Page 124 – Electronics-Lab.com

Increasing the Computing Power with AI Smart Module for AIoT Applications

Posted on 31 October, 2022 by Saumitra Jagdale

With the increased amounts of data generated by IoT applications, the scope of integrating AI technologies tends to create some excellent opportunities for handling smart ecosystems. As the computing power of an IoT system decides the compatibility of implementing AI functionalities, it becomes significant to include AI-specific hardware in designing these AIoT (AI + IoT) applications. Fibocom launches a high-performance AI smart module SCA825-W offering a computing power of up to 15 TOPS dedicated to AIoT applications requiring extensive processing.

Qualcomm’s QCS8250 SoC is the heart of SCA825-W AI smart module, which makes it suitable for applications with real-time processing requirements, such as video conferencing and live streaming, cloud gaming, edge computing, robotics, drone imaging and various use cases of AR/VR. Additionally, it could be a great fit for 5G use cases due to its capability to accommodate applications with quicker data transmission and reception.

Technical Specifications of Fibocom’s SCA825-W AI Smart Module

SoC: Qualcomm’s flagship IoT solution QCS8250 SoC
CPU: Octa-core [email protected]+ [email protected]+ [email protected]
GPU: Adreno™ 650 GPU
Triple Display: 4K @ 60 fps Miracast DisplayPort over Type-C
Memory: 8GB LPDDR5+128GB UFS
Operating System: Android 10

“With AIoT continuously transforming every industry, high-performance AI modules will become an important pillar” as highlighted by Eden Chen, General Manager of MC Product Management at Fibocom. “The newly-launched AI smart module SCA825-W is an exceptional part of the Fibocom smart module family, which will take a big step forward in the AIoT field with technological advancements.”

SCA825-W AI smart module comes with dynamic image processing and deep processing units for implementing AI algorithms. Furthermore, it supports up to seven cameras, triple 4K display and video encoding at up to 4K resolution at 120 fps, 8K at 30 fps (frames per second) dedicated to extensive image processing and displaying use cases.

The module has flexible wireless connectivity as it supports WiFi and Bluetooth technologies for the communication of data generated or required for AIoT applications. Talking specifically about the connectivity, SCA825-W accommodates 5G, Wi-Fi 6.0, Bluetooth 5.1 as well as 2×2 Wi-Fi MIMO multi-antenna technology. Additionally, it integrates a wide range of peripheral interfaces such as MIPI-DSI, I2S, PCIe, UART, USB, I2C, SPI, which increases the compatibility of the module for machine-to-machine applications.

You can read more about the SCA825-W AI smart module on the official product page of Fibocom.

Vecow EAC-5000 Features NVIDIA Jetson AGX Orin SoM, Offering Server-Class Performance for Edge AI Applications

Posted on 28 October, 202228 October, 2022 by Emmanuel Odunlade

Vecow has launched a feature-rich edge AI computing system that guarantees power-efficient AI productivity, flexible integration, and mobile availability: the Vecow EAC-5000.

Designed to deliver faster and/or more energy-efficient performance for machine vision and automotive applications, the Vecow EAC-5000 embedded computer is powered by NVIDIA’s Jetson AGX Orin system-on-module, the latest and most powerful module in the Jetson family, capable of delivering 275 TOPS of AI performance (more than 8 times the AI performance of the earlier NVIDIA Jetson AGX Xavier module). The NVIDIA Jetson AGX Orin module also boasts of an Ampere architecture GPU with state-of-the-art graphics and parallel computing techniques, a 12-core Arm Cortex-A78AE CPU running at up to 2.2 GHz, and the second generation of NVIDIA DLA (Deep Learning Accelerators) with 9x times the performance of the previous generation. With other features like a multi-standard encoder and decoder, JPEG Codec, and the second generation of a Programmable Vision Accelerator supporting common computer vision kernels, the module no doubt is on a higher level when it comes to the performance achievable by embedded system-on-module.

