The Kynix Blog

In recent years Artificial intelligence (AI) has become a technology that global companies are desperately trying to take advantage of, as it is one of the most emerging and competitive technologies. However, a lot of AI technologies focus on the software, with operating speeds low which makes them a poor fit for mobile devices. For this reason big companies are focusing on developing AI with low power and high speeds, hoping to make AI fit for mobile use.   Professor Hoi-Jun Yoo of the Department of Electrical Engineering, along with his research team and collaboration with start-up company, UX Factory Co, has developed a semiconductor chip, CNNP (CNN Processor), which runs AI algorithms with ultra-low power, and K-Eye, a face recognition system using CNNP. Consisting of two different formats, the K-Eye series is available as a wearable type and a dongle type. The wearable type device can be used with a smartphone via Bluetooth, and it can operate for more than 24 hours with its internal battery. By conveniently hanging the K-Eye around their necks users can check information about people by using their smartphone or smart watch, which connects K-Eye and allows users to access a database via their smart devices. A smartphone with K-EyeQ, the dongle type device, can recognise and share information about users at any time.  It works by recognising an authorised user looking at the screen, which then automatically turns the smartphone on, without a fingerprint, passcode or iris authentication. The smartphone cannot be tricked by the user’s photograph, as it can distinguish whether an input face is coming from a saved photograph versus a real person. Other distinct features are carried out by the K-Eye series. Detecting a face at first and then recognising it is one, and it is possible to maintain ‘Always-on’ status with low power consumption of less than 1mW. The research team devised two key technologies to complete this: an image sensor with ‘Always-on’ face detection and the CNNP face recognition chip.  The ‘Always-on’ image sensor, the first key technology, is able to determine if there is a face in its camera range. Then, it can capture frames and set the device to operate only when a face exists, reducing the standby power significantly. Additionally the face detection sensor combines analogue and digital processing to reduce power consumption. Using this approach, the analogue processor, combined with the CMOS Image Sensor array, distinguishes the background area from the area likely to include a face, and the digital processor then detects the face only in the selected area. Therefore, it becomes effective in terms of frame capture, face detection processing, and memory usage.    Following this the second key technology, CNNP, is able to achieve incredibly low power consumption, by optimising a convolutional neural network (CNN) in the areas of circuitry, architecture, and algorithms. Specially designed to enable data to be read in a vertical direction as well as in a horizontal direction, the on-chip memory integrated in CNNP also has immense computational power with 1024 multipliers and accumulators operating in parallel and is capable of directly transferring the temporal results to each other without accessing to the external memory or on-chip communication network. Additionally, convolution calculations with a two-dimensional filter in the CNN algorithm are approximated into two sequential calculations of one-dimensional filters to achieve higher speeds and lower power consumption.  CNNP achieved 97% high accuracy but consumed only 1/5000 power of the GPU thanks to these new technologies. Face recognition can be performed with only 0.62mW of power consumption, and the chip can show higher performance than the GPU by using more power.  Developed by Kyeongryeol Bong, a PhD student under Professor Yoo, these chips were presented at the International Solid-State Circuit Conference (ISSCC) held in San Francisco earlier this year. CNNP, which has the lowest reported power consumption in the world, has achieved a huge amount of attention, which has led to the development of the present K-Eye series for face recognition.  Professor Yoo commented: “AI - processors will lead the era of the Fourth Industrial Revolution. With the development of this AI chip, we expect Korea to take the lead in global AI technology.”  Ref.MT9V022 OV05633

The CC2640R2F SimpleLink ultra-low-power wireless microcontroller from Texas Instruments (TI) is in stock at Mouser Electronics. Part of TI’s CC26xx SimpleLink family of 2.4GHz devices, the CC2640R2F microcontroller features a small, single-chip system that integrates a flash-based microcontroller and Bluetooth Smart radio to target Bluetooth 4.2 and Bluetooth 5 low-energy applications. The microcontroller combines a 61μA/MHz ARM Cortex-M3 microcontroller and a rich peripheral set that includes an 8.2μA/MHz sensor controller. The 48MHz ARM microcontroller offers 128 kBytes of flash and 28 kBytes of SRAM and supports over-the-air (OTA) updates. The sensor controller is ideal for interfacing external sensors and for collecting analog and digital data autonomously while the rest of the system is in sleep mode. The device includes a 12-bit analogue-to-digital converter, up to 31 