Here are some of my collection of newest trends and predictions for year 2018. I have not invented those ideas what will happen next year completely myself. I have gone through many articles that have given predictions for year 2018. Then I have picked and mixed here the best part from those articles (sources listed on the end of posting) with some of my own additions to make this posting.This article contains very many quotations from those source articles (hopefully all acknowledged with link to source).
The general trend in electronics industry is that the industry growth have been driven by mobile industry. Silicon content in smartphones and other mobile devices is increasing as vendors add greater functionality. Layering on top of that are several emerging trends such as IoT, big data, AI and smart vehicles that are creating demand for greater computing power and expanding storage capacity.
Manufacturing trends
According to Foundry Challenges in 2018 article the silicon foundry business is expected to see steady growth in 2018. The growth in semiconductor manufacturing will remain steady, but there will be challenges in the manufacturing capacity and expenses to move to the next nodes. For most applications, unless you must have highest levels of performance, there may not be as compelling a business case to focus on the bleeding-edge nodes. Over the last two years, the IC industry has experienced an acute shortage of 200mm fab capacity (legacy MCU, power, sensors, 6-micron to 65nm). In 2018, 200mm capacity will remain tight. An explosion in 200mm demand has set off a frenzied search for used semiconductor manufacturing equipment that can be used at older process nodes. The problem is there is not enough used equipment available. The profit margins in manufacturing are so thin in markets served by those fabs that it’s hard to justify paying current rising equipment prices, and newcomers may have a tough time making inroads. Foundries with fully depreciated 200mm equipment and capacity already are seeing increased revenues in their 200mm business.The specialty foundry business is undergoing a renaissance, thanks to the emergence of 5G and automotive.
300mm is expected to follow a similar path for lack of capacity because 300mm fabs already produce leading-edge chips and more mainstream 300mm demand is driven by MCUs, wireless communications and storage applications. Early predictions are for solid growth in 2018, fueled by demand for memory and logic at advanced 10/7nm.
In 2017, marking the first time that the semiconductor equipment market has exceeded the previous market high of US$47.7 billion set in 2000. Fab tool vendors found themselves in the midst of an unexpected boom cycle in 2017, thanks to enormous demand for equipment in 3D NAND and, to a lesser degree, DRAM. In 2018, equipment demand looks robust, although the industry will be hard-pressed to surpass the record growth figures in 2017. In 2018, 7.5 percent growth is expected to result in sales of US$60.1 billion for the global semiconductor equipment market – another record-breaking year. Demand looks solid across the three main growth drivers for fab tool vendors—DRAM, NAND and foundry/logic.
Rising demand for chips is hitting the IC packaging supply chain, causing shortages of select manufacturing capacity, various package types, leadframes and even some equipment. Spot shortages for some IC packages began showing up in 2017, but the problem has been growing and spreading since then, so packaging customers may encounter select shortages well into 2018. Apple Watch 3 shipment growth to benefit Taiwan IC packagers in 2018.
Market for advanced packaging begins to diverge based on performance and price. Advanced Packaging is now viewed as the best way to handle large amounts of data at blazing speeds.
Moore’s law
Many recent publications say Moore’s Law is dead. Though Moore’s Law is dead may be experiencing some health challenges, it’s not time to start digging the grave for the semiconductor and electronics market yet.
Even smaller nodes are still being taken to use in high end chips. The node names are confusing. Intel’s 10nm technology is roughly equivalent to the foundry 7nm node.In 2018, Intel is expected to finally ramp up 10nm finally in the first half of 2018. In addition, GlobalFoundries, Samsung and TSMC will begin to ship their respective 7nm finFET processes. On the leading edge, GlobalFoundries, Intel, Samsung and TSMC start migrating from the 16nm/14nm to the 10nm/7nm logic nodes. It is expected that some chip-makers face some challenges on the road. Time will tell if GlobalFoundries, Samsung and TSMC will struggle at 7nm. Early predictions are for solid growth in 2018, fueled by demand for memory and logic at advanced 10/7nm. 7nm is projected to generate sales from $2.5 billion to $3.0 billion in 2018. Over time 10nm/7nm is expected to be a big and long-running node. Suppliers of FPGAs and processors are expected to jump on 10nm/7nm.
South Korea’s Samsung Electronics said it has commenced production of the second generation of its 10nm-class 8-Gb DDR4 DRAM. Devices labeled 10nm-class have feature sizes as small as 10 to 19 nanometers. With the continued need for shrinking pattern dimensions, semiconductor manufacturers continue to implement more complex patterning techniques, such as advanced multi-patterning, for the 10nm design node and beyond. They also are investing significant development effort in readying EUV lithography for production at the 7/5nm design nodes. Samsung is planning to begin transitioning to EUV for logic chips next year at the 7nm node, although it is unclear when the technology will be put into production for DRAM.
There will be talk on even smaller nodes. FinFETs will get extended to at least to 5nm, and possibly 3nm in next 5 years. The path to 5nm loks pretty clear. FinFETs will get extended at least to 5nm. It’s possible they will get extended to 3nm. EUV will be used at new nodes, followed by High NA Lithography. New smaller nodes challenges the chip design as abstractions become more difficult at 7nm and beyond. Models are becoming more difficult to develop, integrate and utilize effectively at 10/7nm and beyond as design complexity, process variation and physical effects add to the number of variables that need to be taken into account. Materials and basic structures may diverge by supplier, at 7 nm and beyond. Engineering and scientific teams at 3nm and beyond will require completely different mixes of skills than today.
Silicon is still going strong, but the hard fact is that CMOS has been running out of steam for several nodes, and that becomes more obvious at each new node. To extend into new markets and new process nodes Chipmakers Look To New Materials. There are a number of compounds in use already (generally are being confined to specific niche applications), such as gallium arsenide, gallium nitride, and silicon carbide. Silicon will be supplemented by 2D materials to extend Moore’s Law. Transition metal dichalcogenides (TMDCs), a class of 2D materials derived from basic elements—principally tellurium, selenium, sulfur, and oxygen—are being widely explored by researchers. TMDCs are functioning as semiconductors in conjunction with graphene. Graphene, the wonder material rediscovered in 2004, and a host of other two-dimensional materials are gaining ground in manufacturing semiconductors as silicon’s usefulness begins to fade. Wide-bandgap semiconductor materials like gallium nitride (GaN) and silicon carbide (SiC) are anticipated to be used in many more applications in 2018. Future progress increasingly will require a mix of different materials and disciplines, but silicon will remain a key component.
Interconnect Materials need to to be improved. For decades, aluminum interconnects were the industry standard. In the late 1990s, chipmakers switched to copper. Over the years, transistors have decreased dramatically in size, so interconnects also have had to scale in size leading to roadblock known as the RC challenge. Industry is investing significant effort in developing new approaches to extend copper use and finding new metals. There’s also some investigation into improvements on the dielectric side. The era of all-silicon substrates and copper wires may be coming to an end.
Application markets
Wearables are a question mark. Demand for wearables slowed down in 2017 so much that smart speakers likely outsold wearable devices in 2017 holiday season. eMarketer is estimating that usage of wearable will grow just 11.9 percent in 2018, rising from 44.7 million adult wearable users in 2017 to 50.1 million in 2018. On the other hand market research firm IDC estimates that the shipments of wearable electronics devices are projected to more than double over the next five years as watches displace fitness trackers as the biggest sellers. IDC forecasts that wearables shipments will increase at a compound annual growth rate of 18.4 percent between 2017 and 2021, rising from 113.2 million this year to 222.3 million in 2021. At the same time fitness trackers are expected to become commodity product. Tomorrow’s wearables will become more fully featured and multi-functional.
The automotive market for semiconductors is shifting into high gear in 2018. Right now the average car has about $350 worth of semiconductor content, but that is projected to grow another 50% by 2023 as the overall automotive market for semiconductors grows from $35 billion to $54 billion. The explosion of drive-by-wire technology, combined with government mandates toward fully electric powertrains, has changed this paradigm—and it impacts more than just the automotive industry. Consider implications beyond the increasingly complex vehicle itself, including new demands on supporting infrastructure. The average car today contains up to 100 million lines of code. Self-driving car will have considerably more code in it. Software controls everything from safety critical systems like brakes and power steering, to basic vehicle controls like doors and windows. Meeting ISO 26262 Software Standards is needed but it will not make the code bug free. It’s quickly becoming common practice for embedded system developers to isolate both safety and security features on the same SoC. The shift to autonomous vehicles marks a major shift in the supply chain—and a major opportunity.