Vecow EAC-5000 is also characterized by various wireless connectivity options such as WiFi, 5G, 4G LTE, 6x antennas, support for up to 8x GMLS2 cameras, 1x PCIe x8 slot, 2 GbE LAN, 5x USB, 2x SIM card sockets, 1x software ignition control, a wide range input DC of 9V to 50V, and temperature range of -20 to 70°C. It can be wall mounted with a mounting bracket, and the company offers an optional DIN Rail mount too.

The EAC-5000 system comes in two variants, depending on the version of the module it comes with. The specifications highlighted below give in-depth details about that.

Specifications:

SoM: NVIDIA Jetson AGX Orin 32GB/64GB
- CPU:
  - EAC-5000-R32: 64-bit 8-core Arm Cortex A78AE v8.2 processor @ 2.2 GHz with 2MB L2, 4MB L3 cache
  - EAC-5000-R64: 64-bit 12-core Arm Cortex-A78AE v8.2 processor @ 2.2 GHz with 3MB L2, 6MB L6 cache
- GPU:
  - EAC-5000-R32: NVIDIA Ampere architecture with 1792 NVIDIA CUDA cores and 56 Tensor Cores @ 1GHz
  - EAC-5000-R64: NVIDIA Ampere architecture with 2048 NVIDIA CUDA cores and 64 Tensor cores @ 1GHz
- System Memory:
  - EAC-5000-R32: 32GB 256-bit LPDDR5 @ 204.8 GB/s
  - EAC-5000-R64: 64GB 256-bit LPDDR5 @ 204.8 GB/s
- DL Accelerators:
  - EAC-5000-R32: 2x NVIDIA v2.0 @ up to 1.4 GHz
  - EAC-5000-R64: 2x NVIDIA v2.0 @ up to 1.6 GHz
- Vision Accelerator: Programmable Vision Accelerator (PVA) v.2.0
- AI Performance:
  - EAC-5000-R32: 200 TOPS @ 50W / 100 TOPS
  - EAC-5000-R64: 275 TOPS @ 50W / 100 TOPS
- Video Encode:
  - EAC-5000-R32: 1x 4K@60; 3x 4K @30; 6x 1080p @60; 12x 1080p @30
  - EAC-5000-R64: 2x 4K@60; 4x 4K @30; 8x 1080p @60; 16x 1080p@30
- Video Decode:
  - EAC-5000-R32: 1x 8K @30; 2x 4K @60; 4x 4K @30; 9x 1080p @60; 18x 1080p @30
  - EAC-5000-R64: 1x 8K @30; 3x 4K @60; 7x 4K @30; 11x 1080p@60; 22x 1080p @30
2x M.2 Key M 2280 socket
MicroSD card socket
1x 64GB eMMC 5.1
1x Digital Display interface up to 8K60
8x Fakra-Z connectors for GMSL2 automotive cameras
2x GbE ports
2x SIM card socket
Optional WiFi, 5G, 4G LTE connectivity through M.2 sockets
6x antennas for WiFi, 5G, 4G LTE, UMTS, GPRS
1x USB 3.1 Gen 2
4x USB 3.1
2x COM RS-232/422/485
2x isolated CAN BUS
1x PCIe Gen4 x8 slot
1x M.2 Key B socket
1x M.2 Key E socket
8x DI, 8x DO
1x Micro USB console debug port
1x Micro USB OS flash port
Power and Rest Buttons
2x User-programmable LEDs
16-mode Software Ignition control
3-pin Remote Switch terminal block
Remote power on/off and remote reset control
Power Input: DC 9V to 50V via 3-pin terminal block
Dimension: 260 mm x 182 mm x 69 mm
Weight: 3.8 kg
Temperature:
- -20°C to 70°C (30W power mode)
- -20°C to 55°C (Maximum power mode)
- -40°C to 85°C (Storage)
Humidity: 5% to 95%, non-condensing
Operating System: Linux Ubuntu 20.04 OS with JetPack SDK
Certifications: CE, FCC, EN50155, EN50121-3-2

The rugged AI computing system is ideal for advanced edge AI applications such as robotic control, intelligent video analytics, in-vehicle computing, machine vision, mobile robots, etc.

Other useful details can be found on the product page, but no details on pricing and availability yet.

