At CDNLive Japan, Tom Beckley gave the keynote Enabling the Fourth Industrial Revolution (Industry 4.0) . At one point he announced Sigrity 2018, the latest version of the Sigrity power and signal integrity analysis tool, now merging in a lot of mechanical design and rigid-flex support to address the increasingly complex multi-board systems. Later in the day, Brad Griffin explained in a lot more detail just what Sigrity 2018 contains in a presentation Overcoming Simulation Challenges in Multiboard Systems . Tom Beckley In a view of our dystopian future when all our jobs have been taken over by robots, Tom Beckley's keynote was hijacked by Pepper the robot, who welcomed everyone to CDNLive Japan before eventually letting Tom onto the stage and meekly going and standing at the back while Tom talked about data and...well, robots. Even people outside the electronics industry have become aware that something is happening with big data and artificial intelligence. I'm sure you've seen plenty of charts showing in 2020 we'll have 4 billion connected people, with 25 billion embedded and intelligent systems, generating 50 trillion gigabytes of data. Or you can use a different set of numbers, nobody really knows, but they are all large. You may have heard that "data is the new oil". But just as a barrel of crude oil isn't very useful until it has gone through extensive processing, data isn't very useful either. You can argue precise dates, but the first industrial revolution started in the late 18th century in Northern England, using initially water power and then steam to automate the iron and textile industries. In the unlikely event you find yourself at Manchester Airport (in England, not New Hampshire), and with hours to spare, then I recommend a visit to Quarry Bank Mill nearby. It contains lots of original spinning and weaving equipment and you can see viscerally the impact on society. It has probably never crossed your mind that much as to why women and girls in fairy stories are always at spinning wheels, but that's because it took a lot of women to make enough thread for a single weaver. Women really did spend a lot of their time spinning since it was slow and not automated. The weavers loved it when spinning was automated, but weaving wasn't, because suddenly they could get enough thread to keep busy full time. They even loved the flying shuttle, which had been patented in 1733, since it allowed them to weave fabric faster (and wider than their arm span). However, the flying shuttle was also the missing ingredient to automating weaving, too. The term "Luddite" today is used as a sort of catchall for people against technology. They were actually weavers who believed the automated looms would put them out of business, which they did. The second industrial revolution was the era of mass production, with electricity and the internal combustion engine as the "driving" force. Oil, as the fuel for the automobiles, trucks, planes and ships was the lifeblood. If we call the third industrial revolution the computing and digital revolution, that unfolded over a long period of time, from the invention of the first computers, to mainframes and PCs. The internet. But what I like to think of as the big transformative event has a very precise date: on January 9th 2007 a man made a call to a Starbucks barista called Hannah Zhang and ordered 4,000 lattes to go. That man was, of course, Steve Jobs and that was the first iPhone call made (beyond internal testing). The smartphone changed everything. One thing it changed was the value of all the companies involved. The first and second industrial revolutions were dominated by oil (well, probably more coal in the first). In 1975, the top companies were all either oil companies, or oil-consuming companies: Exxon Mobil General Motors Ford Motor Texaco Mobil ChevronTexaco Gulf Oil Today, as we are start to go through the 4th industrial revolution, they are: Apple Alphabet Alibaba Microsoft Amazon Facebook Samsung Intel TSMC Softbank Data is certainly showing up as the new oil in stock market caps anyway. What is in our future? Robotics ("Hi Pepper"), AI, nanotechnology, quantum computing, IoT, autonomous cars, biotechnology. These are (mostly) about being able to take advantage of large amounts of data, and use it for training new classes of algorithms. We're not programming like it's 1999 anymore; we're training neural networks to do things we wouldn't even know how to program. One area that has made amazing progress is automated vehicles. The first DARPA grand challenge for this was only in 2004. See my post Ten Years Ago Self-Driving Cars Couldn't Go Ten Miles . And that was true, the furthest any vehicle got over the 120-mile course was an unimpressive 7.3 miles. But as I said in that piece: In 2005, things had improved by an almost unbelievable amount. There was a $2M prize and a new course involving 100 turns, three