Bits&Chips 3 | 14 May 2021 by Bits&Chips

14 MAY 2021 | 2 5 JUNE 2021

Clearing the critical software path

BITS-CHIPS.NL

Tackling the MEMS testing bottleneck

NEXPERIA IS GROWING INTO A NEW SET OF CLOTHES

Falen is geen optie! In het dagelijks leven ben je afhankelijk van een heleboel verschillende producten. De auto waarin je rijdt, het vliegtuig waarmee je vliegt, of de ECG-apparatuur waarmee de activiteit van je hart wordt gemeten. Je verwacht dat alles goed functioneert – simpelweg omdat het moet. In alle elektronische producten zit een printplaat. Op het eerste gezicht lijken die allemaal op elkaar. Er zit echter een wereld van verschil tussen een normale printplaat en een High Reliability PCB. Het komt allemaal aan op de details, de nauwkeurigheid. Het begint met het ontwerp, de juiste specifi caties en het kiezen van de juiste productiepartner. Het omvat ook de logistiek, levernauwkeurigheid en het zoveel mogelijk verduurzamen van het gehele proces. High Reliability PCBs. Omdat falen geen optie is! PB

ncabgroup.com

pinion EDITORIAL Paul van Gerven is an editor at Bits&Chips.

Will Intel save Europe’s semiconductor ambitions?

urope and Intel – it would make for an interesting alliance. Both have been backed into a corner recently. For Europe, the blanket of the liberal international trade order doesn’t seem to be as comfortable as it used to be. Feeling the cold creep in, the European Commission, as well as France and Germany, have started pushing for ways to reduce Europe’s technological dependence. And Intel offered to help. But Intel itself is in a bit of a pickle, having dropped the ball in process technology. After leading the semiconductor industry for a decade or two, it has forfeited technological leadership to TSMC. Adding insult to injury, the once X86-dominated leading-edge semiconductor landscape has become a much more varied ecosystem. Arm, GPU and other architectures are slowly encroaching on Intel’s bread and butter. Faced with these existential threats, Intel was about to wave the white flag. Pressured by investors, it even hinted at moving out of manufacturing entirely. The new CEO, Pat Gelsinger, however, went on the offensive recently, unfolding a bold (some say foolish) new strategy to take Intel back to the top. This includes bolstering manufacturing operations and starting an independent foundry business offering not only X86 but also Arm and Risc-V cores and other IP. Given its struggles of late, the big question is whether Intel really is the godsend Europe has been hoping for. The company has to get its manufacturing woes sorted out, make up ground technologically, get customers to sign up and build up scale while at the same time keeping those fabs filled. And that’s while battling

TSMC, a formidable opponent that has executed flawlessly over the past years and that isn’t exactly going to sit on its hands. The biggest challenge by far, however, is successfully setting up a foundry. This will require a massive culture reset. Historically, Intel has never been good at anything else than doing one thing and doing it very well. That’s a long way from the mother of all customer-centric business models. On top of that, Intel’s

Here’s a crazy idea: what if Intel teams up with Samsung? recent track record likely doesn’t inspire confidence with potential customers. And how eager will they be to share their sensitive designs, especially since many compete with Intel in one way or another? There are also things working in Intel’s favor, of course. Many believe that the upcoming Silicon Everywhere era is more likely to create a semiconductor supply squeeze than a surplus. Governments are doling out a lot of money to anyone willing to build fabs in the West; for the US government, in fact, Intel may be too strategic to fail. And, surely, the Western fabless community wouldn’t mind if TSMC got some serious competition – Samsung is hanging in there, but it doesn’t challenge the Taiwanese (hey, here’s a crazy idea: what if Intel teams up with Samsung?). Technologically, even if Intel can’t catch up with TSMC in density any time soon, there are plenty of oppor-

tunities to compete on performance by introducing disruptive transistor and advanced packaging technologies. Intel now also has IBM’s innovation potential to draw upon: the two former rivals have forged an intimate R&D partnership. IBM may not be in manufacturing anymore, but it still has some excellent semiconductor R&D to offer – just ask Samsung. All in all, Europe would do well to not immediately give in to Intel’s charms. Better be patient and see how things progress – getting 7nm out without any further delays might be a good start. Europe may be desperate, but there’s no reason to be. Getting a leading-edge fab somehow has become the focal point of the technological sovereignty ambitions, but the fact of the matter is that Europe’s industries don’t even need that many leading-edge chips right now. It still makes much more sense to invest in semiconductor technologies that European (and other) industries will actually need. And in activities that will create demand for leading-edge semiconductors, such as IC design. That said, European demand for leading-edge chips is bound to increase, considering future applications such as automotive electronics, communications infrastructure, AI and Industry 4.0. Being able to manufacture them ‘in-house’ would be highly desirable, for which a leading-edge partner is a virtual sine qua non. As long as it’s complementary to evolving Europe’s own semiconductor base and semiconductor-using industries, and it’s looking like Intel might actually pull off the stunt of the century, it would be worth risking some taxpayers’ money to fund a European Intel fab. 3

CONTENTS IN THIS ISSUE OF BITS&CHIPS

News

Background

Nexperia is growing into a new set of clothes

Clearing the critical software path

Liberated from its cash-cow yoke, Nexperia has cranked up the ambition level, venturing out beyond standard products.

ASML, ESI and TUE have developed a modelbased methodology to diagnose, predict and optimize the timing of complex systems.

Will Intel save Europe’s semiconductor ambitions?

News

7 ASML’s sales to surge 30 percent this year 10 Nexperia is growing into a new set of clothes 12 Avular and Sorama team up to soothe the buzz of drones

Opinion 3

9 17 39 57 62 4

Will Intel save Europe’s semiconductor ambitions? – Paul van Gerven Partial de-globalization appears inevitable – Maarten Buijs The headhunter – Anton van Rossum Industry 4.0 is a means to an end, not an end in itself – Robert Howe Wi-Fi 6E: the new kid on the block – Cees Links The power of the why – Hans Odenthal 3

Avular and Sorama team up to soothe the buzz of drones

Background 14 20 26 30 32 36 40 44 47

Seven Dutch hopefuls in battery technology ASML reduces DUV overlay error to 1 nanometer Clearing the critical software path Creating software to keep naval systems always-on Prodrive streamlines software development with its own platform Machine learning isn’t hard with OpenML How to measure a planet Dutch Meteoriet consortium tackles MEMS testing bottleneck Hot topics in IC and electronic system testing – from all angles

ONLINE EVENT SERIES SESSION 3 & 4 27 MAY & 22 JUNE

Background

Dutch Meteoriet consortium tackles MEMS testing bottleneck ‘Neighboring’ companies and research institutes are developing an all-electric, universal MEMS testing solution suitable for all volumes.

ASML reduces DUV overlay error to 1 nanometer

50 Photonﬁrst ready to claim world leadership in integrated photonic sensing 52 Creating digital superhighways with optical wireless communication 54 Building the internet of the future 58 RF PCB designs – challenges, solutions and tips

SESSION 2 - 7 18 MAY, 15 JUNE, 31 AUGUST, 28 SEPTEMBER, 2 NOVEMBER & 30 NOVEMBER

LIVE EVENTS AFTER SUMMER

6 OCTOBER with online kick-off on 29 June BITS&CHIPS

BENELUX RF CONFERENCE

24 NOVEMBER with online kick-off on 2 September

Dates to be announced soon for

Interview

18 “Prepare to be frustrated and try to remember, it gets better” 60 “Sometimes we forget, soft skills can be the hardest skills of them all”

bits-chips.nl/events

simulation case study

How do you adapt the real world for electromagnetics simulations? When the ultimate goal is to design more efficient and productive electronic devices, design engineers need to run antenna measurements. If you know what attributes of the real world are important, you could instead test the designs with simulation. learn more comsol.blog/EM-simulations

The COMSOL Multiphysics® software is used for simulating designs, devices, and processes in all fields of engineering, manufacturing, and scientific research.

Credit: ASML

NEWS SEMICON

ASML’s sales to surge 30 percent this year Due to the chip shortage and exploding demand for 5G, AI and high-performance computing ICs, ASML’s sales might break the 18-billion-euro barrier this year. Paul van Gerven

lready the global chip shortage is driving up ASML’s revenue. The Veldhoven-based semiconductor equipment manufacturer reported Q1 sales of 4.4 billion euros, above the 3.9-4.1 billion euros guided. The main reason for this better-than-expected result, CEO Peter Wennink explained, is that chipmakers rushed to upgrade the software of their tools to increase productivity. “That’s the most effective and efficient way to increase wafer output,” Wennink said in a video posted on his company’s website. ASML’s Installed Base Management (IBM) business grew to 1.24 billion euros in Q1, up from 1.06 billion euros in Q4. For the remainder of the year, however, Wennink doesn’t expect IBM to grow faster than originally anticipated. “The software upgrades have been pulled in. What has been installed you don’t need to install the rest of the year.” Hardware upgrades would increase productivity even more, but they’re not very popular right now, since they require the tools to go offline. “In a time where you can’t make enough wafers, you don’t want to put the tool down. This is why we don’t see a lot of opportunity above the 10 percent that we guided before.” There’s plenty of opportunity for additional growth this year, however, and not just because of the current chip shortage. “Our view for 2021 has definitely changed since the beginning of the year. We’ve seen a significant spurt in terms of customer demand, driven by the digital transformation.