Seeed Studio sells A203 Mini PC with Jetson Xavier NX for $858.00

Posted on 28 October, 202228 October, 2022 by Abhishek Jadhav

Seeed Studio has launched a Jetson Xavier NX-based A203 Mini PC designed for artificial intelligence applications deployed at the edge. Given its smaller form factor than the Nvidia Jetson NX Developer Kit, the hardware platform is capable of delivering the same performance of up to 21 tera operations per second. For those enterprises looking for miniature hardware that can support applications such as smart cities, security, industrial automation, and smart factories, the A203 Mini PC is set to be an ideal option in the market.

Manufactured in a 100x44x59 mm sized edge AI box equips the Nvidia Jetson Xavier NX modules integrated with 8GB of memory and 128GB of NVMe SSD storage. For its deployment in industrial applications, the manufacturer has provided a rich set of input/output ports, including RJ45, RS232, I2C, HDMI, and USB. Additionally, the A202 Mini PC can operate efficiently in extreme temperatures of 20°C to 80°C for industrial and functional safety applications.

When it comes to AI capabilities, the Nvidia Jetson Xavier NX module has been one of the famous hardware platforms in the market. It isn’t surprising to see more such edge hardware boxes equipped with Nvidia AI modules. The advanced embedded AI system comes with 384 Nvidia CUDA cores, 48 Tensor cores, 6 Carmel Arm CPUs, and two Nvidia Deep Learning Accelerator engines.

In the age of machine vision AI applications, the A203 Mini PC is combined with over 59.7GB/s of memory bandwidth, which is required for video encoding and decoding capabilities. Given its ability to deliver up to 21 TOPS, it makes it ideal for high-performance computing and AI in embedded and edge systems. The module is capable of running modern neural networks in parallel and also processing high-resolution data from multiple external sensors simultaneously.

Seeed Studio A203 Mini PC supports AlwaysAI accelerated edge deployment, which means that the customer can train a new model in hours, choose from 130+ pre-trained models, and also deploy it to the mini PC. AlwaysAI is known to reduce entry barriers for building machine learning applications and creating computer vision models fast and effectively.

For more information on the $858.00 USD A203 Mini PC edge computer, head to the official product page.

U-Blox launches development kits with centimeter-level positioning accuracy for evaluation and prototyping

Posted on 28 October, 202228 October, 2022 by Abhishek Jadhav

U-Blox, a Swiss embedded electronic device manufacturer, has announced the launch of explorer kits for easy prototyping and faster time to market. The explorer kit is designed to evaluate applications that require centimeter-level positioning capabilities. The XPLR-HPG-1 and XPLR-HPG-2, high-precision GNSS positioning devices feature real-time kinematics, dead reckoning, cellular, Wi-Fi, and Bluetooth communication.

Both the explorer kits are designed to be easily integrated with the company’s complementary services, such as PointPerfect GNSS augmentation service, and the ubxlib software component. The PointPerfect GNSS augmentation service gives 3 to 6-centimeter accuracy and convergence within seconds. Low-bandwidth data reduces transmission costs and offers 99.9 percent uptime availability through the internet and L-band satellites. The ubxlib simplifies developer experience for rapid application development. This is a portable, open-source C library that is available on GitHub and offers seamless integration of product features and cloud services.

“Centimeter-level positioning accuracy is becoming increasingly important for both consumer and industrial applications. It relies on getting GNSS technology working hand-in-hand with wireless communication capabilities to link in with augmentation data services such as PointPerfect.

The XPLR-HPG-1 explorer kit is based around the U-Blox NORA-W106 wireless microcontroller that offers rich wireless connectivity through Wi-Fi and Bluetooth LE capabilities. As part of the explorer kit, the manufacturer has provided users with three MIKROE click boards– the ZED-F9R high-precision RTK GNSS module, the LARA-R6001D LTE Cat 1 module, and the NEO-D9S L-band correction data receiver module. Meanwhile, the XPLR-HPG-2 explorer kit incorporates the NINA-W106 microcontroller with the same wireless connectivity as the other module.

As mentioned earlier, the explorer kits can be used for centimeter-level positioning because the GNSS receiver provides an initial position reading, which is then refined using correction data from the PointPerfect service. U-Blox has aimed to make the process as simple as possible for design engineers, providing them with all the required technologies and services in a unified solution.