general-purpose inputs and outputs (GPIOs), and built-in robust security on chip with one of the simplest radio frequency (RF) and antenna designs available. Minimal RF expertise is required to implement the device, which helps make development and layout extremely easy. The wireless microcontroller is available in 2.7×2.7 mm WCSP and 4×4, 5×5 and 7×7 mm QFN packages, and is designed for a board array of wireless Internet of Things (IoT) applications, including health and fitness, industrial, and home and building automation. With ready-to-use protocol stacks (including the SIMPLELINK-CC2640R2-SDK software development kit for Bluetooth 5), the SimpleLink portfolio of wireless connectivity solutions not only offers designers maximum flexibility and support but also delivers multi-standard capabilities with code- and pin-compatibility across Bluetooth Smart, 6LoWPAN, ZigBee and ZigBee RF4CE.   Ref: KY32-MB91F376GPMCR-GS KY32-MB90F548GSPFV-G KY32-HD6417604SVF20

Bigger is always better, isn’t it? That’s not necessarily the case when it comes to specifying an AC/DC power supply. One of the most important aspects of designing a power supply into a system is ensuring that it is sized appropriately. Erring on the side of caution by trying to ensure that the supply’s maximum output exceeds that of the load is no longer the right answer in many cases. Customers increasingly need to focus on energy efficiency. The trend is partly driven by the need to cut operating costs and partly by legislation such as the European Union’s EcoDesign Directive.Under the directive, manufacturers of energy related products need to be able to demonstrate they have taken environmental factors into account. The efficiency of the power delivery sub-system is one of the key factors. It will play a large part in determining how energy will be lost through heat. As a result, choosing a high efficiency PSU (Power Supply Unit) is an important consideration in the design process.A 200W PSU operating at full load with an efficiency of 85% will lose 30W in waste heat. Not only is that heat wasted, there may be an additional energy cost in forced air cooling to prevent the rest of the system overheating. A PSU that is 90% efficient will cut the power wastage by 10W.If that PSU is operated below full capacity, it will run at a lower temperature. That allows usage in higher ambient temperatures or with less forced air cooling. However, there can be a trade off between efficiency and headroom. Many PSUs are designed to provide peak efficiency when they are driving a load close to full capacity. But this efficiency can roll off dramatically beneath 70 or 80% of full load. Using a power supply that is oversized for a particular application may result in an undesirable loss in efficiency and excessive heat production.A potential problem for system designers is that the focus on energy efficiency in electronics has led to the adoption of power saving modes. The resulting load demands can vary widely during operation. Responding to this trend, PSU designers working in the data centre space have embraced initiatives such as 80 PLUS.Launched in the mid 2000s at a Market Transformation Symposium organised by the American Council for an Energy-Efficient Economy, the 80 PLUS idea was quickly adopted as the basis for PSU efficiency marking by the US Energy Star programme. Supported worldwide, the idea behind 80 PLUS was to make PSUs deliver high efficiency over a larger proportion of the load curve. Recognising that many data centre system PSUs are operated using 1+1 redundancy and current sharing, the maximum efficiency point was centred on 50% capacity.Ratings range from Bronze to Titanium. At 50% load, Bronze offers an efficiency of 85%. Titanium pushes the peak efficiency to 96%, rolling off towards 94% when operating at 20% load and 91% at full load. Bel and GE provide wide ranges of PSUs that are graded according to the 80 PLUS standards.An alternative way to approach the issue of variable loads is to use the idea of boost power. This concept is gaining popularity in industrial designs where engineers have to deal with highly capacitive and inductive loads such as motors. As these systems shift between modes, there may be short term peak loads that go some way above normal operation. Motor start up also needs careful handling to deal with high current inrush conditions.Built for the DIN-rail format commonly used in industrial systems, the Cliq-II and Cliq-M series of DIN-rail supplies from Delta offer an ‘Advanced Power Boost’ of 120% for three seconds or 150% for five seconds, respectively. Alternatively, if a PSU has been derated to operate at a lower output level so that it does not need forced air cooling, it can be ramped up to peak load for short periods of time without necessarily demanding additional cooling. However, this usage of a PSU does call for attention to the thermal conditions to ensure that the short term peaks in heating are dissipated.To deal with situations such as high current inrush, PSUs such as the Artesyn LCM600 offer constant current modes that limit how much power is delivered to the load during the kinds of demand surges seen during motor startup. As one of a growing number of PSUs that employ digital control, the LCM600’s firmware can be programmed to support a number of different protection strategies so that the integrator can pick the mode best suited to the application.Reference:C300HC650HXBA-01

