acts as a mechanical adapter, converting the Raspberry Pi 5’s 16-pin, 0.5mm-pitch FPC connector to an M.2 M key edge connector. It supports devices in the 2230 and 2242 form factors and can supply up to 3 A of power to connected peripherals. This ensures compatibility with a wide range of high-performance components, making it an essential tool for both hobbyists and professionals.

Conforming to the Raspberry Pi HAT+ specification, the M.2 HAT+ is designed for seamless integration with the Raspberry Pi 5. The device is autodetected by the latest Raspberry Pi software and firmware, simplifying the setup process. It also features power and activity LEDs, providing clear status indicators during operation.

The launch of the M.2 HAT+ is the result of extensive testing and refinement. The Raspberry Pi engineering team worked diligently to ensure compatibility with various NVMe drives and other peripherals. During the development process, they addressed numerous issues, such as startup timing problems in drive firmware, to ensure reliable performance with the BCM2712 PCIe controller on the Raspberry Pi 5.

The M.2 HAT+ comes with a comprehensive set of accessories, including a ribbon cable, 16 mm stacking header, threaded spacers, screws, and a knurled, double-flanged screw. These components facilitate easy installation and ensure that the M.2 HAT+ can be securely attached to the Raspberry Pi 5, even with the Raspberry Pi Active Cooler in place.

Key Features and Specifications

Single-lane PCIe 2.0 interface with 500 MB/s peak transfer rate
Supplies up to 3A to connected M.2 devices
Includes power and activity LEDs
Conforms to Raspberry Pi HAT+ specification
Comes with ribbon cable, 16mm GPIO stacking header, threaded spacers, screws, and a knurled double-flanged drive attachment screw
M.2 M-key socket for 2230 or 2242 modules (2280 modules are not supported)

It wouldn’t be a Raspberry Pi product announcement without their usual commitment to long-term support, maintaining the M.2 HAT+ in production until at least January 2032, and providing users with access to the accessory into the next decade. Additionally, the Foundation has published detailed schematics and documentation to support further innovation and development within the Raspberry Pi community.

The release of the M.2 HAT+ not only enhances the capabilities of the Raspberry Pi 5 but also fosters a broader ecosystem of high-quality PCIe accessories.

GIGAIPC PICO-N97A: Pico-ITX SBC with Intel Processor N97

Posted on 21 May, 2024 by Aditi Ambadkar

GIGAIPC PICO-N97A Pico-ITX SBC features an Intel Processor N97 quad-core Alder Lake-N processor coupled with up to 16GB DDR5 SO-DIMM memory and M.2 SATA or NVMe storage designed for passively cooled and enclosed systems for Industry 4.0 applications in smart cities, retail, and healthcare sectors.

The single board computer supports up to two independent displays via HDMI and LVDS interfaces. It also provides dual Gigabit Ethernet, two USB 3.1 ports, an additional M.2 Key-B socket for wireless, and various headers for RS232/RS422/RS485, GPIO, USB 2.0, and more.

GIGAIPC PICO-N97A specifications:

SoC – Intel Processor N97
CPU – Alder Lake-N quad-core/quad-thread processor @ up to 3.6 GHz
Cache – 6MB
GPU – Intel UHD graphics with 24 execution units @ up to 1.20 GHz
TDP – 12W
System Memory – Supports up to 16GB of DDR5-4800 via a single SO-DIMM slot
Storage – Compatible with SATA or NVMe SSDs through M.2 M-Key socket (See expansion)
Video Output
- HDMI 2.0 port supports up to 4096×2160 @ 60Hz
- LVDS connector supports up to 1920×1200 @ 60Hz and includes a backlight control header
- Supports up to 2x independent displays
Networking – Dual Gigabit Ethernet ports using RealTek controllers
USB – 2x USB 3.2 Gen1 Type-A ports, plus 2x USB 2.0 via headers
Serial – RS232/422/485 header for COM port
Expansion
- M.2 2230 E-Key socket (PCIe x1, USB 2.0)
- M.2 2280 M-Key socket (PCIe x2, SATA 6Gb/s)
- 4x GPIO pins
Security – Infineon SLB9670VQ2.0 TPM 2.0 chip
Misc – Front panel header
Power Supply – 12V DC via 2-pin box connector
Dimensions – 100 x 72mm (Pico-ITX form factor)
Temperature Range –
- Operating: 0°C to 60°C
- Storage: -40°C to 85°C
Humidity – 0-90% (non-condensing)