“Our aim with these new kits is to make the process as simple as possible for design engineers, by providing all the key technologies and services in one place. This will enable them to create rapid proofs of concept of their solutions, and ultimately get the finished product into customers’ hands more quickly.”

For more information, take a look at the press release.

The new kits will be available in early 2023. Engineers can find out more on the XPLR-HPG-1 and XPLR-HPG-2 pages on the u-blox website.

Renesas expands the RZ/V series of microprocessors for improved machine vision AI applications

Posted on 28 October, 202228 October, 2022 by Abhishek Jadhav

Recently, Renesas Electronics Corporation expanded its existing RZ/V family of microprocessors to add a new device that enables AI processing of image data from multiple cameras. This data process offers improved accuracy of image recognition for machine vision AI applications. The new hardware platform, RZ/V2MA microprocessor, is equipped with two 64-bit Arm Cortex cores and delivers high computing performance with an operating frequency of 1GHz.

Renesas RZ/V2MA MPU comes with high-speed interfaces, including Ethernet, USB, and PCIe expansion ports that allow image input from multiple camera sources. Alongside the DRP-AI accelerator, the RZ/V2MA MPU also includes an OpenCV accelerator that allows rule-based image processing. This brings high accuracy for image recognition for a wide range of applications, such as AI-equipped gateways, video servers, security gates, POS terminals, and robotic arms.

“With the new DRP-AI TVM tool, we are offering designers the option to expand AI frameworks and AI models that can be converted to executable formats, allowing them to bring the latest image recognition capabilities to embedded devices using new AI models,” said Shigeki Kato, Vice President of Renesas’ Enterprise Infrastructure Business Division.

Other AI capabilities include the addition of DRP-AI TVM, an open-source deep learning compiler of Apache TVM technology. The DRP-AI translation is designed to convert the AI models into DRP-AI executables, while the DRP-AI TVM compiler lets the DRP-AI accelerator work in combination with the CPU. At the time of writing, Renesas supports ONNX and PyTorch AI models and has plans to support TensorFlow in the future.

The company has also developed a vision AI gateway solution, which is an AI-object detection and recognition platform that uses multiple cameras to collect and transmit data wirelessly. The solution uses the newly announced RZ/V2MA MPU in combination with other Renesas products such as power ICs, VersaClock clock generator, and communication modules for Wi-Fi, Bluetooth, and LTE. This solution provides users with flexible connectivity options and an optimized power supply system that has been tested to accelerate the development of AI devices.

“Renesas’ RZ/V series is ideal for embedded devices since it does not need fans or heat sinks, due to its extremely low power consumption and low heat capability when running AI,” said Chiharu Nakabayashi, President of Amnimo Inc., a provider of IoT and AI-based services and a wholly owned subsidiary of Yokogawa Electric Corporation.

For more information on the Renesas RZ/V2MA microprocessors, head to the official product page.

Home Assistant SkyConnect USB Stick comes with rich wireless connectivity got home automation

Posted on 28 October, 2022 by Abhishek Jadhav

Back in June 2022, according to the Home Assistant’s community blog post, Nabu Casa’s Zack Barett announced on Twitter that Home Assistant founders are working on the SkyConnect USB Stick, a physical radio adapter and dongle hardware. Seeed Studio has now started taking pre-orders for the Home Assistant SkyConnect USB Stick, compatible with Zigbee, Thread, and Matter, ideal for home automation applications.

In terms of wireless connectivity, which is an essential part of any automation application, the hardware dongle supports Zigbee and Matter specifications. The device is based on the Silicon Labs’ EFR32MG21 system-on-chip with an IEEE 802.15.4 multi-protocol 2.4GHz radio microcontroller that is capable of parallelly running both Zigbee 3.0 and Thread protocols. This is done using firmware in RCP RCP mode.