Many applications have need for a long service life — for example those deployed within industrial, scientific and military industries. In these applications, the service life may exceed that of component availability. Replacing an advanced, obsolete components in a design can be very costly, potentially requiring an entire redesign of the electronic hardware and software. The use of programmable devices helps designers not only to address component obsolescence, but also to reduce the cost and complexity of the solution. Programmable logic devices are provided in a range of devices of different types, capabilities and sizes, from FPGAs to System on Chips (SoC) and Complex Programmable Logic Devices (CPLD). The obsolete function can be emulated within the device, whether it is a logic function implemented in programmable logic in a CPLD, FPGA or SoC, or a processor system implemented in an FPGA or SoC.
Become familiar with USB type C connector. USB type C connector is becoming quickly more commonplace than any other earlier interface. In the end of 2016 there were 300 million devices using a USBC connection – a big part was smartphones, but the interface was also widespread on laptops. With growth, the USBC becomes soon the most common PC and peripheral interface. Thunderbolt™ 3 on USBC connector promises to fulfill the promise of USB-C for single-cable docking and so much more.
Power electronics
The power electronics market continues to grow and gain more presence across a variety of markets. 2017 was a good year for electric vehicles and the future of this market looks very promising. In 2017, we saw also how wireless charging technology has been adopted by many consumer electronic devices- including Apple smart phones. Today’s power supplies do more than deliver clean and stable dc power on daily basis—they provide advanced capabilities that can save you time and money.
Wide-bandgap semiconductor materials like gallium nitride (GaN) and silicon carbide (SiC) are anticipated to be used in many more applications in 2018. At the moment, the number of applications for those materials is steadily increasing in the automotive and military industry. Expect to see more adoption of SiC and GaN materials in automotive market.
According to Battery Market Goes Bigger and Better in 2018 article advances in battery technologies hold the keys to continuing progress in portable electronics, robotics, military, and telecommunication applications, as well as distributed power grids. It is difficult to see lithium-ion based batteries being replaced anytime soon, so the advances in battery technology are primarily through the application of lithium-ion battery chemistries. New battery protection for portable electronics cuts manufacturing steps and costs for Lithium-ion.
Transparency Market Research analysts predict that the global lithium-ion battery market is poised to rise from $29.67 billion in 2015 to $77.42 billion in 2024 with a compound annual growth rate of 11.6 %. That growth has already spread from the now ubiquitous consumer electronics segment to automotive, grid energy, and industrial applications. Dramatic increase is expected for battery power for the transportation, consumer electronic, and stationary segments. According to Bloomberg New Energy Finance (BNEF), the global energy-storage market will double six times between 2016 and 2030, rising to a total of 125 G/305 gigawatt-hours. In 2018, energy-storage systems will continue proliferating to provide backup power to the electric grid.
Memory
Memory business boomed in 2017 for both NAND and DRAM. The drivers for DRAM are smartphones and servers. Solid-state drives (SSDs) and smartphones are fueling the demand for NAND. Both the DRAM and NAND content in smartphones continues to grow, so memory business will do well in 2018.Fab tool vendors found themselves in the midst of an unexpected boom cycle in 2017, thanks to enormous demand for equipment in 3D NAND and, to a lesser degree, DRAM. In 2018, equipment demand looks robust, although the industry will be hard-pressed to surpass the record growth figures in 2017.
NAND Market Expected to Cool in Q1 from the crazy year 2017, but it is still growing well because there is increasing demand. The average NAND content in smartphones has been growing by roughly 50% recently, going from approximately 24 gigabytes in 2016 to approximately 38 gigabytes today.3D NAND will do the heavy memory lifting that smartphone users demand. Contract prices for NAND flash memory chips are expected to decline in during the first quarter of 2018 as a traditional lull in demand following the year-end quarter.
Lots of 3D NAND will go to solid state drives in 2018. IDC forecasts strong growth for the solid-state drive (SSD) industry as it transitions to 3D NAND. SSD industry revenue is expected to reach $33.6 billion in 2021, growing at a CAGR of 14.8%. Sizes of memory chips increase as number of layer in 3D NAND are added. We’ve already scaled up to 48 layers. Does this just keep scaling up, or are there physical limits here? Maybe we could see a path to 256 layers in few years.
Memory — particular DRAM — was largely considered a commodity business. Though that it’s really not true in 2017. DRAM memory marked had boomed in 2017 at the highest rate of expansion in 23 years, according to IC Insights. Skyrocketing prices drove the DRAM market to generate a record $72 billion in revenue, and it drove total revenue for the IC market up 22%. Though the outlook for the immediate future appears strong, a downturn in DRAM more than likely looms in the not-too-distant future. It will be seen when there are new players on the market. It is a largely unchallenged assertion that Chinese firms will in the not so distant future become a force in semiconductor memory market. Chinese government is committed to pumping more than $160 billion into the industry over a decade, with much of that ticketed for memory startups.
There is search for faster memory because modern computers, especially data-center servers that skew heavily toward in-memory databases, data-intensive analytics, and increasingly toward machine-learning and deep-neural-network training functions, depend on large amounts of high-speed, high capacity memory to keep the wheels turning. The memory speed has not increased as fast as the capacity. The access bandwidth of DRAM-based computer memory has improved by a factor of 20x over the past two decades. Capacity increased 128x during the same period. For year 2018 DRAM remains a near-universal choice when performance is the priority. There has been some attempts to very fast memory interfaces. Intel the company has introduced the market’s first FPGA chip with integrated high-speed EMBED (Embedded Multi-Die Interconnect Bridge): The Stratix 10 MX interfaces to HMB2 memory (High Memory Bandwidth) that offers about 10 times faster speed than standard DDR-type DIMM.
There is search going on for a viable replacement for DRAM. Whether it’s STT-RAM or phase-change memory or resistive RAM, none of them can match the speed or endurance of DRAM. Necessity is the mother of invention, and we see at least two more generations after 1x. XPoint is also coming up as another viable memory solution that could be inserted into the current memory architecture. It will be interesting to see how that plays out versus DRAM.
5G and IoT
5G something in it for everyone. 5G is big. 5G New Radio (NR) wireless technology will ultimately impact everyone in the electronics and telecommunications industries. Most estimates say 2020 is when we will ultimately see some real 5G deployments on a scale. In the meantime, companies are firming up their plans for whatever 5G products and services they will offer. Though test and measurement solutions will be key in the commercialization cycle. 5G is set to disrupt test processes. If 5G takes off, the technology will propel the development of new chips in both the infrastructure and the handset. Data centers require specialty semiconductors from power management to high-speed optical fiber front-ends. 5G systems will drive more complexity in RF front-ends .5G will offer increased capacity and decreased latency for some critical applications such as vehicle-to-vehicle (V2V) or vehicle-to-infrastructure (V2I) communications for advanced driver assistance systems (ADAS) and self-driving vehicles. The big question is whether 5G will disrupt the landscape or fall short of its promises.
Electronics manufacturers expect a lot from Internet of Thing. The evolution of intelligent electronic sensors is creating a revolution for IoT and Industrial IoT as companies bring new sensor-based, intelligent systems to market. The business promise is that the proliferation of smart and connected “things” in the Industrial Internet of Things (IIoT) provides tremendous opportunities for increased performance and lower costs. Industrial Internet of Things (IIoT) has a market forecast approaching $100 billion by 2020. Turning volumes of factory data into actionable information that has value is essential. Predictive maintenance and asset tracking are two big IoT markets to watch in 2018 because they will provide real efficiencies and improved safety. It will be about instrumenting our existing infrastructures with sensors that improve their reliability and help predict failures. It will be about tracking important assets through their lifecycles.
A new breed of designers has arrived that is leveraging inexpensive sensors to build the intelligent systems at the edge of the Internet of Things (IoT). They work in small teams, collaborate online, and they expect affordable design tools that are easy to use in order to quickly produce results. Their goal is to deliver a functioning device or a proof-of-concept to their stakeholders while spending as little money as possible to get there. We need to become multi-functional engineers who can comfortably work in the digital, RF, and system domains.
The Io edge sensor device usually needs to be cheap. Simple mathematical reasoning suggests that the average production cost per node must be small, otherwise the economics of the IoT simply are not viable. Most suppliers to the electronics industry are today working under the assumption that the bill-of-materials (BoM) cost of a node cannot exceed $5 on average. While the sensor market continues to garner billions of dollars, the average selling price of a MEMS sensor, for example, is only 60 cents.
Designing a well working and secure IoT system is still hard. IoT platforms are very complex distributed systems and managing these distributed systems is often an overlooked challenge. When designing for the IoT, security needs to be addressed from the Cloud down to each and every edge device. Protecting data is both a hardware and a software requirement, as more data is being stored and analyzed in edge devices and gateways.
The continued evolution of powerful embedded processors is enabling more functionality to be consolidated into single heterogeneous multicore devices. You will see more mixed criticality designs – those designs which contain both safety-critical and non-safety critical processes running on the same chip. It’s quickly becoming common practice for embedded system developers to isolate both safety and security features on the same SoC.