GIGAIPC did not provide any information about OS support, but it will be Windows and/or Linux depending on driver support. This is not the first Intel N97 Pico-ITX SBC covered, as the AAEON PICO-ADN4 is offered with Intel N50, N97, Core i3-N305, or Atom x7425E Alder Lake-N processors.

Availability and pricing information were not disclosed, but for reference, the AAEON sells samples of the PICO-ADN4 (N97) for $309. A few more details about the PICO-N97A may be found on the product page and press release.

BitNetMCU Slims Down TinyML Models

Posted on 21 May, 202421 May, 2024 by Aditi Ambadkar

The BitNetMCU framework simplifies the process of training accurate neural networks for deployment on low-end microcontrollers, such as the 15-cent CH32V003.

While major releases of new generative artificial intelligence (AI) tools, like large language models and text-to-image generators, tend to grab all the headlines, smaller-scale AI tools are quietly transforming a number of industries. These powerful yet compact AI applications, known as TinyML, are already being utilized to keep manufacturing equipment in optimal condition, monitor the environment, and oversee people’s health.

New TinyML applications are being continually enabled by advancements in hardware that shrink their size, cost, and energy consumption. Furthermore, algorithmic improvements allow more powerful models to run with fewer resources. Most models are trained primarily with accuracy in mind, leading to strategies focused on reducing model size without significant drops in accuracy.

BitNetMCU introduces an approach to facilitate the training and optimization of neural networks for basic microcontrollers.

A new framework called BitNetMCU seeks to streamline the process of training and shrinking neural networks so they can run on even the lowest-end microcontrollers. The initial focus of the project is to support low-end RISC-V microcontrollers like the CH32V003, which has only 2 kilobytes of RAM and 16 kilobytes of flash memory. Moreover, this chip lacks a multiply instruction in its instruction set, which is crucial for the matrix multiplications that are key components of neural network computations. However, the CH32V003 is extremely cost-effective, priced at about 15 cents in small quantities, making it an attractive option for large, distributed networks of sensors.

The BitNetMCU pipeline consists of a series of Python scripts that use PyTorch on the backend to train a neural network, followed by low-bit quantization to reduce the model size so it can fit on a tiny microcontroller. After training, additional scripts are available for testing the model’s performance. Once everything is optimized, another utility assists in deploying the model to a CH32V003 microcontroller. Although initially focused on this specific microcontroller, BitNetMCU‘s inference engine is implemented in ANSI C, ensuring support across a wide range of hardware platforms.

The BitNetMCU system utilizes Python scripts and PyTorch for training neural networks, which are then compressed through low-bit quantization to fit on small microcontrollers.

To test the system, an experiment was conducted to see how well a handwritten digit recognition model would perform against the 16×16 MNIST dataset. Due to the heavy quantization applied before deployment, the model was able to fit within the limited RAM and flash memory of the CH32V003. The chip’s lack of a multiply instruction was mitigated by using multiple additions, a workaround facilitated by the BitNetMCU framework. Testing showed that the model achieved an impressive accuracy level of better than 99 percent.

This high level of performance demonstrates the potential of BitNetMCU to make advanced AI accessible even on the most constrained hardware. The framework not only enables the deployment of sophisticated models on low-cost microcontrollers but also ensures that these models run efficiently and accurately.