The roll-out of 5G across the globe, in Asia, the US and to a lesser extent in Europe, drives artificial intelligence and highperformance computing. All these developments and services and products are driving the need for our systems. That’s been a big change as compared to three months ago.” At the start of the year, ASML expected the logic business to grow 10 percent this year; now it’s guiding 30 percent. The prediction for memory has risen from 20 to 50 percent. The overall revenue growth expectation for 2021 has more than doubled to 30 percent – previously, it was at 12 percent. That would make for a revenue of 18 billion euros in 2021, which would mean that ASML would already make good on its 2018 ‘promise’ to grow revenue to between 15 and 24 billion euros in 2025. Last year, revenue came in at 14 billion euros.

Higher productivity

All that extra revenue won’t be coming from additional EUV sales: that number will stay at the previously guided 30 percent revenue increase. It could have been more, had it not been for customers putting orders on ice Q2 and Q3 last year, prompting ASML to tell its supply chain to slow down production. This cannot be undone now that the appetite for chips turns out to be bigger than ever. The integral lead time between the installation of an EUV tool and the start of module production is 20 months. ASML appears to have found a way to squeeze out a few more EUV tools next

year, however. Last quarter, Wennink told investors that his company would ship 50 systems in 2022. That number is now upped to 55. All these systems will be the NXE:3600D model, which will be introduced in the second half of this year. It features a 15-20 percent higher productivity than its predecessor, the NXE:3400C, which of course is also an excellent way for leading-edge chipmakers to increase their wafer output.

Capital inefﬁciency

Wennink, finally, subtly changed his tune about countries and regions looking for technological sovereignty, or at least reducing their technological dependence on foreign companies. Previously, he had warned that it would be near-impossible to replicate the current global semiconductor ecosystem in a single country or region and that it would increase cost. Only last week, speaking at the Hannover Messe, he argued for technological interdependence over technological sovereignty. Now, Wennink pointed out there’s an upside for ASML if the US or Europe or Asian countries start increasing semiconductor manufacturing capacity, on top of the huge amounts of money that the Big Three are already spending. “This will lead to higher capital intensity because it’s decoupling a worldwide ecosystem. But it also leads to some capital inefficiency. Well, there’s a beneficiary of that capital inefficiency, and that’s us.” 3

Keep it simple! Sense with MEMS.

WE are here for you! Join our free webinars on www.we-online.com/webinars MEMS Sensor Portfolio & Customer Service Sensors are an integral part of every future application. Measuring temperature, humidity, pressure or acceleration has never been easier. Take advantage of services like our Software Development Kit and Evaluation Boards available off-the-shelf. Detailed documentations as well as the direct support by trained engineers will leave no questions open. With excellent measuring accuracy and long-term stability, the sensors provide high precision and accurate output values with intelligent on-chip interrupt functions.

 Support by engineers within 24 h  Excellent measuring accuracy  Factory calibrated & ready to use  On-chip interrupt functions  Implemented algorithms

Combine sensors and wireless connectivity – start your IoT application today: www.we-online.com/sensors

3 Axis Acceleration Temperature Sensor IC WSEN-ITDS WSEN-TIDS

Absolute Pressure WSEN-PADS

Differential Pressure WSEN-PDUS

Humidity WSEN-HIDS

 SPI & I²C digital interfaces

#SensewithMEMS

pinion

SEMICON Maarten Buijs is CEO of Surfix.

Partial de-globalization appears inevitable

y granddaughter recently turned two years old. In the life ahead of her, I see two significant challenges looming: global warming and China being the dominant power in an interconnected world. When her father was her age, I was working as a postdoc in a group at Harvard with a number of graduate students of Chinese background: one from Hong Kong, one from mainland China and three from Taiwan. The father of the one from Hong Kong also had a domicile in British Columbia. Like many of his fellow Hongkongers, he acted on the expectation that the handover of Hong Kong in 1997 would ultimately lead to systematic integration. Events proved them right. The Tiananmen Square massacre happened while we lived in Boston. It shocked everyone, especially the significant Chinese student community. It certainly left nothing to be left to the imagination as to whether the system that produced the Cultural Revolution had changed. The colleague from mainland China, a very thoughtful and gentle scientist of middle age, was clearly scarred from that period. Within that framework, the developments in Tibet and Xinjiang are consistent. The three Taiwanese colleagues went back to their country and contributed each in their own way to the astounding transformation of Taiwan into a prosperous, open and democratic society. Part and parcel of that transformation was the establishment of TSMC (to which Philips had a significant contribution) and its growth to become the world’s largest semiconductor foundry.

TSMC was the first to commercialize ASML’s EUV technology in high volume. Now, since EUV is the dominant force for the future of semiconductors and thus the electronics industry, the US doesn’t want China to get access to this technology. Re-

What if China would invade Taiwan and take control of the world’s largest foundry? cently, they even started mulling the idea that immersion DUV lithography should be off-limits to China. China has been very clear that it will ensure that the future of Taiwan is identical to the one of Hong Kong: full integration. As it looks now, patience, increasing economic entanglement and increasing political pressure aren’t sufficient to achieve this goal within an acceptable time frame, whatever that may mean. Very recently, the US military openly announced that they assume China to use force to accelerate this development within six years. These kinds of announcements have their own purpose. But what if it does happen this way? What if we wake up to the news that China has invaded Taiwan and taken control of the country? The world economy would get another blow, the stock

markets would crash. But most commentators seem to think that no one is willing to go to war over Taiwan. So, we would end up with a situation where China takes possession of the largest semicon foundry in the world, filled with ASML EUV and DUV machines. Does this mean that one should also stop selling these machines to TSMC? Maybe one can encourage and entice TSMC to build more factories outside of China’s reach and fill these with ASML products. Moral principles, in the end, have precedence over commercial interests. This is, of course, an easy statement to make. Yet, a partial de-globalization of commerce and supply chains appears inevitable if we want to continue living in an open, tolerant and democratic society. Climate warming is a bigger existential concern to my granddaughter than which political system calls the shots across the globe, but I would hate to see our societal achievements disappear during her lifetime.

NEWS CHIPS

Nexperia is growing into a new set of clothes Liberated from its cash-cow yoke, Nexperia has been growing healthily since it was divested from NXP in 2017. Now, the semiconductor company has cranked up the ambition level, venturing out beyond standard products. Paul van Gerven

“W

hen I present an investment proposal to our CEO Xuezheng Zhang these days, I don’t just get an approval, I get asked: can you do more? Can you do this faster? Can you double it? In the 30 years I’ve been working in the semiconductor industry, it has never happened to me that I want to hire 15 new engineers, only to be suggested that 20 or 30 might be better,” says Dan Jensen, general manager of Nexperia’s business group Analog & Logic ICs (A&L). To be fair, Jensen adds, that go-getter mentality has been there ever since Nexperia was divested from NXP in 2017. Formerly a cash cow, the Standard Products division grabbed its newfound freedom with both hands and started to invest heavily. Last year, the construction of a brand-new 1.85-billion-dollar power semiconductor fab in China was announced. R&D expenditure went up from 7.1 percent of sales in 2017 to 9.0 percent in 2020 – and that’s with sales climbing from 1.1 billion dollars in 2017 to 1.4 billion dollars in 2019 (the 2020 financial results haven’t been published yet).

The current owner – Chinese semiconductor and communications company Wingtech Technology acquired a majority stake in Nexperia in 2019 – is taking it up another notch, however. Last February, an additional set of investments in R&D, people and manufacturing were made public. “We’re in the process of transforming the A&L organization,” says Jensen. “We have a history of over 50 years with standard logic parts, which by nature are very digital. We’re still very much focused on that, in fact, we aspire to become market leader over the next three to four years. But we want to grow faster and there’s a limit to what we can do in standard logic. That’s why we’re expanding into new territories.” And so Nijmegen-headquartered Nexperia is moving up the value chain. Its traditional offerings may be the ‘nuts and bolts’ of the electronics industry, within a couple of years, we can expect the company to start releasing increasingly sophisticated semiconductor products. Looking to monetize on trends such as the electrification of cars and the adoption of 5G, the company has gone in overdrive.

Sizable chunk

Jensen’s A&L group took its first steps beyond standard logic when it was still part of NXP. “We started investing in interface ICs, primarily voltage translators and level shifters. This is an extension out of logic, but a similar technology. We’ve released 50-60 new devices of this type over the last three years.” The investment has been paying off handsomely. “This business segment has shown a compound annual growth rate of 25-30 percent over the past 5 years. It’s now responsible for 20-25 percent of A&L’s sales.” A bit further along the road, A&L increased its focus on the analog switch market, another natural extension of its portfolio. “This could add as much as a billion dollars in addressable market.” And then, as a part of the investment push following the acquisition by Wingtech, A&L moved into the power management IC business. These devices are a bit more revolutionary, considering Nexperia’s historical bread and butter. The parts are still in development, though the first ones – a series of battery management ICs – were taped out a few weeks ago.