Silicon Labs’ EFR32MG21 system-on-chip is a high-performance, low-power, and secure IoT solution with support for multiprotocol wireless connectivity. The system-on-chip is optimized for line-powered Zigbee, Thread, and Bluetooth mesh applications, including connected lighting, gateways, voice assistants, and smart metering. The hardware features integrated security features and leverages Secure Vault technology, and physically unclonable function hardware to reduce the risk of IoT security breaches.

The widely adopted feature for home automation and IoT ecosystem– over-the-wire firmware update. The Home Assistant SkyConnect USB Stick also comes with an over-the-wire firmware update feature through the USB interface that allows easy and quick firmware updates for both connectivity protocols.

An important thing to note is that SkyConnect will not run Matter connectivity immediately upon launch, but using Home Assistant the firmware will be updated to enable Matter when the firmware is developed.

“This is due to Matter not being released until November and not all of the information being available. The Home Assistant is already working on their Matter implementation, but more information will come after November.”

The Home Assistant SkyConnect USB Stick is currently available in limited quantity and shipment after the official product launch is expected for November/December 2022. The dongle is priced at $29.99 USD and can be bought on the Seeed Studio product page.

Golioth IoT development platform combines several interesting features you’d want to use!

Posted on 28 October, 202228 October, 2022 by Abhishek Jadhav

Earlier this year, the Golioth IoT development platform announced the launch of its new product feature– a web console for creating a custom user interface for your IoT devices. The web console is already part of the Golioth IoT development platform feature suite, but the ability to create our own user interface to access and control the entire IoT fleet is interesting. This capability gives you the flexibility to send and receive data in real-time, check on the state of each device, and read logs from the hardware.

The company has written a detailed blog on using the web console and its custom UI feature to demonstrate this around the image of an Adafruit MagTag, a development board used for the developer training program. As the author notes, “It’s pretty cool to click a button on the image and see the device in your hand react.”

After this, Golioth announced a series of new features– one of which is the Golioth LightDB Stream, a purpose-built for collecting time-based data from devices deployed in the field. Golioth LightDB Stream sends data to the Golioth cloud and as soon as the data hits the server, it is timestamped and placed into a database. Using the Golioth Device SDK, the user only needs to understand the high-level function to send data, and everything else happens “under the hood.”

The team introduced account management, self-service upgrades, and the ability to manage teams on Golioth IoT development platform. The new organization’s feature gives teams, accessibility from the new account page and adds the ability for Golioth users to create and manage organizations within their user accounts. On the other hand, Golioth’s new self-service subscription makes the enterprise grow seamlessly.

Golioth CloudEvents is designed to format event data so that one cloud provider can understand the incoming data from another cloud provider. The standard includes things like the type of event, source of the event, and formatting of the event. Golioth helps users get their data to and from resource-constrained IoT devices in the field. Interestingly, the latest feature that comes into the market from Golioth is updating device firmware over the air, and this has become a critical feature for every IoT deployment.

“As a device management cloud, Golioth delivers all of the tools you need for dependable OTA. That means associating new firmware binaries with specific hardware variants in your fleet, using a rollout process that informs the device a new version is available, and confirming the state of the currently running firmware.”

For more information, head to the official website and start using Golioth IoT development platform for scale.

Intel’s Horse Creek Platform Combines SiFive Performance P550 Quad-core RISC-V Processor with PCIe 5.0 and 8GB of DDR5 Memory

Posted on 28 October, 2022 by Emmanuel Odunlade

Last year after SiFive introduced its Performance P550 quad-core processor – the fastest 64-bit RISC-V processor so far with a SPECInt 2006 score of 8.65/GHz – Intel made it known that it was partnering with SiFive to develop the Horse Creek platform which was to be powered by the processor and manufactured with an Intel 4 process.

The fact that Intel is positioning itself as a foundry that can build chips for clients like SiFive is also part of what makes the Horse Creek platform interesting. The development of the new high-performance RISC-V platform is not just a continuation of SiFive’s HiFive development boards designed to grow the RISC-V ecosystem, but also a part of Intel’s efforts to catalyze the adoption of RISC-V and build their Foundry Services.