AI
There is clearly a lot of hype surrounding machine learning (ML) and artificial intelligence (AI) fields. Over the past few years, machine learning (ML) has evolved from an interesting new approach that allows computers to beat champions at chess and Go, into one that is touted as a panacea for almost everything. Machine learning already has delivered beneficial results in certain niches, but it has potential for a bigger and longer lasting impact because of the demand for broad insights and efficiencies across industries. Also EDA companies have been investing in this technology and some results are expected to be announced.
The Battle of AI Processors Begins in 2018. Machine learning applications have a voracious appetite for compute cycles, consuming as much compute power as they can possibly scrounge up. As a result, they are invariably run on parallel hardware – often parallel heterogeneous hardware—which creates development challenges of its own. 2018 will be the start of what could be a longstanding battle between chipmakers to determine who creates the hardware that artificial intelligence lives on. Main contenders on the field at the moment are CPUs, GPUs, TPUs (tensor processing units), and FPGAs. Analysts at both Research and Markets and TechNavio have predicted the global AI chip market to grow at a compound annual growth rate of about 54% between 2017 and 2021.
Sources:
Battery Market Goes Bigger and Better in 2018
Smart speakers to outsell wearables during U.S. holidays, as demand for wearables slows
Wearables Shipments Expected to Double by 2021
The Week In Review: Manufacturing #186
Five technology trends for 2018
NI Trend Watch 2018 explores trends driving the future faster
Creating Software Separation for Mixed Criticality Systems
Isolating Safety and Security Features on the Xilinx UltraScale+ MPSoC
Meeting ISO 26262 Software Standards
DRAM Growth Projected to be Highest Since ’94
NAND Market Expected to Cool in Q1
Memory Market Forecast 2018 … with Jim Handy
3D NAND Storage Fuels New Age of Smartphone Apps
$55.9 Billion Semiconductor Equipment Forecast – New Record with Korea at Top
Advanced Packaging Is Suddenly Very Cool
Apple Watch 3 shipment growth to benefit Taiwan IC packagers in 2018
Rapid SoC Proof-of-Concept for Zero Cost
EDA Challenges Machine Learning
What Can You Expect from the New Generation of Power Supplies?
Optimizing Machine Learning Applications for Parallel Hardware
FPGA-dataa 10 kertaa nopeammin
Chipmakers Look To New Materials
What the Experts Think: Delivering the next 5 years of semiconductor technology
Programmable Logic Holds the Key to Addressing Device Obsolescence
The Battle of AI Processors Begins in 2018
For China’s Memory Firms, Legal Tests May Loom
Predictions for the New Year in Analog & Power Electronics
Lithium-ion Overcomes Limitations
Will Fab Tool Boom Cycle Last?
The Next 5 Years Of Chip Technology
Chipmakers Look To New Materials
Process Window Discovery And Control
Sensors are Fundamental to New Intelligent Systems
Industrial IoT (IIoT) – Where is Silicon Valley
Internet of things (IoT) design considerations for embedded connected devices
How efficient memory solutions can help designers of IoT nodes meet tight BoM cost targets
1,325 Comments
Tomi Engdahl says:
What’s the Difference Between DC-DC Conversion Topologies?
http://www.electronicdesign.com/power/what-s-difference-between-dc-dc-conversion-topologies?NL=ED-003&Issue=ED-003_20180302_ED-003_82&sfvc4enews=42&cl=article_1_b&utm_rid=CPG05000002750211&utm_campaign=15676&utm_medium=email&elq2=7eabf090ab86460bad32d6471d04b350
When modifying dc voltage from high to low or vice versa in a design, it’s important to have a clear understanding of the different converter types involved in the process.
Power-supply classifications come in many flavors. At the top level are the basic ac-ac, ac-dc, dc-ac, or dc-dc converters. In this article, we’ll discuss the dc-dc conversion category in terms of shifting the power signal down with a low-dropout regulator (LDO) or buck topologies, and up with a boost topology.
Tomi Engdahl says:
Cadence, Imec Disclose 3-nm Effort
64-bit CPU tapeout expected later this year
https://www.eetimes.com/document.asp?doc_id=1333016
Cadence Design Systems and the Imec research institute disclosed that they are working toward a 3-nm tapeout of an unnamed 64-bit processor. The effort aims to produce a working chip later this year using a combination of extreme ultraviolet (EUV) and immersion lithography.
So far, Cadence and Imec have created and validated GDS files using a modified Cadence tool flow. It is based on a metal stack using a 21-nm routing pitch and a 42-nm contacted poly pitch created with data from a metal layer made in an earlier experiment.
Tomi Engdahl says:
Magnesium-ion batteries; watching lithium ions; chemical reactions.
https://semiengineering.com/manufacturing-bits-feb-27/
Texas A&M University and others have discovered a new metal-oxide magnesium battery cathode material—a technology that promises to deliver a higher density of energy storage than today’s traditional lithium-ion (Li-ion) cells.
Magnesium-ion battery technology is promising. A battery consists of an anode (negative), cathode (positive), electrolytes and a separator. In a simple operation, ions are transported from the anode to the cathode and back.
In lithium-ion batteries, a carbon-based material makes up the anode. A compound based on lithium-cobalt-oxide is used for the cathode in mobile phones, laptops and other products. In addition, electric vehicles use lithium-ion batteries. Graphite is used for the anode. The cathode is based on lithium with other metals. For example, Tesla uses a battery based on lithium with a nickel-cobalt-aluminum (NCA) mix.
The problem? The safety and supply of these materials present some challenges.
http://www.science.tamu.edu/news/story.php?story_ID=1929#.WpCfLqrTnIX
Tomi Engdahl says:
Hey Big Spender! (For Semiconductor R&D, That’s Intel)
https://spectrum.ieee.org/view-from-the-valley/semiconductors/design/hey-big-spender-for-semiconductor-rd-thats-intel
The semiconductor industry in general is increasing its investments in research and development, but it will take a long time to challenge Intel’s dominant role.
That’s the conclusion of a report by IC Insights. The research firm indicated that overall industry spending, considering the top ten semiconductor companies (see chart, below), was up 6 percent in 2017 over 2016 to US $34 billion.
Tomi Engdahl says:
Getting EUV Ready for 2020
https://spectrum.ieee.org/nanoclast/semiconductors/nanotechnology/getting-euv-ready-for-2020
Extreme-ultraviolet lithography looks ready for its debut later this year, making it easier to build huge numbers of chips with even more huge numbers of the tiniest circuits you can buy. But will EUV be ready for the next generation, when circuits are slated to be even tinier? The Belgian microelectronics research house Imec has uncovered some problems with using EUV for the so-called 5-nanometer generation, which is expected to go into full production in late 2020. They are fixable, says Kurt Ronse, program director on advanced patterning at Imec, but there’s quite a lot of work to be done.
Tomi Engdahl says:
Computing With Random Pulses Promises to Simplify Circuitry and Save Power
https://spectrum.ieee.org/computing/hardware/computing-with-random-pulses-promises-to-simplify-circuitry-and-save-power
Yet the benefits of digital over analog are undeniable, which is why you see digital computers so often used to process signals with much more exactitude—and using much more energy—than is really required. An interesting and unconventional compromise is a method called stochastic computing, which processes analog probabilities by means of digital circuits. This largely forgotten technique could significantly improve future retinal implants and machine-learning circuits—to give a couple of applications we’ve investigated—which is why we believe stochastic computing is set for a renaissance.
Stochastic computing begins with a counterintuitive premise—that you should first convert the numbers you need to process into long streams of random binary digits where the probability of finding a 1 in any given position equals the value you’re encoding. Although these long streams are clearly digital, they mimic a key aspect of analog numbers: A minor error somewhere in the bitstream does not significantly affect the outcome. And, best of all, performing basic arithmetic operations on these bitstreams, long though they may be, is simple and highly energy efficient. It’s also worth noting that the human nervous system transfers information by means of sequences of neural impulses that strongly resemble these stochastic bitstreams.
Because basic arithmetic operations on such bitstreams are remarkably easy to accomplish.
An AND gate is a digital circuit with two inputs and one output that gives a 1 only if both inputs are 1. It consists of just a few transistors and requires very little energy to operate. Being able to do multiplications with it—rather than, say, programming a microprocessor that contains thousands if not millions of transistors—results in enormous energy savings.
Similarly simple circuits can carry out other arithmetic operations on these bitstreams. In contrast, conventional digital circuits require hundreds if not thousands of transistors to perform arithmetic, depending on the precision required of the results. So stochastic computing offers a way to do some quite involved mathematical manipulations using surprisingly little power.
Tomi Engdahl says:
Dana Cimilluca / Wall Street Journal:
Microchip Technology to acquire Microsemi, the largest US commercial supplier of military and aerospace semiconductor equipment, for ~$8.35B
Microchip Technology Agrees to Buy Microsemi
Microchip to pay $68.78 a share for Microsemi, or $8.3 billion
https://www.wsj.com/articles/microchip-technology-agrees-to-buy-microsemi-1519938593?mod=e2tw
It comes amid a wave of consolidation in the semiconductor industry as companies seek to cut costs amid fierce competition and position themselves for new applications. It would form a company with about $6 billion of annual revenue and generate some $300 million of so-called synergies within three years.