Credit: Nexperia

“This is a step-wise process. We’re starting with basic building blocks, such as gate drivers, load switches and e-fuses. These are basically similar to our existing devices with some additional intelligence wrapped around them. As we develop this kind of core IP, we’ll venture into more integrated and sophisticated solutions. The long-term growth prospects for these kinds of devices are very attractive: it’s an addressable market of more than 5 billion dollars,” says Jensen. To lay claim to a sizable chunk of that, A&L has hired 45 new engineers since the start of 2020 – two-thirds of the company’s total local personnel increase in that period. Most of the new hires are working at the Nijmegen headquarters, where most of A&L’s pre-manufacturing activities are stationed. The company has also opened a new design center in Penang, Malaysia, and is in the process of starting one in Shanghai, China. A fourth center in the US is being considered. “This speaks to the aggressive posture we’re taking about investing in the power management IC market. We’re ramping up and

hiring talent as fast as we can, in Nijmegen and around the world. That’s equally true for our other two business groups, by the way. We’re really turning into a company that’s focused on growth. In fact, under the Wingtech ownership, I’d say we’re pursuing a hyperaggressive growth trajectory.”

Fresh ideas

Jensen’s customers have hailed the course change enthusiastically. “I’ve literally had dozens of meetings explaining this new strategy, and every customer applauds it. They want us to move into these areas. They’re very much aware that, because of our history as a supplier of standard products, we know how to squeeze out costs. We typically have long-standing relations with them, so they’re familiar with our quality standards, the reliability of our supply chain and our customer support.” The new path is also changing the relationship with customers, Jensen explains. “To deepen our insight into the customer’s needs and the market in general, we’re also making big investments in what I call system engineers, which were traditional-

ly called product managers. They have an understanding of the market and the applications, and work directly with the customer base. After finding out what the key specs and parameters are, they work out the design definition. In the past, Nexperia invested in a very limited amount of that kind of skill set, but now we’re looking all over the world for it.” “Customers don’t just want us to make the parts, they want to work with us to get the best parts made. This kind of access to the customer is new to us. In the NXP days, it was much more difficult to get that kind of information, because of where the company was focused. We would typically be talking to purchasing, maybe to component engineering. Now, we’re having discussions at a much deeper level, with design architects and R&D leads. And the great thing about these meetings is that we came to talk about product A, but soon we’re talking about product B, C and D. We’re always walking away with fresh ideas. That’s why I’m very confident that our growth strategy will be very successful.” 3 11

NEWS DRONES

Avular and Sorama team up to soothe the buzz of drones From Amazon package deliveries to an unprecedented birds-eye view of ground operations, drones offer countless applications on which businesses can capitalize. But as the number of drones continues to explode, so too is the raucous sound of the blades piercing through the air. By taking a page out of nature’s book, sound imaging specialist Sorama and robotics expert Avular are teaming up to calm the shrill of the machines and limit noise pollution. Collin Arocho

ost people have a good understanding of the damaging effects that air pollution poses on the planet and the population. Somewhat less known, however, is that the second most damaging environmental impact on humans actually stems from noise pollution. There are a number of studies showing that psychoacoustic effects in the environment can have major effects on people’s health. “There are certain sounds that can really trigger the fight-orflight response that releases hormones in the human body,” says Rick Scholte, founder and CEO of the Eindhoven-based acoustic imaging expert Sorama. “Studies show that over time, high levels of these stress hormones can actually lead to a number of ailments – including heart disease.”

Insightful

However, according to Scholte, not all sounds have the same effect on people. In fact, many of the sounds of nature – the ocean, forests, or in his research, the hum of hummingbirds can actually be therapeutic. “What we find is that many of the sounds we encounter naturally in the world, don’t trigger the stress hormone response, rather they can have beneficial effects on stress levels,” explains Scholte. “It’s the man-made mechanical noises that we find ourselves surrounded by daily that have the most damaging effects. This is 12

Credit: Nexperia

Credit: Avular/Sorama

Sorama has developed systems that combine microphone arrays to record sounds and AI to process, analyze and understand how the noise is generated.

what we’re just starting to understand and has become the core business of Sorama as we aim to make sound insightful.” To gain this insight, the Eindhoven University of Technology (TUE) spinoff, has developed systems that use microphone arrays, ranging from just a few up to over 1024 microphones, to record sounds, which then uses AI to process, analyze and understand not only how the noise is generated but also the characteristics that make the sounds pleasant or stressful.

basis for modular robots that can be easily modified to fit in any application domain,” illustrates Yuri Steinbuch, COO of Avular. “Rather than creating application-specific robots or drones, we work with domain experts to show them how our robotics knowledge and modular robotic brains can be integrated into their systems for near immediate use, especially within our drone and driving platforms.”

Application

Credit: Sorama

While there are several applications for this technology, spanning smart-city solutions to the industrial imaging domain where Sorama’s technology currently helps detect gas and air leaks, one of the focuses of the Eindhoven startup has been in collaboration with drone and robotics expert and Strijp-T neighbor Avular. In its younger days, Avular gained a lot of experience in the drone market as a designer and builder of several application-specific drones, ranging from flying machines used in the agricultural sector to explosion and hazardous environments (ATEX). But in the last few years, the company has taken a pivot, focusing purely on robotics. “Our focus today has really become on creating a

Understanding how propellers ﬂap and how exactly that results in speciﬁc sounds can have a big impact on designing next-gen drones.

Collaboration

With Sorama’s insights into the effects of sound and how it’s generated, Avular is keen to find ways to improve its drone and robotics solutions. “To have additional awareness and understanding of how drone propellers flap and how exactly that results in specific sounds can have a big impact on designing next-gen drones, which can lessen the noise pollution. Of course, that’s a very interesting path for us at Avular,” says Steinbuch. “But we’re looking to achieve more than that. In our collaborative efforts with Sorama, we’re really looking into how to utilize the technology from both sides to create new and improve existing solutions – like creating mobile sound cameras that can fly or drive into different environments and give us those insights that Rick was referring to. Those can tell us a lot.” “Exactly. And we believe that it’s not simply about reducing the noise of drones or other mechanical sounds we encounter. Rather, our interest really lies in understanding the noise at a deeper level and finding ways to not just make it quieter but to make it more soothing and enjoyable,” highlights Scholte. “By mimicking nature, like the sound of the hummingbirds that can help reduce noise pollution and stress.” 3 13

B a c kg r o u n d

Batteries

Seven Dutch hopefuls in battery technology The Dutch high-tech ecosystem has sprouted seven companies that are looking to improve lithium-ion battery technology, or market completely different battery designs. Paul van Gerven

he battery has entered a golden age. It has already been indispensable for a range of applications, but with the anticipated transition to electric driving and increasing adoption of renewable energy sources, the world is rushing to increase battery manufacturing capacity. For example, global lithium-ion cell manufacturing is expected to rise from 95.3 GWh per year in 2020 to 410.5 GWh per year in 2024, according to market research firm Globaldata. The world will also need better batteries. Researchers and companies are frantically trying to improve the lithium-ion cells or come up with alternative technology that’s better suited for particular applications. After all, buffering supply and demand in the electrical grid is a different ball game than getting an electric car to drive as far as possible on a single charge. The Netherlands plays no role in battery manufacturing. There are several companies, such as Cleantron, SuperB and Eleo, that assemble battery modules for niche applications. VDL Nedcar is considering automotive battery pack assembly as part of its efforts to replace the impending loss of the BMW business. But there’s no battery cell or A-to-Z battery manufacturing in the Netherlands. On the technology front, however, seven Dutch companies have emerged that have something to contribute. The applications they target are remarkably varied, ranging from materials and components to full-fledged batteries and manufacturing equipment. 14

Battolyser

A spinoff from Delft University of Technology (TU Delft), Battolyser has developed, well, a battolyser. This device stores electricity like any battery, but it can also split water into hydrogen en oxygen when it’s fully charged. A simple combination of a regular battery and an electrolyzer could do the same, but the battolyser “does it better and at lower costs in situations where it really matters,” says its inventor, Fokko Mulder, professor at TU Delft’s Chemical Engineering department. The key is the ability to quickly react to electricity price fluctuations, which are expected to become more pronounced as more renewable energy sources are installed. While conventional electrolyzers can’t easily be turned on and off, the battolyser can instantly switch between hydrogen production and discharging the battery. So, when elecCredit: Holst Centre

Lionvolt’s 3D solidstate lithium-ion battery on foil.

tricity prices are low, the battolyser is put to work for producing (green) hydrogen, a valuable compound used in a range of (cleantech) applications. When prices are becoming too high, the device can not only discontinue hydrogen production, but it can actually start selling electricity by discharging the battery. Battolyser, headquartered in Schiedam, is backed by Koolen Industries, Proton Ventures and Delft Enterprises (which is part of TU Delft) and has been commissioned to install a device at Nuon’s Magnum power plant in Eemshaven.