The broader implications of BitNetMCU are significant. By making it easier to deploy powerful AI models on inexpensive and widely available microcontrollers, BitNetMCU opens up new possibilities for embedding intelligence in a multitude of devices. This can lead to smarter, more responsive technology in everything from industrial sensors to consumer electronics, enhancing functionality while keeping costs low.

BitNetMCU’s approach of leveraging existing AI frameworks like PyTorch and focusing on quantization techniques aligns well with the needs of developers working with constrained devices. The availability of comprehensive scripts and tools for training, testing, and deployment simplifies the workflow, allowing developers to focus on innovation rather than the intricacies of hardware limitations.

As the framework continues to evolve, it is expected to support an even wider range of microcontrollers and AI models, further democratizing the use of TinyML. With BitNetMCU, the promise of bringing AI to every corner of our world moves closer to reality, enabling smarter and more efficient technology across diverse applications.

Source code for the project is available on GitHub, and there is some fairly good documentation as well. So, if you have a few dimes burning a hole in your pocket, BitNetMCU could be a fun way to get started in the world of tinyML.

SparkFun M7E Hecto RFID Reader – A Simultaneous RFID Reader with USB-C Interface

Posted on 21 May, 2024 by Debashis Das

The SparkFun M7E Hecto RFID Reader is a high-performance RFID reader for reading UHF RFID tags. The Tag is built around a powerful M7E-HECTO module by JADAK and can read up to 300 tags per second and write at 80 milliseconds.

The module’s power is also adjustable, so it offers a 1-2 foot range with the onboard antenna, there’s also an option for a solderable external antenna through which the reader can read tags from 16 feet or 4.9 meters away. The reader gets connected with the PC or any other SBCs with a USB-C port and can communicate via a built-in CH340C converter for easy, solder-free setup.

The reader also offers flexibility in power options. You can choose between USB-C or a 0.1in.-header for connecting to an external 3.3V source or microcontroller of your choice. A convenient switch on the board lets you effortlessly switch between these interfaces.

SparkFun M7E Hecto RFID Reader Specifications

Main Chip – M7E-HECTO module by JADAK
Read Rate: Up to 300 tags/sec to read 96-bit EPC format
Write Rate: 80ms for standard write of 96-bit EPC format
Connectivity and Interface:
- Serial interfaces:
  - USB-C connector
  - 0.1″-spaced PTH header (3.3V logic)
- 2-way switch for serial interface selection
- Enable and GPIO PTH pins
Antenna Options:
- Integrated PCB trace antenna (default)
- u.FL connector for external antenna connection
Performance:
- Adjustable read and write power levels: 0dBm to 27dBm in 0.01 dB steps
- Read capability: Up to 300 tags/sec for 96-bit EPC
- Write speed: Typical 80ms for standard 96-bit EPC
Power Specifications:
- Supply voltage: 3.3V-5V
- Maximum supply current: 1A
- Power mode consumption (@5V):
  - Full: 0.665W
  - Minsave: 0.140W
  - Sleep: 0.080W
- The internal current limiting circuit
Environmental:
- Operating temperature range: -40°C to 60°C
- Built-in thermal management
Standards and Protocols:
- EPCglobal Gen 2 (ISO 18000-6C) compliant
- Nominal backscatter rate: 250kbps
Dimensions: 2.4 x 1.4 inch

External Antenna Connected to M7E Hecto Reader

The M7E Hecto has a lot of software support, including Jadak’s Universal Reader Assistant for Windows users and SparkFun’s Arduino library for seamless integration with microcontrollers. This Arduino library simplifies serial communication, byte manipulations, and CRC verifications, making it easy for users to get started.

The M7E Hecto is fully open-source, meaning all the hardware schematics, production files, and documentation are available on GitHub. This open-source nature encourages customization and community collaboration, making the M7E Hecto a versatile tool for a wide range of RFID projects. The module can be purchased from the SparkFuns website for $299.95.