The two companies have complementary strengths, focusing to a large extent on different areas. Microsemi, based in Aliso Viejo, Calif., makes chips for communications and for aerospace-and-defense applications. Microchip, meanwhile, specializes in semiconductors for the industrial, auto and home-appliance markets.
Tomi Engdahl says:
Miniaturized Robots Combine Engineering, Biological, and Manufacturing Technologies
http://www.powerelectronics.com/robotics/miniaturized-robots-combine-engineering-biological-and-manufacturing-technologies
Shrinking miniature robots even further to the millimeter scale with conventional manufacturing techniques and components has proven difficult.
A milliDelta robot was developed by Robert Wood’s team at Harvard’s Wyss Institute for Biologically Inspired Engineering and John A. Paulson School of Engineering and Applied Sciences (SEAS). By integrating their microfabrication technique with high-performance composite materials that can incorporate flexural joints and bending actuators, the milliDelta can operate with high speed, force, and micrometer precision, which make it compatible with a range of micromanipulation tasks in manufacturing and medicine.
Tomi Engdahl says:
China Plans $31.5 Billion Fund to Reboot Chip Industry
https://www.eetimes.com/document.asp?doc_id=1333027
The Chinese government is planning a new 200 billion yuan ($31.5 billion) fund aimed at renewing efforts to kickstart its domestic chip industry and offset a huge trade deficit in imported semiconductors.
The state-backed China Integrated Circuit Industry Investment Fund Co. is in talks with government agencies and corporations, targeting the new financing, according to press reports, citing unidentified people familiar with the matter. Under the reported plan, the fund would begin disbursing money in the second half of 2018.
Tomi Engdahl says:
92 IC Wafer Fabs Closed or Repurposed From 2009-2017
150mm and 200mm wafer fabs accounted for two-thirds of total closures.
http://www.icinsights.com/news/bulletins/92-IC-Wafer-Fabs-Closed-Or-Repurposed-From-20092017/
Since the global economic recession of 2008-2009, the IC industry has been on a mission to pare down older capacity (i.e., ≤200mm wafers) in order to produce devices more cost-effectively on larger wafers.
Given the flurry of merger and acquisition activity seen in the semiconductor industry recently, the skyrocketing cost of new wafer fabs and manufacturing equipment, and as more IC companies transition to a fab-lite or fabless business model, IC Insights expects more fab closures in the coming years—a prediction that will likely please IC foundry suppliers.
Tomi Engdahl says:
Drones, Augmented Reality, UHD TV – High-End Video SoCs Need Emulation
https://www.mentor.com/products/fv/techpubs/download?id=102191&contactid=1&PC=L&c=2018_02_27_veloce_drones_emulation_wp_v3
New markets for multimedia and high-definition video chips are quickly becoming the next wave fueling the electronics industry. One market that demonstrates a mega-appetite for electronics is the drone…
Tomi Engdahl says:
“Swiss Army Knife for Engineers”
https://www.redpitaya.com/
Tomi Engdahl says:
OLED Pioneers Win NIHF Honor
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333031
The National Inventors Hall of Fame (NIHF) will recognize two chemists who pioneered OLED technology now used widely in the hottest consumer electronic products–from smartphones to TVs.
Chemists Ching Wan Tang and Steven Van Slyke pioneered organic light-emitting diodes. OLED displays represent an advance in flat panels that provides increased power efficiency, longer battery life and improved display quality. They are among 15 innovation pioneers who will be honored in May as part of the newest class of Inductees in the National Inventors Hall of Fame.
Tang joined Eastman Kodak in 1975. He hired Van Slyke, and together they applied the organic heterojunction–a bilayer structure of an electron donor and an electron acceptor invented by Tang–to various applications including OLEDs.
Tomi Engdahl says:
Compound semiconductor suppliers set to benefit from 5G
https://www.digitimes.com/news/a20180305PD210.html
Taiwan suppliers of compound semiconductor devices are poised to gain significant growth momentum from the increasing market demand for radio frequency (RF) and power amplifier (PA) components along with the efforts of chipmakers to roll out related solutions to support 5G devices such as smartphones, various IoT (Internet of Things), and base stations in the second half of 2018 and 2019 before 5G technologies and applications start to be commercialized in 2020, according to industry sources.
The upcoming 5G mobile communication era will not only witness the extension of frequency bands for the 5G handheld devices to the low range of sub-6GHz, but will also see sharp increases in demand for high-frequency, high-power compound semiconductor components such as RF, PA, integrated millimeter wave (mmWave) and GaN (gallium nitride) devices to support 5G infrastructure deployments, especially base stations, the sources said.
While worldwide telecom operators are aggressively proceeding with 5G deployments, the sources continued, Taiwan chipmaker MediaTek and suppliers of GaAs epi wafer and other compound semiconductor suppliers as Win Semiconductor, Advanced Wireless Semiconductor, Visual Photonics Expitaxy, Global Communications Semiconductors and IntelliEPI have directly or indirectly tapped into the supply chains of China telecom operators including China Mobile, and they are expecting significant business opportunities from tender projects presented by China’s telecom carriers.
Tomi Engdahl says:
Better Living Through Microelectronics
https://semiengineering.com/better-living-through-microelectronics/
From energy harvesting to prosthetics, semiconductor technologies have the potential to improve the human condition.
Tomi Engdahl says:
Finding Faulty Auto Chips
The road to zero defects requires some new tactics.
https://semiengineering.com/finding-faulty-auto-chips/
The next wave of automotive chips for assisted and autonomous driving is fueling the development of new approaches in a critical field
called outlier detection.
Outliers or faulty chips arise for several reasons, including the advent of latent reliability defects. These defects do not appear
when a device is shipped, but they are somehow activated in the field and could end up in a system.
Tomi Engdahl says:
Intel, TSMC vying to recruit engineering talent in Taiwan
https://www.digitimes.com/news/a20180305PD209.html
Intel and Taiwan Semiconductor Manufacturing Company (TSMC) are in a race for seeking new engineering talent in Taiwan, with the former aiming to support the capacity expansion at its 3D NAND flash plant in Dalian, China and the latter to ready new manpower for its new 5nm fab under construction in southern Taiwan, according to industry sources.
The sources said that Intel will kick off intensive interviews with applicants at National Taiwan University starting mid-March, claiming to offer pays that can well rival those offered by TSMC. The move has drawn attention from semiconductor players in Taiwan and China.
Tomi Engdahl says:
Intel Needs New Strategies, ASAP
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333043
After losing the No. 1 spot in semiconductor revenue for the first time since 1992, Intel needs new strategies to narrow the gap or surpass Samsung.
Samsung’s semiconductor division generated $69.1 billion in total revenue in 2017, eclipsing Intel’s $62.8 billion. Thanks to high memory prices, Samsung knocked Intel from the top spot in chip sales for the first time since 1992.
Tomi Engdahl says:
Cree Buys Infineon’s RF Power Biz for $430 Million
https://www.eetimes.com/document.asp?doc_id=1333042
LED and RF chip vendor Cree said it paid about $430 million to acquire Infineon Technologies’ RF Power business in a move to bolster the offerings of Cree’s Wolfspeed subsidiary in wireless, including faster 4G and 5G networks.
In 2016, Cree struck a deal to sell Wolfspeed to Infineon for $850 million, but the companies called it off last year after the Committee for Foreign Investment in the United States (CFIUS) raised concerns about the deal’s implications for U.S. national security.
Tomi Engdahl says:
10 Ways Digitization Will Drive the Electronics Supply Chain
https://www.eetimes.com/document.asp?doc_id=1333041
Tomi Engdahl says:
Is Advanced Packaging The Next SoC?
https://semiengineering.com/is-advanced-packaging-the-next-soc/
Why uncertainty about new applications coupled with splintering markets will redefine how chips are developed and sold.
Device scaling appears to be possible down to 1.2nm, and maybe even beyond that. What isn’t obvious is when scaling will reach that node, how many companies will actually use it, or even what chips will look like when foundries actually start turning out these devices using multi-patterning with high-NA EUV and dielectrics with single-digit numbers of atoms.
There are two big changes playing out across the semiconductor industry right now that ultimately will have a big impact on the answers to those questions. One involves a growing number of new markets and opportunities around artificial intelligence, whether that’s AI itself or related subcategories such as deep learning or deep neural networks or machine learning.
There are two phases to AI, DL and ML-the training and the inferencing. Training can happen inside a data center using standard hardware, whether that’s GPUs or ASICs. Inferencing needs to happen much closer to the edge, and it requires massive compute power in applications such as automotive because they have to be done almost in real-time. This will require blazing fast processing, but it also will require flexibility because AI and its subcategories today are so inaccurate that they need to be constantly modified.