Delft IMP

Batteries are just one application for the powder coating process developed at TU Delft and currently being commercialized by spinoff Delft Intensified Materials Production (Delft IMP). Coating cathode and anode materials enhances their durability, resulting

in extended battery life. Other advantages include higher energy density, increased battery safety and the option to use cheaper materials without performance loss. Delft IMP adapted the atomic layer deposition (ALD) process to give powder particles a ‘nanoshell’ of desired thickness. While traveling through a tubular reactor, the particles come into contact with gaseous precursors, which react with their surface. Thus, a coating is formed atomic layer by atomic layer, the thickness of which depends on the diameter and length of the reactor’s tubes. This process is continuous and capable of industrial production rates. The company sells the reactors, not the coated materials.

Elestor

Elestor’s mission is to build a storage system with the lowest possible storage costs per kWh. To accomplish that, the Arnhem-based company is building on technology that was developed by NASA decades ago: the redox flow battery. Lithium-ion cells are no match for these devices when it comes to grid-scale energy storage, on account of their scalability, superior lifetime and cost of ownership. In a nutshell, these redox flow batteries ‘store’ electrons in a chemical compound, which is synthesized whenever there are electrons (ie electricity) in excess. The electrons are released when needed through the same chemical reaction in reverse. Elestor’s flow battery turns hydrogen bromide into hydrogen and bromine when charging. These active materials are readily available, cheap and enable both a high energy and power density. Supported by the European InnoEnergy fund, Elestor was founded in 2014. The Arnhem-based company later attracted investments from Koolen Industries and Enfuro Ventures.

E-magy

The focus of Broek-op-Langedijk-based E-magy is on lithium-ion battery anodes. These are traditionally made from graphite, even though it has been known for a long time that silicon would be vastly superior in terms of energy density. The reason is that silicon can’t handle the mechanical stress associated with repeatedly taking in and letting go of lithium ions. E-magy’s parent company, RGS Development, created a casting process that results in ‘nanosponge’ silicon, which doesn’t crack under the pressure of entertaining

Electron microscope image of E-magy’s nanoporous silicon.

Credit: E-magy

guests. This material can achieve a 40 percent higher energy density than traditional graphite anodes while shortening charging times and reducing the cost of production. Focusing on the electric-vehicle market, E-magy currently has one production facility operational and is making preparations to expand. The company is targeting to churn out 3,000 tons of nanoporous silicon annually, enough to supply up to half a million EVs each year.

Leydenjar

Like E-magy, Leydenjar has set its sights on silicon anodes for lithium-ion batteries. Unlike E-magy, however, it’s not selling the material but the manufacturing equipment. The company’s core technology was originally developed by PV research institute ECN in hopes of obtaining better-performing solar cells. The nanotextured silicon, produced using a plasma-enhanced chemical vapor deposition (PECVD) process, didn’t do well in that particular application but, as it turns out, makes for fine anodes. Leydenjar claims an increase in energy density of up to 70 percent compared to graphite anodes. The Leiden-headquartered company initially spent a lot of time to prove the real-world advantages of its anode manufacturing process, as well as its commercial viability. Next, it established a pilot production line in Eindhoven, allowing for customers to put in a sample order to get a taste of the technology. The final step will be to develop roll-to-roll deposition equipment optimized for production.

Lionvolt

Lionvolt wants to become the first Dutch manufacturer of battery cells – but not just any battery cells. The Holst Centre spinout has successfully demonstrated a proof of concept of a 3D solid-state thin-film

lithium-ion battery. It consists of a foil, covered with an array of micropillars, each coated with thin layers of battery materials: lithium-storing electrodes sandwiching an electrolyte. For simplicity’s sake, each pillar can be considered a tiny battery. This design has three main advantages. One: the lithium ions need to travel relatively short distances, translating into faster charging and de-charging times. Two: there’s no liquid electrolyte involved, meaning a longer lifespan and little to no danger of fires or explosions. And three: the design is inherently lightweight. Leveraging these advantages, Lionvolt will initially target the wearables market. In the longer term, larger versions of the 3D solid-state batteries will be developed for automotive and other markets.

SALD

Eindhoven-based SALD has a lot of similarities with Delft IMP. Both companies are developing ALD production equipment for a variety of applications. When used for contemporary lithium-ion battery manufacturing, the tools of both companies are used to apply a protective coating to the electrode materials. The processes involved are quite different, however. SALD’s core technology is another ALD variety capable of high-throughput production, called spatial ALD. It involves leading a moving substrate past different precursor-gas zones, thus step by step building up a nanolayer. Unlike Delft IMP, SALD applies the coating after depositing the powdered electrode materials on a substrate. This means that the particle surfaces are not covered entirely. For example, where particles touch, no coating can grow. According to SALD, this improves the ease with which electrons can move through the electrode. 3 15

‘Even if your opinion contradicts your boss or management, they want you to speak up.’ Joining the Dutch workforce can be fraught with challenges, especially when coming from another country. While some cultural norms are easy to notice, learn and understand, others can be a little shocking or even frustrating for those with green behind their ears. In the Dutch work culture, it’s often that you need to look no further than communication. According to ASML design engineer Marco Allegri, who joined the Dutch work culture from Italy, the transition can be a little frustrating. But, he says, once you learn where it stems from, you’ll learn to appreciate the typical Dutch communication style. hightechinstitute.nl/dutch-culture

pinion

THE HEADHUNTER Anton van Rossum anton.van.rossum@ir-search.nl

Ask the headhunter B.T. asks: I’ve been working as an analog IC designer in the Netherlands for almost two years now. I did my bachelor’s at a university in Tamil Nadu, India, where I was born. I then worked for about four years as a PCB and layout engineer, before I had saved enough to move to Delft and do my master’s there. After completion, I started in my current position. Recently, my manager told me my department will be closed in six months. The activities are relocated to Arizona, where my company is now recruiting plenty of staff. Of course, I think it’s a shame. I have great colleagues and I’ve learned a lot here. On the other hand, I know that there’s always work to be found for my specialism and I’m open to anything. When I applied for a position at a large chip company in Eindhoven, I soon received a phone call from an HR employee, inviting me for a screening. Initially, he was very positive about a follow-up video interview, but the next day, he called me to cancel. Under the existing knowledge migrant legislation, they would be obliged to give me a higher salary than allowed by their classification based on education and experience. I didn’t take it completely seriously at first, but it turns out to be the case. According to the applicable rules, because I only have two years of work experience after my mas-

ter’s and turned thirty this year, I’d have to earn 4,752 euros gross per month with a new employer in the Netherlands, which is 61,586 euros per year. This salary is considerably higher than what I get now. According to the HR employee who brought it up, this could make it very difficult for me to find a job in the Netherlands. For most companies, I’m ‘too expensive’ compared to other engineers with my experience. Can you help me find an employer for whom the salary limit isn’t a problem, or can you offer an alternative solution?

The headhunter answers: The Netherlands enforces an agedependent salary criterion for highly skilled migrants: a minimum of 45,153 euros under the age of 30 and 61,586 euros above that. This criterion is a nuisance to you and many others. You’ll be subject to it as long as you continue to work for the same employer – even after you turn 30. When moving to another company, your age at that time applies. That’s what the HR employee saw on the IND website and what he based his conclusion on. What he’s overlooking, however, is that there’s an exception to this rule for highly skilled migrants who qualify for the reduced salary criterion. This continues to apply, even if a migrant subsequently moves to another

employer. The reduced salary criterion isn’t age-dependent; it applies if you meet the condition for the “orientation year for highly educated persons” and if you apply for a residence

There’s an exception to the rule: the reduced salary criterion permit for work as a highly skilled migrant within three years after your graduation or promotion date or the date on which the residence permit for scientific research expired. The reduced salary criterion also applies if you didn’t get a residence permit for the orientation year for highly educated persons, but do meet the conditions for that purpose of residence. Because you graduated less than three years ago, you’re eligible. If I were you, I’d contact that HR employee again – maybe you can still get an interview for the position.