NANO-6064-ASL: Nano-ITX Industrial Embedded Board

Posted on 21 May, 202421 May, 2024 by Saumitra Jagdale

The industrial setting often calls for advanced SoCs and Dev boards that can operate at a wide range of temperatures. In this context, NANO-6064-ASL provides a processing platform that not only can handle industrial temperatures ranging from -40°C to 85°C but also performs at minimal power consumption. Featuring the most recent Intel Atom® x7000RE series processors, NANO-6064-ASL is well-suited for developing efficient IoT solutions in medical, retail, and digital signage. Intel Atom x7000RE Series CPUs are an advanced choice for industrial, communication, and edge computing applications. These processors can perform more tasks at once with less power consumption with up to 8 Gracemont cores, which helps them run faster. They also have better abilities for both the main computing part (CPU) and the graphics part (GPU).

Features

Intel Atom® x7000RE series SoC
DDR4 3200 MT/s non-ECC SO-DIMM up to 16GB
Triple displays by DP, HDMI and LVDS
High-speed I/O via dual 2.5GbE and four USB 3.2 Gen2
Multiple expansions with M.2 E Key and M.2 B Key
Support real-time performance: Intel® TSN, TCC (selected SKUs)
Support onboard TPM 2.0
Support a wide -40°C to 85°C industrial temperature range

NANO-6064-ASL Specifications:

The NANO-6064-ASL integrates an Intel Atomx7000RE Series Processor, delivering efficient performance within a power range of up to 12W. Managed by an AMI UEFI BIOS, it supports a single non-ECC DDR4 SO-DIMM slot, expandable up to 16GB. It also offers In-Band ECC support. Storage options include SATA III, M.2 B Key 2280 (SATA III), and a micro SD socket. Hardware features comprise a programmable watchdog timer and system monitoring capabilities for voltage, fan speed, and temperature. The interface is expandable thanks to M.2 E Key 2230 and M.2 B Key 3052+2280 (shared with M.2 2280 SATAIII socket).

The I/O interface features an HDA controller integrated into the Intel SoC, Realtek ALC897 HDA codec which supports 7.1 channel high-definition audio playback and recording, dual Intel 2.5GbE Controllers (I226-IT), and a single RS-232/422/485 port. USB connectivity includes two USB 3.2 Gen2 ports on the rear I/O and two USB 3.2 Gen1 ports accessible via onboard pin headers. Notably, one USB 3.2 Gen1 port shares connectivity with the M.2 B key socket. Additional features encompass onboard TPM 2.0, 8x GPIO, and support for Real-time Performance features on selected SKUs. On the other hand, the Intel UHD Graphics Gen12 controller, including LVDS, HDMI, and DP ports facilitates display. Mechanically, it measures 120mm(L) x 120mm(W) and operates with a DC 12V input. What’s more, is that the board doesn’t get affected by changes in humidity.

NANO-6064-ASL Development Board

NANO-6064-ASL Architecture and Features:

With up to 16GB DDR4:3200MT/s non-ECC SO-DIMM support, it ensures smooth multitasking and efficient operation. Users can also enjoy versatile display options via HDMI, DP, and LVDS connections, facilitating multiple display setups. High-speed I/O capabilities are provided through dual 2.5GbE LAN ports and four USB 3.2 ports, enhancing connectivity and data transfer. Expansion options are available with M.2 E Key and B Key slots, accommodating various add-on modules. Additionally, onboard TPM 2.0 support enhances security measures.

The NANO-6064-ASL is a new advanced industrial and IoT embedded board featured by Intel Atom®x7000RE Series processors with exceptional features like supporting M.2, microSD, SATA III, TPM, triple display support including HDMI, DisplayPort, and LVDS. It is specifically designed for industrial use applications, featuring wider temperature ranges, dependable performance, and real-time capabilities via Intel TCC and TSN.