There are a number of ways to achieve that flexibility. One involves programmability, whether that is in software (easier but slower) or hardware (eFPGAs, FPGAs and PLDs), which is faster but still not as fast as an ASIC. The second involves packaging, where the majority of a device is pre-designed and only a few of the components are swapped out to deal with future needs or protocol updates.
The second big change that is underway is the splintering of markets under the umbrella of the IoT. The growth in cloud computing (regardless of whether that involves internal or external clouds) has fostered much more differentiation at the edge. This is what enables the umbrella terms IoT and industrial IoT.
Tomi Engdahl says:
Control ICs Modulate Power in an Explicit Way
http://www.powerelectronics.com/pmics/control-ics-modulate-power-explicit-way?NL=ED-003&Issue=ED-003_20180307_ED-003_769&sfvc4enews=42&cl=article_2_b&utm_rid=CPG05000002750211&utm_campaign=15756&utm_medium=email&elq2=ca9286772cd2418bbd864a30d1109399
Power Integrations introduces a family of ICs designed to dynamically control the power and voltage of a power supply.
The InnoSwitch3-Pro family of ICs launched by Power Integrations at APEC 2018 this week simplifies the development and manufacturing of fully programmable power supplies. A basic I2C interface is used to configure, control and supervise operation of the power subsystem, enabling dynamic, adjustable control of output voltage (in 10-mV steps) and current (in 50-mA steps) when paired with a microcontroller or with inputs from the system CPU.
Tomi Engdahl says:
Executive Perspective: The Future Isn’t What It Used To Be
https://blog.globalfoundries.com/executive-perspective-future-isnt-used/
It may be tempting to view the strong demand for semiconductors as just one more up-cycle in our traditionally cyclical industry, but what’s really driving things right now is the opening of entirely new horizons, made possible by the increased capabilities of today’s chips.
Chip demand is no longer only being driven by the needs of computer and smartphone manufacturers. Now, a mushrooming number of new and varied applications within many different industries is both creating demand and pushing chip technology in new directions. Therefore, while the traditional goal of developing faster, denser semiconductors remains very important, it is no longer the only path forward.
We see the following sectors as major drivers of semiconductor demand going forward, in addition to traditional computing and smartphone applications: sophisticated Internet of Things (IoT) applications; 5G and wireless networking; automotive; and artificial intelligence/machine learning (AI/ML).
In the IoT area, a high degree of sensing, processing and communications capability is increasingly being embedded into physical objects to bring “intelligence” – powerful data analysis and processing capabilities – to their operation. The goals are to improve performance, efficiency and cost, and to develop entirely new ways of solving problems.
For 5G and wireless networking, bandwidth requirements are becoming incredibly stringent so as to create more capable, reliable and secure networks. For example, while achieving just 50 milliseconds of latency in networking equipment was an impressive technical achievement not that long ago, it now seems almost quaint because projected requirements call for latency of 1ms or less for many networking applications.
Tomi Engdahl says:
Wave Computing Chooses MIPS 64-bit RISC
https://www.eetimes.com/document.asp?doc_id=1333046
MIPS, a storied but beleagured RISC processor core company, is coming back to life. Breathing new life into MIPS are a new customer — Wave Computing — and a number of existing clients that include Intel/Mobileye, NetSpeed, Fungible, ThiCI and Denso. All have pledged to use MIPS 64-bit multi-threaded processor core to handle device management and control functions inside their respective AI processors — many either in development or ready for rollout.
Wave Computing is a designer of a massively parallel dataflow architecture called Wave Dataflow Processing Unit (DPU) for deep learning. Wave Computing, which is getting ready to roll out its beta system in the next few weeks by using the company’s first-generation processor, has decided to use MIPS 64-bit CPU for the company’s second-generation DPU, Derek Meyer, Wave Computing CEO and a veteran of MIPS, told EE Times.
Tomi Engdahl says:
32GT/s PCI Express Design Considerations
How to successfully design systems with the new PCIe 5.0 interface.
https://semiengineering.com/32gt-s-pci-express-design-considerations/
Today’s networking and rapidly emerging artificial intelligence (AI) applications are requiring more bandwidth in accelerators and GPUs, as well as faster interconnects to transmit and receive greater amounts of data.
Towards the middle of 2017 the PCI-SIG industry consortium announced its latest specification, PCIe 5.0, which raised the data rate from 16GT/s to 32GT/s and doubled the link bandwidth from 64GB/s to 128GB/s.
However, moving to 32GT/s designs comes with several challenges that both system and PHY designers must consider. This article describes such challenges and how designers can successfully design systems with the new PCIe 5.0 interface.
Tomi Engdahl says:
The Analog Design Gap
https://semiengineering.com/the-analog-design-gap/
Aggregating data is the basis of modern production systems and should be considered from system requirement to analog layout.
Sensors are everywhere. In the context of Industry 4.0 and IoT, we face an ever-increasing demand for high-quality sensing. Data acquisition is fundamental to adaptive production chains.
So aggregating data isn’t just some nice-to-have feature. It is the basis of modern production systems. But don’t we have sensors already? Isn’t everything fine?
When collecting data from the “real world,” there is no other way than integrating analog components into Systems-on-Chip (SoC). But digital components provide a noisy environment for sensitive analog components. In addition, the design flows of both analog and digital are very different.
The digital design flow, although being far from simple push-button, is highly automated with flow components that utilize complete automation. Analog design, on the other hand, is something of an art. Just like analog designers differ, their design approaches (and also verification approaches) do, as well. These differences can be seen, for example, in the design of testbenches, the design of schematics, the affinity for whether or not they use models, the approach of how parasitic effects from the layout are estimated, as well as in the way layouts are designed.
Tomi Engdahl says:
Intelligence At The Edge Is Transforming Our World
https://semiengineering.com/intelligence-at-the-edge-is-transforming-our-world/
https://semiengineering.com/too-many-nodes/
Machine learning already plays a part in everyday life, but efficient inference will keep it moving forward.
Tomi Engdahl says:
DRAM prices may rise 5-10% on datacenter demand in 1H18
https://www.digitimes.com/news/a20180307PD212.html
Bolstered by robust demand from datacenter and smartphone sectors, the global production value of the DRAM industry is estimated to surge over 30% to US$96 billion in 2018, and price quotes are expected to rise by 5-10% in the first half of the year, according to industry sources.
The sources said that the tight supply of memory chips for servers at datacenters actively under installation by US web giants since the second half of 2017 has eased slightly in the first quarter of 2018, but price quotes have lingered at high levels.
Tomi Engdahl says:
Power Electronics Products on Display at APEC 2018
http://www.powerelectronics.com/power-management/power-electronics-products-display-apec-2018?NL=ED-003&Issue=ED-003_20180308_ED-003_256&sfvc4enews=42&cl=article_2_b&utm_rid=CPG05000002750211&utm_campaign=15783&utm_medium=email&elq2=df5049844fa54dc69da5add9bf0cbe8a
Tomi Engdahl says:
The incredible multi-dimensional chess of Qualcomm vs. Broadcom
https://techcrunch.com/2018/03/10/qualcomm-vs-broadcom/?ncid=rss&utm_source=tcfbpage&utm_medium=feed&utm_campaign=Feed%3A+Techcrunch+%28TechCrunch%29&sr_share=facebook
Game of Thrones may be out of season, but the complex and multi-dimensional strategic drama at the heart of the acclaimed series can still be witnessed in today’s on-going showdown between Broadcom and Qualcomm.
Mega-mergers happen occasionally, and hostile takeovers are also not rare. What makes Qualcomm vs. Broadcom unique though is the incredible amount of chess playing that is taking place not only by the two companies, but other companies and governments as well in a simultaneous strategic game for tech domination.
Tomi Engdahl says:
Intel May Intervene in Broadcom’s Effort to Buy Qualcomm
https://www.wsj.com/articles/intel-considers-possible-bid-for-broadcom-1520633986
Intel, facing a threat, considers deals that could include a giant bid for Broadcom
Intel Corp. is considering a range of acquisition alternatives in reaction to Broadcom Ltd.’s hostile pursuit for Qualcomm Inc. that could include a bid for Broadcom, according to people familiar with the matter.
Intel is watching the takeover battle closely and is eager for Broadcom to fail as the combined company would pose a serious competitive threat,
Tomi Engdahl says:
Microchip inked an agreement to acquire Microsemi, provider of chips for defense and aerospace, for $68.78 per share in cash. The acquisition price represents a total equity value of about $8.35 billion and a total enterprise value of about $10.15 billion, according to Microchip. The deal is expected to close in the second quarter of 2018.