3 17

INTERVIEW MARCO ALLEGRI (ASML)

“PREPARE TO BE FRUSTRATED AND TRY TO REMEMBER, IT GETS BETTER” For even the most culture-savvy expats, Dutch directness inside the workplace can serve as a bit of a shock. According to ASML design engineer Marco Allegri, who joined the Dutch work culture from Italy, the transition can be a little frustrating. But, he says, once you learn where it stems from, you’ll learn to appreciate the typical Dutch communication style. Collin Arocho

oining the Dutch workforce can be fraught with challenges, especially when coming from another country. While some cultural norms are easy to notice, learn and understand, others can be a little shocking or even frustrating for those with green behind their ears. In the Dutch work culture, it’s often that you need to look no further than communication. Not so much in terms of language-ability barriers, as the Dutch are extremely talented in a number of languages, but in their style of communication – where the “Dutch way” can feel a little, well, ouch. “Working in the Netherlands has been a relatively smooth transition for me. ASML has gone out of its way to provide me and other expat employees with all the necessary help, resources and a number of onboarding activities to feel part of the team from the very start,” explains Marco Allegri, a mechanical design engineer who joined the Dutch semiconductor equipment giant after moving to Belgium from Italy. But, despite his positive start with the company, even he has to admit: there are certainly some cultural differences. “Compared to my previous job in Italy, I’ve noticed that the Dutch workplace has a very no-nonsense 18

approach to work, with extreme attention to process, procedures and details, which was all a little new to me. I’ve also found that this downto-business approach you find in the Netherlands can often result in communication or feedback that’s both instant and rather harsh.” Can you recall a specific moment when you experienced this? “Oh yes, definitely. It was the first time I received direct feedback from my previous team leader. We were in a meeting having a discussion, when suddenly he cut me off, almost mid-sentence, in complete opposition to what I was just saying. He totally disagreed,” recalls Allegri. “Let’s just say, this wasn’t something I was used to, and I didn’t dare to try to respond or argue. What would I even say?” In Italy, according to Allegri, 90 percent of the time, people probably wouldn’t speak up in direct opposition. And, if they did, it would have been full of niceties and politeness. “You’d take small steps and ask if you could add something, or mention that you had another perspective to offer, but never would you do it in such an immediate and direct manner,” expresses Allegri. “In Dutch culture, on the other hand, this is

an expectation. Even if your opinion contradicts your boss or management, they want you to speak up – you just need to be sure you have supporting facts and evidence. That’s what drives people here. In Italy, it’s very hierarchical. Even if the boss is wrong, he’s right – because he says he’s right and he’s the boss. There’s not really room for discussion and it would never be so direct.” Giving feedback This experience served as a real eye-opener for Allegri. As he continued to grow within his role and the company, he saw this sort of communication style being used by nearly all his colleagues, especially those that were Dutch. “At first, you know, it’s really a bit of a shock. But that’s how it’s done here, and I’ve really come to enjoy it. It’s this style of direct communication that gives me a clear understanding of where things stand, what has to be done and how to achieve it,” remarks Allegri. “It’s never personal, it’s always facts first. When you have data to back up your opinion, you can be sure that the people here are open and will actually hear what you have to say. That’s kind of a new idea for me.” However, for Allegri, there was still a real challenge to this sort of

communication. In his experience and with his cultural background, giving this type of feedback was no simple task. That’s when he registered with High Tech Institute for the training: “How to be successful in the Dutch high-tech work culture.” “This training provided us with a really good theoretical overview on why the Dutch communicate in this manner. By far the most impactful information I received though, was in learning to provide this sort of direct feedback as well as how to deal with the vast number of stakeholders in meetings and in our day-today work,” highlights Allegri. “The most helpful aspect of the training was learning how to structure my feedback, being sure to kick the ball, not the man – so to speak. Also, the techniques for dealing with disagreements between or influencing stakeholders and creating buy-in from a position without power. This was really enlightening and again put a real emphasis on using facts, data and figures to support ideas – that’s central to Dutch-style communicating. I especially found the exercises and scenarios that put the theory into practice to be useful. I wish we could have done even more because that’s something I’m still implementing in my work today.”

Credit: Marco Allegri

What advice would you give to other expats that are looking to work in the Netherlands? “Sometimes, the Dutch struggle to put themselves in your shoes. You have to remember that they’ve grown up being integrated into the ‘Dutch way,’ which I’ve come to really appreciate and even favor. But sometimes, they lack perspective from the other side,” illustrates Allegri. “So, my advice to other expats coming to work in the Dutch high-tech is rather simple. Prepare to be frustrated. Prepare yourself for tones that will seem harsh and procedures that will seem endless. But also try to remember, it gets much better. That’s just the way things are done here, and they have a very strong track record.” 3 19

B a c kg r o u n d

Lithography

ASML reduces DUV overlay error to 1 nanometer In a balancing game between precision and productivity, ASML has increased the throughput to 295 wafers per hour, while reaching an overlay error milestone of 1 nanometer in its new Twinscan NXT:2050i. The ﬁrst 15 systems have already been shipped.

Credit: ASML

René Raaijmakers

SML’s engineers continue to improve the overlay in its most advanced immersion scanners. Since the introduction of the NXT:1950i platform in 2010, the overlay error dropped from 3 nanometers to 1.4 nanometers in the NXT:2000i two years ago. At the same time, the throughput increased from 175 to 275 wafers per hour. “As always in our industry, we needed to do better,” says Bart Paarhuis, who presented the latest overlay results on ASML’s most recent DUV system, the Twinscan NXT:2050i, 20

at the online Bits&Chips High Tech Systems 2021 conference. “Customers requested additional improvements. The logic manufacturers want better DUV-to-EUV matching for their 3-nanometer node and our memory customers want better overlay when patterning successive layers on the same chip by the same DUV scanner. And everyone is asking for higher productivity.” To meet these requests, ASML introduced the NXT:2050i, with a new thermally improved wafer handler, a more accurate wafer stage, a more wear-resistant wafer

table, an improved light source, a new immersion hood for lower defectivity and a 2050 projection lens with reduced overlay fingerprint as main features. The reticle stage now has reduced distortion of the reticle during clamping. Also, many software improvements were made. For the development of this new immersion scanner, ASML collaborated with various teams and suppliers on the wafer handler (VDL), the wafer stage (VDL and Kyocera), the wafer table (Berliner Glass), the production of the immersion hood

Credit: ASML

One key task of a lithography scanner is the ability to image dense lines and spaces from the mask to the wafer. Here, the critical dimension (the width of the line), the pitch (the distance between the lines) and the straightness of each of the lines are important – shown in green. The horizontal placement error between the layers also impacts the quality of a chip. The structures that are impacted by overlay errors are shown in red.

(AAE), the reticle stage (ASML’s Wilton factory) and the projection lens (Zeiss). Cymer and Gigaphoton from Japan delivered a pulse stretcher on the lasers that reduced the speckle, resulting in much lower scanner-induced line-width roughness. ASML’s software team took care of the required software intensity around measurement, analysis and control.

Relay race

Going down to the molecular level of a lithographic machine, we don’t see a rigid structure but a pudding, a coherent set of elastically moving parts. Lenses, frame, stage, sensors and wafers – at the nano level, everything is acting like rubber. These microscopic dynamics make the flawless production of nanochips increasingly difficult. Printing ever-smaller details is already a challenge;

Credit: ASML

placing billions of minuscule parts where they’re needed is even more demanding. The correct and predictable stacking of, say, sixty different layers is an absolute necessity for a properly functioning chip. This comes together in the concept of overlay – the accuracy with which a lithographic scanner can place billions of structures in a layer on top of the billions of elements in the previous layer. “If you make an overlay error, you have less contact area between features in the layers. As a result, you get more resistance, and when you get more resistance, this results in less speed and more energy consumption,” explains Paarhuis. The better the overlay, the better the yield and chip performance. Paarhuis: “As customers currently split layers into multiple exposures, the overlay performance is even more critical.”

In the past, all layers were exposed in optical scanners with deep ultraviolet (DUV) laser light. Currently, part of these layers is exposed in scanners that use extreme ultraviolet light (EUV). ASML claims its EUV equipment will keep Moore’s Law alive beyond the next decade. This continued scaling puts quite some pressure on the demands for overlay. In the most advanced fabs, the first layers on bare silicon wafers are structured by EUV lithography. Subsequent layers are printed by DUV immersion scanners, followed by less critical layers that are produced by DUV dry and older-generation machines. The sequence of layering is a relay race where handing over the baton needs attention, especially the step between EUV and DUV (the switch between litho technologies is called cross-matching). “We need The dedicated chuck overlay (DCO) is one of several tests ASML uses to measure overlay performance. Using a special overlay reticle with 7x7 measurement patterns, the scanner prints 49 marks in the ﬁrst layer of one exposure ﬁeld. These marks can be read out later using the alignment sensor. This is repeated for the subsequent layer. With the help of the alignment sensor, the positions in layer one and layer two are compared, which results in a plot of the overlay errors as vectors from the measured locations. The graph shows maximum overlay errors of 0.8 and 0.7 nm for champion DCO wafers on the NXT:2050. 3 21

Sponsors

SOFTWARECENTRICSYSTEMS.COM #BCSC2

B a c kg r o u n d

Software engineering

Creating software to keep naval systems always-on Imagine a system that, once turned on, will stay on for the rest of its life. Neither hardware failures nor code glitches can bring it down. At Thales, chief software architect René van Hees is putting his shoulders to the wheel to make this dream a reality. Nieke Roos

Credit: Thales

“C

ompanies like Tesla have set a new standard in software development. With a mere update over the air, they’re able to cut 6 meters off their cars’ braking distance at 100 km/h. That’s really something – we’re talking about millions of vehicles that have to remain safe to drive,” says René van Hees from Thales. “Who’d have thought five years ago that we’d be changing the functionality of our critical systems on the fly?” Chief software architect Van Hees is pushing hard to adopt a similar approach for naval systems. “We’re becoming much more software-intensive as well, with software playing an increasingly important role in our radars. New features are more and more enabled by software, while the hardware remains more or less the same. We’re heading towards a future where, once we’ve deployed a system on a ship, we’ll be adding functionality exclusively through software updates.”