SummaryAn innovative new method to engineer computer chips more easily and cheaper than conventioanl methods have been developed by researchers who from the University of Exeter.This new technique to produce cutting-edge,versatile microchips could revolutionize the speed,efficiency and capability of the next-gen of computers.About the researchThe discovery could revolutionise the production of optoelectronic materials – or devices that produce, detect and control light – which are vital to the next generation of renewable energy, security and defence technologies, the researchers said.Dr Anna Baldycheva, from Exeter's Centre for Graphene Science and author of the paper said:"This breakthrough will hopefully lead to a revolution in the development of vital new materials for computer electronics. The work provides a solid platform for the development of novel next-generation optoelectronic devices. Additionally, the materials and methods used are extremely promising for a wide range of further potential applications beyond the current devices." This innovative new research focused on developing a versatile,multi-functional technology to significantly enhance future computing capabilities. The team used microfluidics technology, which uses a series of minuscule channels in order to control the flow and direction of tiny amounts of fluid. For this research, the fluid contains graphene oxide flakes,that are mixed together in the channels, to construct the chips.While the graphene oxide flakes are two-dimensional- consisting of length and width only- the research team used a new sophisticated light-based system to drive the assembly of the three-dimensional chip structures.Crucially, the research team have analysed their methodology to not only confirm the technique is successful, but also to provide a blueprint for others to use to help manufacture the chips. "We are very excited about the potential of this breakthrough and look forward to seeing where it can take the optoelectronics industry in the future." added by professor Monica Craciun, co-author of the paper and Associate Professor of Nanoscience at Exeter.  This article provide by University of Exeter,and the research is  published in the respected journal Scientific Reports.Article edited by kynix. 

The global lamp market has undergone tremedous changes over the last several years. Even government regulation,declining packaged LED and other component prices, and technological advancements have all played a major role in the continued penetration of LED into the overall market. Obvioulsy, LED lamps are the future of lighting, and there is a new invention using LEDs has been pushing out that Osram,a lighting company that offers innovative and sustainable lighting solutions, has revealed that he will stitch LEDs into workwear.  About this amazing news,Osram has taken a first step toward weaving smart lighting into clothes, announcing workwear that lights up with LEDs, while strongly hinting that interactive apparel is coming including a cycling jacket that illuminates when you hit the brakes, and lights that flash when your pulse rate rises too high. At the same time,Osram decided to stitching LEDs into safety vests and work jackets,and choosing Fvrth,a Germany-based safety and sportswear company uvex to operate with.Osram's first stab at textile illumination stops short of interactivity. It simply focuses on giving visibility to workers on job sites. “The textile illumination is incorporated into the safety clothing and ensures greater visibility and hence safety in day-to-day work, for example, on construction sites or in road traffic,” Osram said. “The key advantage of the new technology: Reflector strips on conventional work clothing only reflect incident light, while the light modules ensure active illumination at all times, thus improving safety when working in the dark or in poor visibility conditions.” Osram has been testing the technology for some time.The company used it to help illuminate ice hockey players, sticks, and pucks in an outdoor night game nearly10,000 ft high in the German Alps last year, for example. With sportswear as part of the mix, Osram plans to eventually add sensors that will enable integrated LEDs to respond to physical stimuli, providing health alerts, safety measures, and more.“This will in the future allow various applications to be controlled using an app,” Osram said. “Possible examples include sports clothing that warns the wearer about a high pulse rate via the light guides, or a cycling jacket with an integrated brake light.” Like many lighting companies, Osram is trying to establish LED lights and luminaires as nodes and backbones of information technology networks.While some industry observers expect that one day, OLED technology will prevail for integrated textile illumination, Osram has chosen instead to stitch LEDs, as have other illuminated clothing providers. OLED (organic light-emitting diode) is a patch of material that emits light in response to a current, whereas LEDs are single light points. Osram told LEDs Magazine that it has no plans to use OLEDs for lighting textiles. 