Source: https://semiengineering.com/the-week-in-review-design-120/
Tomi Engdahl says:
Cyclicality in the Age of IoT
https://www.eetimes.com/author.asp?section_id=40&doc_id=1333057
Could the rise of the Internet of Things and other new opportunities for electronics actually minimize the volatility of semiconductor industry cycles? At least one exec thinks so.
The semiconductor industry has been enjoying something of a renaissance as of late. After establishing a new high-water mark for revenue in 2017 — growing by more than 20% and passing the $400 billion mark for the first time — market watchers are predicting more market expansion for 2018.
Some parts — particularly memory chips — are in short supply, lead times are getting longer across the board, and chip suppliers have found amid boom times a swagger that seemed to have been missing for years.
What’s more, even as PC shipments continue to spiral downward, prospects for the future look bright, with evermore semiconductor content being designed into cars and technologies associated with artificial intelligence and the Internet of Things opening up whole new swaths of applications.
It’s the best of times, right? Except that anyone who has been around the semiconductor industry for any length of time is just kind of waiting for the other shoe to drop. Inevitably, healthy cynicism tells us, euphoria will lead chipmakers to expand production capacity too fast, lead times will shrink, market drivers will stall, and the industry will suddenly find itself in that dreaded state of overcapacity. The sunlight of prosperity will slip below the horizon, plunging the industry into another dark, dreary downturn.
Tomi Engdahl says:
Global Resistor Shortage, Economics, and Consumer Behavior
https://hackaday.com/2018/01/31/global-resistor-shortage-economics-and-consumer-behavior/
The passive component industry — the manufacturers who make the boring but vital resistors, capacitors, and diodes found in every single electronic device — is on the cusp of a shortage. You’ll always be able to buy a 220 Ω, 0805 resistor, but instead of buying two for a penny like you can today, you may only get one in the very near future.
Yageo, one of the largest manufacturers of surface mount (SMD) resistors and multilayer ceramic capacitors, announced in December they were not taking new chip resistor orders. Yageo was cutting production of cheap chip resistors to focus on higher-margin niche-market components for automotive, IoT, and other industrial uses, as reported by Digitimes. Earlier this month, Yaego resumed taking orders for chip resistors, but with 15-20% higher quotes
As a result, there are rumors of runs on passive components at the Shenzhen electronics market, and several tweets from members of the electronics community have said the price of some components have doubled. Because every electronic device uses these ‘jellybean’ parts, a decrease in supply or increase in price means some products won’t ship on time, margins will be lower, or prices on the newest electronic gadget will increase.
Learning a Lesson of Beef Consumption
News of a potential shortage of such a vital commodity like chip resistors and capacitors may come as a shock to some, and of course precedes the obvious question: should companies stock up on these jellybean parts? Should you lock in your prices now, and buy an entire year’s worth of inventory? What would happen if everyone did that? History shows us you shouldn’t.
The moral of the beef shortage story is to buy what you forseeably need, and not what you fear you’ll be unable to get. To do the latter in our current scenario will result in reels of components going unused on racks and in closets around the world instead of being available when needed.
Unnecessary stockpiling can be bad, but the electronics industry is also very weird. There is no other industry on Earth where one random dude can buy the world’s supply of something.
Tomi Engdahl says:
Trump blocks Broadcom’s takeover of Qualcomm
https://techcrunch.com/2018/03/12/trump-blocks-broadcoms-takeover-of-qualcomm/?utm_source=tcfbpage&sr_share=facebook
President Trump has blocked Broadcom’s proposed $117 billion buyout of Qualcomm over security concerns, according to a White House statement. News of the president’s decision was first reported by CNBC.
The move could send shockwaves rippling through the broader global economy, as the president continues his push to “put America first” in trade negotiations with global partners.
Trump administration’s block in Qualcomm vs. Broadcom merger should shake tech to its core
https://techcrunch.com/2018/03/12/trump-and-cfius/?ncid=rss&utm_source=tcfbpage&utm_medium=feed&utm_campaign=Feed%3A+Techcrunch+%28TechCrunch%29&utm_content=FaceBook&sr_share=facebook
Not only did the Trump administration move faster than expected to make a decision on the merger through CFIUS — the Committee on Foreign Investment in the United States (TechCrunch’s overview of the committee here) — but it decided to unilaterally block the transaction from taking place. While not unprecedented in the history of CFIUS, this is an incredible decision on a U.S. tech merger, and has massive ramifications for tech company valuations and strategy going forward.
While there are many issues at stake in the merger, the one that drove interest in Washington has been Qualcomm’s leadership role in 5G
Tomi Engdahl says:
Here Comes the Solid-State RF Energy Evolution
http://www.electronicdesign.com/power/here-comes-solid-state-rf-energy-evolution?NL=ED-003&Issue=ED-003_20180312_ED-003_469&sfvc4enews=42&cl=article_1_b&utm_rid=CPG05000002750211&utm_campaign=15833&utm_medium=email&elq2=47bd8c51af60417f862299025ae3f86f
Costs associated with GaN-based solid-state RF energy will trend downward, and in turn transform how we cook our food, illuminate our environments, treat our sick, and power our vehicles.
At the semiconductor level, the performance limitations of LDMOS have fueled the ascension of GaN-on-silicon (GaN-on-Si) to the forefront of the RF semiconductor industry.
The parallel evolution of solid-state RF energy technology and GaN-on-Si sets a clear path forward for OEMs competing for leadership in the aforementioned commercial markets. As these technologies continue to advance, market awareness continues to grow, and the underlying economics come into favorable alignment, mainstream adoption of solid-state RF energy technology is approaching quickly on the horizon.
Beyond the Magnetron
The ability to generate and amplify RF signals using solid-state semiconductor devices is nothing new—this technology is the cornerstone of modern wireless communications. But solid-state RF energy has enormous potential beyond data transmission applications. It’s increasingly being used for the purposes of heat and energy generation, enabling greater efficiency and control than what’s possible with the conventional magnetron tubes that—among other things—have been powering the microwave ovens in our homes for the past 50 years.
One of the major deficiencies of magnetron tube-based RF energy delivery is the inability to measure and adapt to energy that’s irradiated and reflected within the cavity where the energy is outputted. Magnetrons deliver open-loop, crudely-averaged energy output, whereas with multi-antenna, solid-state RF energy sources, forward and reflected power levels can be easily assessed and adapted to with closed-loop, precision control over the frequency, output power, phase and RF signal modulation (see figure). Solid-state RF transistors can also provide 10X longer lifespans than magnetrons, ensuring significantly higher reliability.
Here we’ll assess some of the identified markets where RF energy is primed for mainstream adoption in the coming years:
Solid-state Microwave Cooking
Solid-State Plasma Lightning, aka Light-Emitting Plasma (LEP)
Medical Care
Automotive Ignition
GaN-on-Si and LDMOS at a Crossroads
The performance benefits that GaN-on-Si delivers compared to LDMOS are well-understood. Where LDMOS compromises in power and frequency capability, GaN-on-Si demonstrates exceptional performance across both of these metrics, while offering several additional technical benefits. GaN delivers raw power density that’s considerably higher than LDMOS, with the ability to scale the device technology to high frequency.
GaN-on-Si is also distinguished by its high efficiency, providing up to 10 percentage points greater efficiency than LDMOS.
GaN-on-Si is expected to yield RF devices that are more cost-effective at scaled volume production levels than LDMOS in absolute $/W before even considering the benefits at a system level
A Coordinated Effort
As with any emerging technology, the speed of RF energy technology’s commercial adoption hinges in large part on collaborative industry efforts to establish common standards. In the solid-state RF energy technology domain, the RF Energy Alliance (RFEA) is leading this initiative with support from industry leaders spanning RF semiconductor vendors, commercial appliance OEMs, and beyond.
http://rfenergy.org/
Tomi Engdahl says:
Moore’s Law Is Dying — So Where Are Its Heirs?
https://www.forbes.com/sites/forbestechcouncil/2018/03/09/moores-law-is-dying-so-where-are-its-heirs/#51fa5e547a7b
Depending on whom you consult here at Forbes, Moore’s Law is either over (a polite way of saying “dead”), no longer holding up or alive and well (if you believe the company and its founder who invented the concept).
The impact of this plateau in performance will certainly be felt in areas like mobile devices, IoT, machine learning and artificial intelligence. More troubling is the impact on hyperscale data centers with massive aggregated compute power that manage cloud-based consumer and commercial products and services.
Most troubling of all is the impact on advanced research done via supercomputing (i.e., high-performance, high-density data processing) that drives progress in genomics, energy, national security, weather forecasting and modeling and big data analytics.
The forward march of progress and the health of the economy rely on continued improvement in semiconductors. I won’t bore you on why nanometer-size microelectronics are reaching their technological limits (let’s just say, “because physics”). It’s become too hard to make those already small circuits any smaller. As transistors shrank they become faster, but the wires now became thinner and slower. The solution is to address the problem from the first principle that will help solve the wire slow-down problem. This problem can be fixed by introducing a wire centric computational mechanism in a way where existing applications do not need to be rewritten.