Design for change

To get to that future, there are still some choppy waters to navigate. “For one, it means that we need to drastically increase the frequency with which we deploy new software versions to our systems in the field,” clarifies Van Hees. “Instead of do30

ing a big bang every six months or so, we need to move to very frequent, very small, very local updates. Within our software development department, we’re already integrating continuously, ie adding new

features to our software baseline every day. We’re looking to roll this out to the system level and eventually to the customer.” The main challenge is to keep the systems qualified at all times. After a system has

passed all operational tests, customers tend to become very wary of even the smallest modifications. Which is perfectly understandable considering that some of the radars are critically connected to heavy weaponry. So, at the very least, Thales needs to be able to give the absolute guarantee that adding a new feature will allow operations to continue. Van Hees: “Such a task is far from trivial for software that’s so dependent on hardware. Design for change is key.” The software engineering team in Hengelo is exploring several means to this end. “We’re looking into the communication between the different software components,” gives Van Hees as an example. “If we update one such component and one of its interfaces changes as a result, this might affect the components on the other side of that interface, in turn affecting the components they’re talking to, and so on. One little modification could thus cause a tidal wave flooding the system. Building on the Comma framework developed in collaboration with Philips and ESI, we’re working on ways to keep the changes local.” Another research project addresses the security considerations of software updates in the field. “Wherever the system is, we need to make sure that it only gets the updates we want it to,” explains Van Hees. “We’re exploring ways to enable the system to verify that Thales is indeed Credit: Thales

the source and if so, to automatically and securely install the new software version while remaining operational and qualified.”

Software ﬂexibility

All these efforts, and more, come together in Van Hees’ BHAG, big hairy audacious goal – building systems that are always-on. “My vision is to have a system that, once turned on, will stay on for the rest of its life. Hardware parts may break or go obsolete, software components may crash or be corrupted – nothing can bring it down. To get there, however, we have our work cut out for us. It has all kinds of ramifications, for software development as well as system design.” Hardware is bound to go obsolete, for example, so at some point, a new processing board or module would have to be put in. “Since the system is always-on, I’d need to add the hardware at run time,” Van Hees points out. “It’s then up to the software architecture to manage this and keep the system up and running.” The system also needs to be prepared for new features and the associated added complexity – like the Tesla update reducing the car’s braking distance. Van Hees: “I could deal with that by expanding the hardware, but I could also be more flexible in the software I’m running on my existing hardware. Based on the circumstances, I could redistribute my processing power from features that are less es-

sential at a given moment to functionality that’s critical. Maybe I don’t need to be able to see 2,000 kilometers all the time and, for a short period, 150 kilometers is far enough. During that specific window, I could then allocate more resources to more demanding tasks.” The same mechanism could be used to fence off cyberattacks, Van Hees goes on to philosophize. “For this to work, my security would need to be completely implemented in software, though, which isn’t yet the case. Periodic updates, design for change and software-defined security all presume that I can reason about my solution and, moreover, coordinate the required changes.”

Opening up

This vision of always-on systems, together with the software strategy to get there, is one of three pillars supporting Van Hees’ work as a chief software architect at Thales. “But I’m going nowhere without skilled people,” he notes, introducing the second pillar. “As good software architects are extremely hard to get by, we’ve decided to train them ourselves. With Luminis, we’ve set up the Accelerate program: in 18 months, select mid-level software engineers have both their technical and human skills cranked up to senior level. This pressure cooker is expected to deliver its first batch of twelve technical leaders come November.” Van Hees’ third pillar is collaboration. “We simply can’t do everything ourselves so we need to work together. Collaboration in education, as illustrated by our Accelerate partnership with Luminis, which was born out of a mutual need for senior people – I’d like nothing more than to open up the program to other companies. But also, collaboration in sharing knowledge and technology. Take Inaetics for example, our dynamic, service-oriented software architecture. It was developed in a publicly funded consortium. We’re now looking to forge similar partnerships with companies to open-source it and broaden its application where feasible.” “Our systems no longer operate on their own,” Van Hees concludes. “As the world around them is growing more complex, so are they. The move to systems that are always-on only adds to the complexity. Keeping them resilient, also over time, is our main challenge.” 3 31

B a c kg r o u n d

Software engineering

Prodrive streamlines software development with its own platform means Prodrive had to build up very specific knowledge. So specific, in fact, that it pushed itself into a niche. “Four years ago, we started building up our own IP. We’ve invested heavily in a catalog full of off-the-shelf products. After all, with a good collection of standard drives and motors, we can develop and deliver mechatronic systems much easier and faster. We still do custom work, but since most of the building blocks are already available, we can move more quickly.”

With the in-house built PMP platform, Prodrive Technologies has succeeded in its efforts to make sure software development is no longer a bottleneck. By modeling and analyzing at an early stage, the dynamic behavior of the physical system can be predicted very accurately. Various industrial parties have already recognized the advantages of PMP and are using it for their own motion applications. Alexander Pil

leap of faith, that’s how Ralph Stassen describes the step Prodrive Technologies took a few years ago in its development strategy. “We started out as an electronics specialist,” recalls the Motion & Mechatronics program manager at the Eindhoven-based company. “We added custom software and we’ve also been doing custom drives and mechatronics for some time

now.” Within Stassen’s group, half of the turnover currently comes from the development of specialized mechatronic systems. Prodrive is very technology-driven by nature. That sounds good, but it has a downside as well. Stassen explains: “Customers come to us with a problem and we try to solve it with the technical capabilities we have in-house.” That much customization

For everyone

To keep up the high development pace, Prodrive also had to make its software designers work more efficiently. “In the seventeen years that I’ve been working here, we’ve set up several motion applications,” says Maik van Kranenburg, a software engineer at

Credit: Prodrive

“Because it’s all done virtually, you can iterate very quickly,” says Milan van den Muyzenberg.

posal and discussed where there was room for improvement. This is how we create a platform that’s interesting for everyone.”

Faster iterations

It turns out that the wishes of all users, both internal and external, show considerable overlap. “With highly software-driven engineering groups, the differences can be overcome quite well,” says Stassen. “Those engineers want to start from scratch with as much freedom as possible so that they can build the most efficient system. We offer that openness with PMP. On the other side of the table are the mechatronic engineers who just want to control their engines and don’t want to program at all. They’re looking for a practical interface with which they can

Analyze early

PMP has been mature for a few years now. Multiple Prodrive customers are already using the tool internally. The Eindhoven-based company isn’t allowed to disclose much about these projects, but it can go into some detail about the development process for a

Credit: Prodrive

Prodrive. “These were always iterations of existing software, but about seven years ago, we concluded that it was too difficult to maintain. That’s when we started to build a new platform so we could easily configure our software. Based on a hardware model and with input about which sensors and actuators are to be included in the system, we can now start developing software via the Prodrive Motion Platform before a physical design is built. Today, we’re really reaping the benefits. Previously, the pressure might have been on the software, but now, with PMP, it has shifted back to the firmware and hardware. The platform enables us to have the software up and running in no time.” Although PMP was originally designed to make internal software development more efficient, Prodrive quickly saw that its clientele could also benefit from the tool. Therefore, it didn’t set up PMP under the radar but invited a number of its largest customers. “It was quite nerve-racking to ask them all for input at the same time in one big meeting,” admits Stassen, “but it was very much appreciated. We asked them how we should design the software, which functionalities are important and let them participate in the decision-making process on the roadmap. This way, we’re working on a generic platform that precisely matches the wishes of our high-tech customers. It’s an ongoing process in which we still meet regularly to make adjustments.” Van Kranenburg gives an example: “One of the requests was about data acquisition. Initially, you could start and stop it manually, but there was a need to be able to base it on triggers. If a certain value gets too high, you want to record all signals one second before and one second after. We made a pro-

conveniently tune their drives. This is also possible with PMP, partly because we integrate with Matlab and Simulink. Through a toolbox and PMP, you can upload the models you make in those tools into the controllers of your system.” Van Kranenburg adds: “We used to sit around the table with a customer to discuss the model, after which we would implement it manually. When the customer would then integrate it into his system, he would sometimes run into problems and we had to come over to solve the issues together. At that time, we didn’t have a model that sufficiently corresponded with reality so that we could develop and test our software against it.” PMP put an end to the separate models for the mechanics and the software. “A customer can model his system himself in Simulink, turn it into a library and upload it. We no longer have to interfere, thus immediately avoiding IP discussions. A customer can do his own thing and no longer has to tell us the details of his machine. You can also simulate in Simulink, in PMP and on the final hardware. By comparing the results, you can significantly shorten the iteration time.”

Prodrive originally wanted to make internal software development more efﬁcient, but its customers also beneﬁt from the PMP platform. 3 33

B a c kg r o u n d

Software engineering Ralph Stassen: “We’re working on making PMP even more accessible.”