“We are pleased to be a lead development partner with SiFive to showcase to mutual customers the impressive performance of their P550 on our 7nm Horse Creek platform,” says Amber Huffman, Intel’s Fellow and CTO of IP engineering group. “By combining Intel’s leading edge interface IP such as DDR and PCIe with SiFIve’s highest performance processor, Horse Creek will provide a valuable and expandable development vehicle for cutting-edge RISC-V applications.”

A prototype of the dev platform was showcased for the first time during the Intel 2022 Innovation Developer Conference held earlier this year. The Horse Creek is a Raspberry-Pi-inspired RISC-V software development board integrated with 8GB of DDR5 memory, PCIe 5.0 slot, SPI Flash, an SD card reader for storage, Linux-based software, and of course, a Cortex-A75 quad-core RISC-V processor running at up to 2.2 GHz, each core featuring private L1 and L2 caches with a shared last-level cache. No built-in GPU but the CPU should be powerful enough to perform basic computing tasks on its own since Intel demonstrated it executing a video game along with various other applications.

Specifications:

Processor
- SiFive P500 quad-core processor @ up to 2.2 GHz with:
- a 13-stage, 3-issue high-performance out-of-order pipeline
- Each core has private 32KB + 32KB L1 cache and 256KB L2 cache
- Up to 4MB L3 cache in a quad-core cluster
- a SPECint 2006 score of 8.65/GHz
Memory
- DDR5-5600 interface
PCIe
- Intel PCIe 5.0 PHY with x8 lanes
- Synopsys PCIe Root Hub Controller
Others
- I3C, Quad and Octal SPI, UART, peripheral DMA
Package
- 19 x 19 standard FBGA Package
Process
- Most advanced Intel 4 Process Technology
Software
- Supports Ubuntu 20.04 with Linux 5.17.4

Further Details

There is no detail on how much it will cost or when it will be available for purchase yet but you can get other useful details on WikiChip. Intel might also choose to not sell the chip but just use it as a foundry platform for others to design their own chips.

Meet the Modern Mini CB-11 Waterproof Doorbell from Avantek

Posted on 26 October, 202228 October, 2022 by Emmanuel Odunlade

Introduction

Going wireless has become a perfectly convenient way to do a lot of things these days including adding a doorbell wherever it may be needed around your homes, offices, and workspaces. There are currently a number of wireless mini doorbells out there in the market and for a non-technical person, identifying and making the right choice amongst them might actually be a bit hard. So, we thought it necessary to review one of the coolest we have ever come across, named the Mini CB-11 Waterproof Doorbell Chime.

Made by Avantek, a US-based engineering company specializing in making advanced solutions for the home and business, the Mini CB-11 Waterproof Doorbell is a high-quality compact-sized digital doorbell system designed for full home coverage, with an extensive working distance that ensures you never miss any important guests or urgent deliveries, even if you are a few rooms away or on another floor.

Built with convenience in mind, the mini wireless doorbell comes in a wireless easy mounting design that ensures your interior decorations are not affected. The receiver is made compact enough so that it won’t block off any neighboring power sockets.

One of the best features of the Avantek digital doorbell is its innovative memory function. There are actually up to 52 tune melodies that you can choose from depending on your taste and preference. You can use the “Previous” and “Next” buttons located by the side of the device to go back and forth when going through the list. These chimes are amplified through a built-in 28 mm speaker that delivers rich and crisp sound, along with up to 5 adjustable volume levels with a maximum of 115 dB. Once the melody and volume are set, the device’s memory stores it and is able to recall it when required. So you don’t have to worry about setting the device all over again each time you put it on, even after a power outage. The memory will automatically take you to the last settings on the device. This makes it not only very convenient to use but also very friendly to operate.

The doorbell is also environmentally-friendly, having an IP55 dust-proof and water-resistant transmitter that is strong enough to work well in almost all outdoor environments. The transmitter will even work in the harshest weather conditions given its very wide -16°C to 60°C operating temperature range.

Working Range

The doorbell also boasts a 300m/1000ft ultra-long wireless working range and strong signal penetration that makes it almost impossible not to reach you wherever you are in your house. This range, however, depends on your operating environment as we are learned that metals affect the signal and shortens the transmission range, so you might want to keep it away from metals. The doorbell can be used in homes, factories, offices, classrooms, apartment blocks, villas, and large areas.