The problem of device physics is, however, a solvable one, and that’s the silver lining to the death of Moore’s Law. The industry’s performance plateau creates a market space and opportunity for new ways of thinking, new designs and new inventions. We need radical, not incremental, change. We cannot afford to be bound by the dogma that has dictated traditional Silicon Valley product development (this dogma, by the way, is why the alternatives we’ve been promised are far behind schedule).
When it comes to using specialized processor systems such as TPUs and GPUs to save energy, we face headwinds from the growing non-recurring cost of designing chips that must be amortized against increasingly higher sales volume to maintain economic viability. A TPU or GPU can only run a small percentage of applications (algorithms) more efficiently than a CPU. In any computer architecture, it takes a lot more energy to fetch and schedule an instruction than it does to execute that instruction. A GPU has better efficiency than a more general CPU only if it’s mindlessly executing the same instruction on a larger set of data.
In this example, this means a GPU has a five-fold power and a five-fold area advantage over a standard CPU, but it comes at a price, because GPUs can be applied on only a very few problems, while the applications allow the GPU to repeat the same work of one instruction on 32 pieces of data at a time. Simple convolutional AI is an example where GPUs work well. However, a GPU approach has a much lower performance for general AI or symbolic AI, where mindless parallelism doesn’t work well.
Piecemeal advances won’t get the job done. We’re going to need 10-fold improvements in processing performance, and we’re going to need it at a third or half of the cost. Otherwise, U.S. companies may not be able to perform on the global stage. Trillions of dollars are riding on high-performance chips, and stagnation is not a risk we can take.
Tomi Engdahl says:
Why Inductance Is Good for Area, Power and Performance
https://semiengineering.com/why-inductance-is-good-for-area-power-and-performance/
Utilizing inductance rather than trying to suppress it can have a significant impact on leading-edge designs.
For chips designed at advanced technology nodes, interconnect is the dominant contributor towards delay, power consumption, and reliability.
Major interconnects such as clock trees, power distribution networks and wide buses play a significant role in chip failure mechanisms such as jitter, noise coupling, power distribution droops, and electro-migration. Buffers used in designing global interconnects to handle RC delay easily can contribute up to 60% of the total chip power. Therefore, robust interconnect design and modeling is critical for today’s leading-edge chips in order to meet performance and reliability targets.
For example, a designer may use differential switching on buses, where complementary signals are routed adjacent to each other. This effectively reduces the inductive coupling range and magnitude, because opposite currents close to each other produce opposite magnetic coupling fields that cancel each other out. However, it can consume four times more power and twice the delay compared to actively coupled wires with average switching.
Fundamentally, suppressing the inductive/magnetic effect can result in significant power and performance loss. While inductance is a reactive element that does not consume power, resistance is an active element that consumes power. In an interconnect network, inductance and resistance appear in series, so by suppressing inductance the resistive effect is amplified, resulting in greater power loss.
With current design trends, such as 10GHz+ clock / 10Gbps+ data line speeds, lower noise margins, tighter integration of analog/RF and digital blocks, and decreasing feature sizes, inductive/magnetic effects are becoming too difficult to suppress and can no longer be avoided. Ignoring signal integrity and crosstalk issues due to inductive coupling also can result in undetected reliability problems.
Designing with an understanding of inductive/magnetic effects, rather than suppressing them, can result in chips with better performance, lower power consumption and smaller area.
Tomi Engdahl says:
Handling the Heat: Materials Make the Difference
http://www.mwrf.com/materials/handling-heat-materials-make-difference?NL=MWRF-001&Issue=MWRF-001_20180312_MWRF-001_486&sfvc4enews=42&cl=article_1_b&utm_rid=CPG05000002750211&utm_campaign=15796&utm_medium=email&elq2=369c86cac660455cb182216d1f46548f
Ever-smaller electronic systems with increasingly greater functionality depend on efficient thermal management to maintain performance and reliability.
Thermal management can be an important step in achieving a long operating lifetime for the circuits and devices in a defense electronics system. Active devices such as power transistors and laser diodes are never 100% efficient, which means some amount of power applied to these devices will be converted to heat. The lower the efficiency, the more heat is generated for a given amount of applied power.
Tomi Engdahl says:
What PCB material do I need to use for RF?
https://www.edn.com/design/analog/4398951/What-PCB-material-do-I-need-to-use-for-RF-
Is plain old FR-4 (also known as “Glass Epoxy”) PCB material suitable for use in RF designs[1]? This question comes up time and again. Many say no, fewer say yes – who’s right?
As I have published before [2], I have been using FR-4 material for years to build not only protoboards, but wireless radios, RF test fixtures and RF test equipment. This is not to say that FR-4 does not have limitations, but when you understand the limitations you can make better cost/performance tradeoffs for all your designs.
So what are the limitations of FR-4? Er (dielectric constant) stability from lot to lot and over frequency is one of them [4]. Loss is another, then there is the concern of lead-free processing temperatures and perhaps thermal conductivity as even low-power RF can consume a lot of power if the active circuits are biased to provide very high linearity.
Since many FR-4 materials are not not really specified for RF performance the Er can and will vary from manufacturer to manufacturer and from lot to lot; sometimes Er is not even specified by some material suppliers! Does this all mean that FR-4 and “Glass Epoxy”-like materials can’t be used for RF?
Some of the above may be applicable to your project and some may not. Then there are the choices of material itself,
Plain old FR-4, with a higher loss and not tightly controlled Er[2]
Better specified Er[2] FR-4 derivatives (these may have better loss also)[3]
Specialized low-loss RF Materials with well specified Er values and much lower loss
It is a simple matter to go through the data sheets and make a spreadsheet that compares these items above one by one for comparison.
In the high-volume cases you will be hard pressed to find anyone using really exotic materials in the under 6-GHz world. Take apart all the items that I just mentioned and you will find materials that look just like regular old FR-4. In the low-volume but high-performance category you will find board material that again looks like FR-4 and you will find higher-frequency materials, especially when the operating frequency exceeds 6 GHz.
In the low-volume cases, performance may be paramount and the circuit designs might be more complex. Many of these products do use a tighter specified type of “Glass Epoxy” or exotic RF materials. Mainly for their repeatability and for the trace losses.
In addition to my standard FR-4 prototype boards (Figure 1) I also make quick turn prototypes on Rogers RO4350B material [5] (a low-loss, high-GHz material)
If you were building a 2.5-GHz Bluetooth module and the RF traces were about an inch long total – would you really care about a 0.3-dB signal loss, especially in light of the fact that the antenna matching circuit will probably exhibit more loss than this? Probably not. Even if you used Rogers RO4350B with its loss at 2.5 GHz of 0.13 dB/inch you would only be saving 0.17 dB.
Incrementally Improving on Plain Old FR-4
The first thing that can be done to improve on the generic FR-4 is to use a FR-4-like material that has a specified and controlled Er range. This material won’t be called FR-4 but will be made out of the same type of “Glass Epoxy” technology. PCB shops typically like these materials because they process with the same FR-4 manufacturing flow. There are many manufacturers that supply materials like this [3] and you should also make sure that the completed board uses material that can survive the high-temperature lead-free assembly processing temperatures, if you have this requirement also.
The Er of these materials is usually also much more stable with frequency than FR-4
High-Performance RF PCB Materials
The next step up in improving on FR-4 is to use a high-performance material like the mentioned Rogers RO4350B[5] and others. As can be seen in figure 2, the RO4350B PCB loss is less than half the loss of FR-4 at 6 GHz. While this may not be too important and not worth the extra cost if your circuit operates at less than 6 GHz, at 10 GHz the losses are even less and FR-4 really starts to show its weakness.
These really high-performance materials work well up to the 20-GHz-plus range and have a very stable and repeatable Er. The Er of these materials is also usually much lower, being on the order of 3.6. As with the higher-grade “Glass Epoxy” materials, the Er is essentially flat with frequency.
If your circuit design uses distributed elements or matching networks in the multi-GHz range then there is really no better choice than these types of materials for lot-to-lot consistency.
Another option to building a multilayer PCB with all high-performance material is to build a hybrid Glass Epoxy/high-performance material type board. This is where you use a material like the high-performance Rogers RO4350B on the outside layers – where the RF components and Microstrip traces are – and use a lower-cost Glass Epoxy inside where the power and control traces reside.
This Hybrid-type construction works out quite well and can save a substantial amount on your board costs.
So what’s the verdict?
Well I hope that I have shown that plain old FR-4 or improved Glass Epoxy can indeed be used at all the common RF/wireless frequencies up to 7 GHz or more. If plain old FR-4 won’t work for you for some reason, then you have the option of using a high-frequency, better specified “FR-4 like” Glass Epoxy material that won’t ruin the budget. If total loss and circuit stability are of paramount importance to you, or if you need to go above 10 GHz where FR-4 is really getting pretty lossy per inch, then you can always use the exotic high-performance microwave materials.