Credit: Prodrive

manufacturer of PCB production equipment. Prodrive designs and builds all the mechanics for that client, which it delivers turnkey. The customer only needs to add his own application module on top. “We started with a clean sheet and the customer’s requirements,” describes Prodrive system architect Milan van den Muyzenberg. “Some features of the machine are crucial to the customer. For example, the settling behavior determines how successful the system will be in the market. You want to secure these kinds of things in advance. We immediately started with a model in PMP so that we could properly analyze the system. In such a model, you can inject interference forces and see what influence they have on the dynamic behavior. You can also assess the stiffness by checking that the eigenfrequencies are what they should be. This knowledge can then be linked back to the properties that are relevant to the customer. As this is all done virtually, you can iterate very quickly, internally and also with the customer.” With PMP, Prodrive engineers can build, compile and run the necessary models. “You can simulate in Simulink and load the resulting code into our simulator,” explains Van Kranenburg. “This way, we can already perform the machine’s critical movements in a virtual setting. We can test the dynamic behavior in PMP and compare it with the customer’s model.”

“It’s a close collaboration between the control engineer and the mechanical engineer,” Van den Muyzenberg continues. “Together, they ensure that the system behaves as intended. As the process evolves and we have more hardware available, we can test the software developed on the real hardware. We’re currently still in the phase of deploying the

models on a PC, but we can switch almost seamlessly. After all, we use the same tooling and software that will later run on the actual system.” Prodrive has provided a simplified model of the mechanics to the customer. “They are, of course, more interested in the core of their machine than in the mechanical behavior,” notes Van den Muyzenberg. “But in the end, everything has to be integrated. The model allows them to assemble and test their entire system at an early stage.”

With the Prodrive Motion Platform, you can start software development before a physical design is built.

Credit: Prodrive

Staying ahead

Where can PMP still be improved? “We’re already very close to first time right,” states Van den Muyzenberg. “The mechanics match the model so well that we don’t expect any problems in machine dynamics. Because we’re talking about expensive, complex parts with very long lead times, your first shot must be right. Modeling in PMP has helped us enormously.” Stassen: “The platform’s user-friendliness can be improved a bit further. We’re working on making PMP even more accessible so that you don’t have to be a software engineer to get the most out of it. We’re targeting the high-end market where the wishes and requirements are very challenging. We need to improve constantly to be able to keep up and stay ahead.”

ONLINE EVENT SERIES Four online afternoon sessions with in-depth presentations and virtual networking Two more sessions to go: 27 May and 22 June Previous sessions: 23 March and 29 April

P R OGRAM 27 M AY

Sponsor

Expert sessions by:

Sander Bohté (CWI)

Gernot Eggen & Patrick Bronneberg (Philips IGT Systems)

Buy a full event ticket to get access to the presentations of all sessions (live and afterwards) or register for one or more separate sessions.

MLCON.NL

#BCML

B a c kg r o u n d

Artificial intelligence

Machine learning isn’t hard with OpenML Developed in the heart of the Brainport region, boasting more than 150 thousand users worldwide and yet, the Open Machine Learning platform is largely unknown in the local industry. With the support of initiator Joaquin Vanschoren, Georgo Angelis from TUE’s High Tech Systems Center and Eindhoven AI Systems Institute wants to change that with his startup PortML. Alexander Pil

hen Joaquin Vanschoren started developing the Open Machine Learning platform about six years ago, it was out of need and out of frustration. As a researcher at the KU Leuven, he kept running into the same walls when he wanted to use machine learning techniques. “How can I get access to many datasets? And how can I properly compare different machine learning algorithms?” describes Vanschoren some of his daily obstacles. “The challenge was – and still is – that most datasets aren’t accessible or, at least, require weeks of work before they’re useful. Moreover, what’s published in research papers is often very difficult to reproduce, if it’s even possible at all. Especially when there’s a commercial company behind it, they contain a lot of marketing. When you try for yourself, it often doesn’t work.” Vanschoren started the OpenML platform as an open-source project because his ambition was too big for one person to achieve. His initiative was quickly picked up by the research community and now about twenty people are contributing to the tool. “Mostly volunteers,” says Vanschoren, currently an associate professor at Eindhoven University of Technology. “Initially, they were predominantly PhDs who, like myself, were struggling with the same challenges but now, more and more people from the industry are getting involved.” 36

Credit: TU Eindhoven

Joaquin Vanschoren from OpenML Foundation: “OpenML is a gathering place for all machine learning research worldwide.”

Credit: TU Eindhoven

Georgo Angelis, High Tech Systems Center: “Within PortML, we’re working on a version of OpenML for the industry.”

Credit: OpenML

SMEs will beneﬁt from the machine learning recipes developed by Michelin chefs.

The basic idea behind OpenML is that it should be an open platform where datasets are easily available and where you can find algorithms that are relevant to your problem. “An accessible interface to all machine learning research,” summarizes Vanschoren. At the moment, OpenML serves a community of about 150 thousand users worldwide. Understandably, a similar tool wasn’t available. “Commercial parties have little interest in transparency. They rather hold their cards close to their chest. However, they can benefit from having such a platform for internal use – something we realized quickly during the development. Large companies like Amazon have their own tools, of course, but for most companies and organizations, it’s unfeasible to do it themselves.”

version,” tells Angelis. “Currently, we’re in the pilot phase with several companies to really understand which features are required from an industrial point of view.” “Openness is pivotal for OpenML, but companies want to safeguard their data,” Vanschoren adds. “PortML tries to find the middle ground by building a platform that combines the advantages of access to the

latest research with the requirements from industry.”

Michelin chefs

Angelis expects the first beta version of OpenML for industrial users to become available this quarter. “Data scientists are specialists who use every tool they can find to optimize their machine learning flow,” An-

How does OpenML work?

OpenML is an online and open platform for machine learning that consists of thousands of datasets, algorithms and tasks. “The core of the platform is formed by more than 21 thousand, well-annotated datasets,” says Georgo Angelis from High Tech Systems Center. “That’s a seriously big pool of data directly available for all kinds of machine learning experiments.” Users can upload new datasets themselves, making sure the OpenML platform keeps growing continuously. The second axis in OpenML are the tasks, for instance, for classification, regression or clustering. “You can use the results others have shared to get to an algorithm for your problem,” Angelis explains. “Imagine you want to distinguish apples from pears. Then you go and find the right dataset, look for a similar classification task and see what algorithms others have found. You’ll get an overview of which recipes scored

Enterprise version

It’s precisely that last point that triggered Georgo Angelis from TUE’s High Tech Systems Center and the Eindhoven AI Systems Institute (EAISI) to start his own company, PortML, in collaboration with the OpenML Foundation and the Eindhoven university. “In the academic world, OpenML has many users but in industry, the platform is still largely unknown,” says Angelis. Although the tool is available for anyone, and companies could start right away, there’s some reluctance. “That’s understandable considering the open character of OpenML. The uploaded datasets, the models, the algorithms – it’s all public. Commercial companies aren’t too keen on sharing these kinds of data with everyone.” Speaking with potential industrial users, Angelis notices that many are interested but that they indeed aren’t happy with releasing all their valuable data. “Within PortML, and supported by the OpenML community, we’re working on an enterprise

Credit: OpenML

OpenML shows which algorithms performed best.

best. You can try those on your own dataset and maybe even improve them. Finally, you share the new recipe in the OpenML database.” Joaquin Vanschoren from the OpenML Foundation: “OpenML is a gathering place for all machine learning research worldwide. You’ll find problems similar to your own and see what solutions work best.” A luxury problem for OpenML is that it contains so many datasets and tasks that it’s difficult to see the forest for the trees. “Largely, we’ve solved this with an advanced search functionality but we also use machine learning to better organize all data.” This last step is essential because sometimes it’s not enough to cluster datasets and tasks around topics like healthcare, sports or industry. An algorithm working splendidly to map the stars might also be the best solution for recognizing skin conditions. “We tackle that machine learning challenge with machine learning”, says Vanschoren.

3 37

B a c kg r o u n d

Artificial intelligence

gelis points out. “They can approach OpenML from their toolset through an API. They can keep using their own environment – whether it’s Python, or R, or any other suite – and use the stand-alone version of OpenML to organize all data, algorithms and models, and make them suitable for reuse.” Does OpenML also make sense for smaller companies that more often than not lack an in-house data scientist? Angelis definitely thinks so: “Bigger companies can use OpenML to improve the efficiency of their process, increase the quality and automate several steps. For smaller companies, OpenML lowers the threshold to start with machine learning. With some mouse clicks, they have access to what’s already available and reach a sufficiently good solution. That result may not be next-level, but it will surely be a big step in the right direction. They’ll benefit from the recipes created by Michelin chefs.” Vanschoren: “When you want to build a machine learning model, you need to take an incredible amount of decisions. Which algorithms, which models, which parameters, to name a few. Currently, you need a PhD – or at least someone with a lot of experience in machine learning – to create efficient models. Because there’s so much data and metadata available in OpenML, we can learn from ourselves. We use machine learning to decide what will work and

OpenML is an open platform to easily share datasets, algorithms and models. Credit: OpenML

what won’t.” That’s the research field of automatic machine learning, or AutoML, focusing on good search algorithms that find the best solution for a given dataset. “The resulting solution may not be a panacea but for many small and medium-sized companies, it can be very insightful to experience what they can do with machine learning, and it will give them a perfect starting point to build on.” Angelis is looking emphatically for collaboration with the outside world. The

companies interested in the pilot phase are involved in healthcare, manufacturing and mobility. “We at HTSC/EAISI want to contribute to the power of the Brainport region. So obviously, we’ve started with our own network in the high-tech,” he explains. “Later, we plan to expand to, for instance, telecom or finance. To me, this article is a pitch for the industry. We want to work together with companies to get to the most optimal version of OpenML for that target audience.”