It's hardly a character flaw, but organic transistors—the kind envisioned for a host of flexible electronics devices—behave less than ideally, or at least not up to the standards set by their rigid, predictable silicon counterparts. When unrecognized, a new study finds, this disparity can lead to gross overestimates of charge-carrier mobility, a property key to the performance of electronic devices.If measurements fail to account for these divergent behaviors in so-called "organic field-effect transistors" (OFETs), the resulting estimates of how fast electrons or other charge carriers travel in the devices may be more than 10 times too high, report researchers from the National Institute of Standards and Technology (NIST), Wake Forest University and Penn State University. The team's measurements implicate an overlooked source of electrical resistance as the root of inaccuracies that can inflate estimates of organic semiconductor performance.Already used in light-emitting diodes, or LEDs, electrically conductive polymers and small molecules are being groomed for applications in flexible displays, flat-panel TVs, sensors, "smart" textiles, solar cells and "Internet of Things" applications. Besides flexibility, a key selling point is that the organic devices—sometimes called "plastic electronics"—can be manufactured in large volumes and far more inexpensively than today's ubiquitous silicon-based devices.A key sticking point, however, is the challenge of achieving the high levels of charge-carrier mobility that these applications require. In the semiconductor arena, the general rule is that higher mobility is always better, enabling faster, more responsive devices. So chemists have set out to hurry electrons along. Working from a large palette of organic materials, they have been searching for chemicals—alone or in combination—that will up the speed limit in their experimental devices.Just as for silicon semiconductors, assessments of performance require measurements of current and voltage. In the basic transistor design, a source electrode injects charge into the transistor channel leading to a drain electrode. In between sits a gate electrode that regulates the current in the channel by applying voltage, functioning much like a valve.Typically, measurements are analyzed according to a longstanding theory for silicon field-effect transistors. Plug in the current and voltage values and the theory can be used to predict properties that determine how well the transistor will perform in a circuit.Results are rendered as a series of "transfer curves." Of particular interest in the new study are curves showing how the drain current changes in response to a change in the gate electrode voltage. For devices with ideal behavior, this relationship provides a good measure of how fast charge carriers move through the channel to the drain."Organic semiconductors are more prone to non-ideal behavior because the relatively weak intermolecular interactions that make them attractive for low-temperature processing also limit the ability to engineer efficient contacts as one would for state-of-the-art silicon devices," says electrical engineer David Gundlach, who leads NIST's Thin Film Electronics Project. "Since there are so many different organic materials under investigation for electronics applications, we decided to step back and do a measurement check on the conventional wisdom."Using what Gundlach describes as the semiconductor industry's "workhorse" measurement methods, the team scrutinized an OFET made of single-crystal rubrene, an organic semiconductor with a molecule shaped a bit like a microscale insect. Their measurements revealed that electrical resistance at the source electrode—the contact point where current is injected into the OFET— significantly influences the subsequent flow of electrons in the transistor channel, and hence the mobility.In effect, contact resistance at the source electrode creates the equivalent of a second valve that controls the entry of current into the transistor channel. Unaccounted for in the standard theory, this valve can overwhelm the gate—the de facto¬ regulator between the source and drain in a silicon semiconductor transistor—and become the dominant influence on transistor behavior.At low gate voltages, this contact resistance at the source can overwhelm device operation. Consequently, model-based estimates of charge-carrier mobility in organic semiconductors may be more than 10 times higher than the actual value, the research team reports.Hardly ideal behavior, but the aim of the study, the researchers write, is to improve "understanding of the source of the non-ideal behavior and its impact on extracted figures of merit," especially charge-carrier mobility. This knowledge, they add, can inform efforts to develop accurate, comprehensive measurement methods for benchmarking organic semiconductor performance, as well as guide efforts to optimize contact interfaces.Reference:2SA1987C4706FJA4213RTU