Currently the most popular wireless RF frequencies are around the 0.3 to 2.5-GHz range, and FR-4 will work just fine at those RF frequencies in many applications, especially when considered in light of the other variable parameters like board thickness, which has a greater effect on trace impedance than does even a widely varying Er.
Tomi Engdahl says:
Suppress slot antenna effect to diminish radiated EMI
https://www.edn.com/electronics-blogs/living-analog/4460378/Suppress-slot-antenna-effect-to-diminish-radiated-EMI
Way back when, I described how a blundering engineering manager sabotaged a project by ordering that a microwave assembly housing be constructed with only four screws to hold an aluminum case together with one screw in each corner instead of with a multiplicity of screws along each edge. The resulting slot antennas that were created along each seam between the box and its cover plate radiated like gang busters and the innards of that box could not function.
Tomi Engdahl says:
Less will be Moore, but it wont come cheap
https://www.electropages.com/2018/03/less-will-be-moore-but-wont-be-cheap/?utm_campaign=&utm_source=newsletter&utm_medium=email&utm_term=article&utm_content=Less+will+be+Moore%2C+but+it+wont+come+cheap
The road to 3nm may well be paved with good, groundbreaking technical intentions but there are some pretty deep financial potholes that need to be negotiated on the way.
So when news breaks that the industry’s first 3nm test chip tapeout has been achieved it is an attention grabber and certainly could perpetuate the scaling predictions of Moore’s Law.
But the harsh financial facts involved in turning 3nm into a feasible production reality make grim reading for the moneymen counting the beans at wafer fabrication companies.
Take the Taiwanese Semiconductor Manufacturing Corporation (TSMC) for example. It’s costing that company a cool $20 billion to build a manufacturing unit capable of producing 3nm devices.
The thing is that creating smaller chip geometries means cramming increasingly high numbers of transistors into shrinking spaces and there are only a handful of companies with wallets of sufficient size to cope with the costs involved with that. Amongst those are the likes of Samsung, Intel, TSMC IBM and STMicro.
But despite the cost, TSMC is pretty bullish about its 3nm prospects and expects to be in production in about four years. But back to that news about the industry’s first 3nm test chip tapeout. It is the result of collaboration between Cadence Design Systems and nanoelectronics research organisation Imec.
Tomi Engdahl says:
APEC 2018: Power Semis Focused on Tomorrow’s EVs
http://www.electronicdesign.com/automotive/apec-2018-power-semis-focused-tomorrow-s-evs?NL=ED-004&Issue=ED-004_20180313_ED-004_464&sfvc4enews=42&cl=article_2_b&utm_rid=CPG05000002750211&utm_campaign=15859&utm_medium=email&elq2=4861b5a13a6246a9a6f2d03bee7d90cf
WBG power semiconductors, including SiC power MOSFETs and GaN transistors, were out in full force at APEC 2018—but they are only the beginning of a steep adoption curve.
Tomi Engdahl says:
Power Management, Chapter 3: Power Supplies-Make or Buy?
http://www.powerelectronics.com/power-electronics-systems/power-management-chapter-3-power-supplies-make-or-buy?NL=ED-003&Issue=ED-003_20180312_ED-003_469&sfvc4enews=42&cl=article_2_b&utm_rid=CPG05000002750211&utm_campaign=15833&utm_medium=email&elq2=47bd8c51af60417f862299025ae3f86f
Power supplies are necessary in virtually every piece of electronic equipment. Therefore, equipment manufacturers are confronted with the task of deciding whether to make or buy a power supply for their system. DC power management employs a power supply that can either be bought or made by the equipment manufacturer. The make-or-buy decision for power supplies can have a major impact on the cost and time-to-market for the end-item electronic equipment.
The equipment manufacturer has several challenges to consider before making a power supply in-house:
• Can they make it cheaper than a purchased power supply?
• Is time-to-market a consideration?
• Are the necessary people and resources available to make the power supplies, including design and production facilities?
• Does the design and production include the time, costs, and fees associated with getting agency certifications specific to power supplies?
Unless the equipment manufacturer can meet these challenges, it most likely will buy the power supplies and then implement the power management subsystem.
Tomi Engdahl says:
Bloomberg:
Sources say Broadcom will formally abandon its attempt to acquire Qualcomm in an announcement on Wednesday — Broadcom Ltd. will formally abandon its attempt to acquire rival chipmaker Qualcomm Inc. after U.S. President Donald Trump blocked the deal citing national security risks, according to people familiar with its plans.
Broadcom Will Abandon Attempt to Acquire Qualcomm
https://www.bloomberg.com/news/articles/2018-03-14/broadcom-said-to-abandon-qualcomm-bid-on-government-opposition-jeqd4ss6
Tomi Engdahl says:
Ben Thompson / Stratechery:
Qualcomm-Broadcom merger threatened US security more because it would have dulled Qualcomm’s competitiveness beyond 5G than because Broadcom is a foreign firm
Qualcomm, National Security, and Patents
https://stratechery.com/2018/qualcomm-national-security-and-patents/
Tomi Engdahl says:
Analyst Lifts Chip Market Forecast to 15% Growth
https://www.eetimes.com/document.asp?doc_id=1333079
In a move reminiscent of the continual upward revisions to semiconductor market forecasts that characterized 2017, market watcher IC Insights has increased its forecast for semiconductor industry growth this year to 15 percent from a previous projection of 8 percent.
Just as strength in the memory chip market drove the broader chip market last year, the same trend is holding true early in 2018. IC Insights (Scottsdale, Ariz.) dramatically increased its forecast for DRAM sales growth to 37 percent from 13 percent and lifted its forecast for NAND sales growth to 17 percent from 10 percent.
The average selling price (ASP) of DRAM is now expected to be much stronger this year than originally forecast, IC Insights said. The firm expects DRAM ASPs to increase by 36 percent this year after growing by a whopping 81 percent last year.
NAND ASPs are now expected to increase by 10 percent this year, following a 45 percent increase last year, IC Insights said.
Tomi Engdahl says:
What to Expect at 5-nm-and-Beyond and What that Means for EDA
https://www.eetimes.com/author.asp?section_id=36&doc_id=1333075
With EUV finally on the verge of being inserted into volume manufacturing, let’s look at some innovations that have brought us to this point and what ultimately lies ahead for designing at 5nm and beyond.
Lithography scaling has long been the workhorse and enabler for the industry to track the path first proffered by Gordon Moore in the mid 1960’s — a doubling of transistor density every two years. The real thrust however being a halving of cost every node generation and it is this that has been the primary driver for innovation that has been delivered on to date. This cost and area scaling has demanded far more than lithography alone could deliver and it is a testament to the doggedness and engineering talent abound in our industry, that we have been able to maintain a trajectory along this path — albeit with a few bumps along the way.
With EUV finally making its appearance at the latest-generation 7-nm nodes and promising single-patterning benefits at 5-nm (though still likely double patterning at 3 nm without high-NA systems),
Tomi Engdahl says:
For A Successful Analog ASIC, First Weed Out the Pretenders
https://www.eeweb.com/profile/bob-frostholm/articles/for-a-successful-analog-asic-first-weed-out-the-pretenders
In my 45 year analog career I must have heard over one-hundred horror stories about failed attempts to develop analog ICs.
Tomi Engdahl says:
New brain for the Arduino, and…PICkit 4
https://www.edn.com/electronics-products/electronic-product-reviews/other/4460386/New-brain-for-the-Arduino–and-PICkit-4
Microchip has fair claim to the crown of Acquisition King, even ignoring the just announced Microsemi deal. But that’s for another article. Here are some interesting new products of theirs in ye Realm of Microcontrollers.
First off is the ATmega4809, of clan AVR. Heir apparent to the Arduino throne, it shares the basic characteristics of the somewhat-stuck-in-time AVR family. Not that this is necessarily a bad thing. Lower-end applications will always exist (and of course, Arduino is mainly aimed at the hobby crowd anyway), but much more powerful Arduino-compatible boards, official and unofficial, already exist. We can only speculate why Arduino chose to remain with eight-bit AVR. Comfort, 5V capability, a sweet deal from Microchip? But that’s for another article.
What sets the 4809 apart, if not the CPU (or memory complement for that matter (20MHz, 48k, 6k)), is the panoply of peripherals, including what Microchip calls core-independent peripherals, which can accomplish many functions without pesky CPU intervention.
The CCL is basically a teeny-tiny FPGA that can be connected to internal peripherals and external pins. I’m not sure it’s the answer to all of life’s problems, but it could certainly be useful, and seems to impinge on Cypress’ PSoC space a bit.
The ATmega4809 is yours for only USD 1.09 @5k.