Highlights of the Avantek Mini Wireless DoorBell

Mini yet mighty with great performance features
Extreme low power consumption
Super long three-year battery life
Built-in 28 mm stereo speaker that delivers rich and crisp sound
Up to 52 external melodies; you can easily choose between the previous and next.
Offers a massive 300m distance range depending on the environment
A durable and water-resistant transmitter that makes it ideal for various outdoor environments
Convenient to use; it has a memory that remembers your previous melody and volume level.

Specifications:

Best Suited for Doorbell
WiFi Wireless connectivity
52 Melodies
Square-shaped Button
3V Voltage
Type: Digital, Door Chime
Color: Black
Material: ABS
Design: Self-Adhesive or screw-in
Sound level: 115dB
Operating Distance Range: 1000 ft
Thickness: 1.5 cm (0.6 inches)
Batteries: CR2032 Lithium
Switch style: Remote
Bulb Type: LED
Power Source: Yes
Average Battery Life: 3 years
Dimensions: 7.4cm x 8cm (2.8 in x 3.1 in)
Operating Temperature Range: -16°C to 60C
Weight: 136 grams
Warranty: 3 years
Certification and Compliance: FCC, CE, and RoHS

The maximum ring volume of the doorbell is 115dB which will obviously sound loud enough in your eyes even if you are at a party, while the lowest is 0 dB which can not even wake a sleeping baby.

Video

What’s More? You don’t have to go through the stress of regular battery purchase and replacement since the wireless doorbell system has a built-in 230mAh CR2032 lithium battery that can comfortably power the device for an estimated lifespan of 3 years.

The mini doorbell package comes along with the accessories that you need for setting it up, such as anchors, a screwdriver, a piece of double-sided adhesive tape, a plug-in receiver, screws, and a transmitter. The device is easy to set up; all you need to do is to just plug the mini receiver into a socket and fix the transmitter on a wall. A user manual guide is also included in the package to simplify things.

Interested in getting the mini doorbell which sells for $19.99? Well, you are in luck as our readers get a whopping 15% off purchases using the code ATDBCB15

Polyn Technology announces the availability of NeuroVoice, an AI solution for voice processing

Posted on 25 October, 202225 October, 2022 by Abhishek Jadhav

Polyn Technology has been providing tinyML solutions with its NASP technology, a neuromorphic analog signal processing, ideal for next-gen sensor-level devices in real-time edge AI applications. The company has now announced the availability of its NeuroVoice AI solution for on-chip voice extraction from a noisy background.

The news follows its primary testing of AI chips based on its NASP technology back in April. At that time, the company projected its mass production in the first quarter of 2023– but why not, the AI solution hits the market early. The NASP chip enables data processing disaggregation between the sensor node and the cloud– to mimic the tinyML concept.

“Earbuds, smartphones, hearing assistance, gamer headphones, and intercom systems need new technology to bring voice processing to a new level,” said Eugene Zetserov, VP of Marketing and Business Development of POLYN. “Current methods of voice signal processing are power hungry and, in some cases, fall short. A neural network is a perfect tool for voice processing and POLYN offers it on a tiny neuromorphic analog chip.”

This ultra-low power tinyML solution addresses voice extraction problems from a noisy environment. As an always-on AI chip, the hardware platform is adaptive to different types of noises with cloud independence for voice processing. Some of the additional voice features are– voice activity detection, wake word detection, and keyword spotting. The ability to detect the presence or absence of voice in a sound is very helpful for a wide range of edge applications.

NeuroVoice AI solution also supports many other voice processing features, and all of them in combination helps the user to build an advanced voice processing product for any use case requirements. The hardware has a very long battery life due to its ultra-low power consumption when compared to a traditional digital signal processor. Also, comes in a small IC footprint for easy integration into embedded applications.

“AI-based extraction of the voice signal in a noisy environment, including irregular noises, provides a better hearing experience than standard noise cancellation filters. A combination of NASP technology for voice extraction for clear communication and other voice processing features gives NeuroVoice an additional advantage,” Zetserov said.

For more details on the NeuroVoice AI solution, head to the official product page.