tunnels, and, at the end, a pass with a steep drop on one side and a cliff on the other. There were 23 finalists and all but one got further than the 7.3 miles of the previous year's winner. Five vehicles finished the entire 132-mile course. Industry 4.0 offers great promise and great peril for both IC and system companies (and probably lots of others, but let's stick to our industry). There is the fusion of sensors, electrical, mechanical, RF, software, along with, of course, chips. Machine learning is transforming how systems, especially ones involving vision, are assembled. The traditional "system" company boundaries are shattering (the automotive OEM tiers are becoming less relevant: Alphabet (Google) is building cars and isn't even in the car market (yet?)); Apple and Huawei design their own phone chips, they don't buy them from semiconductor companies; startups for smart everything. 5G is going to change a lot more boundaries. There is big opportunity, too. As Wally Rhines said in his keynote at the DARPA ERI event the following week, "many of these deep learning chip companies won't survive...but in the meantime they all need EDA software." But more than ever before, even a company like Cadence can't do everything and System Design Enablement (SDE) needs a lot of partners, different partners in different verticals. Before Pepper got back on stage to wrap up the keynote, Tom had some statistics on robots. The table below shows the number of robots per 100 workers. The most amazing statistic of all, though, is that since 2015 when Japan launched its "New Robot Strategy", they have now reached the stage that they are saving 25% of their labor with robots (which is behind Korea and Singapore). But they export 75% of the robots that they build, accounting for 11% of Japan's total exports. In 2022, robots are expected to reach a $250B opportunity, about half the size of the $500B smartphone market. Time for Pepper to make her own contribution to Japan exports by returning Tom's wallet. Brad Griffin One of the challenges with power and signal integrity, any form of full system analysis really, is that it doesn't respect technology and organizational boundaries. In a somewhat complex, but realistic case, a signal may have to go from a chip, through a complex package, across a board, into a cable connector, through a twisted pair, through another connection, across a board, into another package, in into the receiving chip. What I mean by not respecting technology boundaries is that those are a lot of different technologies: silicon, package, board, 3D connectors, cables. By organizational boundaries, I mean that those two chips may be designed by different design teams (or even different companies, especially for memories). The boards may be designed by different design teams. Probably the connectors and cables were simply purchased. Who takes responsibility for the integrity of the system? And who pulls together all the different analysis approaches to actually accomplish the task? You can fill in your own paragraph here about chips getting faster, packaging technology getting more complex, lower power...and shorter schedules with no more manpower. The answer to who pulls all the technologies together is Cadence with the Sigrity 2018 release and Sigrity 3D workbench. This brings together electrical and 3D mechanical structures into a system integrity flow. The pictures above give some idea of the complexity of the type of structure that needs to be analyzed. The GIF at the start of this section shows just how complex this can get with a complex multi-bit plug and socket. Brad had an example of a camera and its enclosure. There is one part of the chassis with the lens, one part with the cable out. The camera electronics consist of two boards. See above. There are connectors, a network interface, and I'm assuming a CMOS image sensor that consists of 2 or 3 very thinned die that are the same size are and attached (these days the light for a CMOS image sensor comes through the back of the thinned wafer, which sounds, when you first hear about it, like trying to watch your TV from the back). The sort of question that we want to analyze is how well the camera will function with ventilation holes. The challenge with system integrity is that everything affects everything else. If the holes are big, the design will be cooler, but EM might be unacceptable. Temperature is affected by power, but power is also affected by temperature, and signal integrity is affected by both. There is much more in the release, but the headline news is: Thermal, signal integrity, power integrity, ESD, EMI analysis Advanced accuracy simulation of multi-board systems using 3D models that merge connector and PCB Optimize PCB signal and power integrity, using system-level effects of cables and connectors Sign up for Sunday Brunch, the weekly Breakfast Bytes email![]()
↧