Leader in Customer Value VECTOR NETWORK ANALYZERS

VALUE =

Uncompromising performance. Advanced software features on Windows or Linux OS without added cost or licensing fees. Timely support from automation and applications engineers. Comprehensive VNA solution from Copper Mountain Technologies.

Need more information?

VNAs to 44 GHz and Frequency Extension to 110 GHz from Copper Mountain Technologies

+31 40 2507 451 tm-nl@acalbfi.nl

pinion

INDUSTRIAL AUTOMATION Robert Howe is an independent management consultant.

Industry 4.0 is a means to an end, not an end in itself

ndustry 4.0 is a subject that’s much hyped these days. Governments are financially encouraging their manufacturing industries to develop Industry 4.0 programs. Business strategists, such as McKinsey and Deloitte, publish a constant stream of articles on its value and importance. And the fact that most definitions of Industry 4.0 are based on other hyper-sexy topics such as AI, IoT, cloud, additive manufacturing and 5G only goes to further bolster its significance. But down at ground level, many people who run manufacturing businesses frequently don’t understand what Industry 4.0 is about. Too often I’ve heard Industry 4.0 referred to as “meaningless container terminology” because it’s usually defined through technologies that make it possible. This is counterproductive and alienates the very people that Industry 4.0 needs to target: the owners of manufacturing business challenges. This audience doesn’t think from a technology perspective. It doesn’t think “If I had AI, what could I do with it?” It thinks in terms of either “How can I make more widgets in less time, with less waste and less cost?” or “How can I better meet the needs of my customers?” or both. I was recently introduced to a company that I was told wanted to integrate the machines they manufacture into an IoT platform. Upon meeting them, I carefully listened to their ideas for this platform, which essentially focused on “how.” They had extensive ideas about adding sensors, collecting data and dump-

ing it into the cloud. Ultimately, quite expensive ideas. When I asked them about the “why” behind their ideas, they more or less dried up. They hadn’t thought about the value they hoped to create with their IoT platform, nor how to

Any Industry 4.0 program not deeply grounded in “why” is doomed to failure monetize it. They hadn’t thought about how to take the idea to market, how their customers might react. They hadn’t stood in the shoes of their customers and considered what they need. More or less their motivation was based on the dazzling promise of exciting technology. And when one stands back and looks at the definition of Industry 4.0, one cannot blame them as the whole thing centers on “how.” This is strange, really, as any Industry 4.0 program that’s not deeply grounded in “why” is doomed to failure. For the owners of manufacturing business challenges, it’s much more meaningful to define Industry 4.0 in terms of the practical benefits it delivers. Getting back to basics, Industry 4.0 is relevant to a business to the extent that it enables that business to create more value. In other words, Industry 4.0 is a means to an

end and not an end in itself, that end being defined by the vision and mission of a business. The true power of Industry 4.0 is its ability to deliver better efficiency, flexibility, sustainability, customer value and financial performance. In each of these areas, it provides rich possibilities for value creation. The technology of Industry 4.0 is – indeed – a means to this end. There’s one caveat to this line of reasoning. Industry 4.0 technology makes possible things that a business might not have considered when framing its mission. Therefore, understanding the possibilities can provide business owners with new insights into how to realize their business goals. So, what is Industry 4.0? Fundamentally, it’s a collection of technologies that, when applied to a manufacturing business, enable a leap in the performance and value creation capability of that business.

3 39

B a c kg r o u n d

Industrial automation

How to measure a planet Angelo Hulshout has the ambition to bring the beneﬁts of production agility to the market and set up a new business around that. His journey is taking a small detour. Angelo Hulshout

n 1998, a not so obscure metal band from ’s-Hertogenbosch released an album called “How to measure a planet.” The band was called The Gathering, and this album was a break away from their previous albums, which were filled with rather bombastic heavy metal. Daring, but they did it – and they were successful. With a startup, it’s no different. At some point, you may have to deviate from your original plan to become successful. I remember the story of an ex-colleague at Philips who started his own company in compiler optimization. He ended up doing completely different things, in the area of static code analysis. The core technology he used stayed the same, but he wound up in a completely different market. You don’t make a switch like this from one day to the next. It requires a bit, or a lot, of trial and error to get there. This is what we’re facing with Shinchoku right now, shortly after we started and even before the first version of our website is live.

Feet on the ground

My original idea is still alive: help small and medium-sized manufacturers optimize 40

their production processes in a data-driven manner. Or in order words, make them part of Smart Industry, Industry 4.0. However, we now know that for this market, the world looks a little bit different than all the fancy Industry 4.0 videos show. There are very few robots, there’s still a lot of manual labor and although we live in the age of digitization, a lot of factories have very little production management software running. That isn’t necessarily a bad thing, but it does require us to define which steps we want to take to realize our original goal. We’re not by default going in to install an internet gateway, hook it up to the PLCs of all production machines and then feed the data acquired to a dashboard. That was a bit of a short-circuited version of our story anyway – you can’t just collect random data and be successful, so we’d have to analyze with the customer what data to collect before taking off. That analysis has put us with our feet firmly on the ground now: in most factories we’ve seen so far, there’s very little data to collect initially. We have to work with the customer, with the operators in the factory, to identify which parts of the process

are candidates for improvement. This has to be done based on anecdotal information and experience because there’s only limited automation and reporting support. Once identified, we need to establish what can be done to (semi)automatically collect and arrange data before we can start working on our actual goal.

Small steps

At one factory, for example, about 60-70 percent of the production process turned out to be based on manual labor. Not only are the machines operated by humans, but the metal parts produced by those machines are also polished by hand. The problem in this factory is that nobody at any point during the production has an overview of how many parts for a certain order are being processed and where they are. In a place like this, going from manual labor to full automation in one big bang is nothing more than a dream. The reality is that we have to take small steps. The first step, in this case, could very well be installing wireless foot pedals with certain workstations, which the operators can simply press every time they finish working on a

single part. Combined with tracking the trays transporting the metal parts through the factory (using RFID or similar), that would help us get a better view of how many parts are produced and in which production stage they are. Obviously, this is lightyears away from just installing a few network cables and a gateway and collecting data that’s already there. We discussed our experiences with both customers and automation specialists from our side of the industry. They all confirm that some important steps need to be taken first if we want Industry 4.0 to come within

reach of the smaller production companies as well. Not everybody is as big as L’Oreal or Volkswagen, where millions can be spent each year to improve automation and human labor in production has been largely eliminated long ago.

Process analysis

In keeping with the song from The Gathering, we need to find out how to measure a planet before we can actually go ahead and measure it. That’s what we’re doing now. Realizing our Industry 4.0 vision requires some big steps, the first of which need to be

taken in our market of small and mediumsized manufacturing companies. So, almost from the start, Shinchoku will not only be a data analysis company, but it will also have a lot of work to do in process analysis and automation – almost literally to be able to generate its own data. Angelo Hulshout is an experienced independent software craftsman and a member of the Brainport High Tech Software Cluster. Edited by Nieke Roos

3 41

CHILDREN IN UTRECHT SEEK SUPPORTING TECH COMPANIES If you think about technology regions, you might not immediately think of Utrecht. Utrecht is famous for its beautiful city center. It’s also the central railroad hub in the Netherlands, which is why many major events are organized in the Jaarbeurs. And... from now on, Utrecht has a real “Hall of Discoveries,” to inspire children and youngsters to pursue a future in tech!

he tech situation: alarming Employers in the construction and technology industries suffer a major shortage of workers. For realizing the climate goals alone, an extra capacity of 70,000 people in the Netherlands is needed in the short term. While the city of Utrecht has high ambitions in this area, its technical education outflow is extremely low: in 2019, only 17(!) VMBO students (preparatory secondary vocational education) chose technology as their major. The solution: Hall of Discoveries In response to this situation, leading companies (including BAM, Heijmans and Eneco), together with the municipality of Utrecht, have built the “Hall of Discoveries” (De Ontdekhal). With the proven concept of “The Inventors” as the cornerstone, children can discover if they have

affinity and talent for technology and innovation. On top of the Inventors programs and workshops in the Hall, all schools (primary and secondary education) in Utrecht are offered ‘adventure programs’ for appealing tech education at their school locations. The challenge: your company’s support The responsible foundation, TEC Utrecht, already has a great partner network. However, for a sustainable continuation of the Hall of Discoveries and expansion of the activities, many more partners are needed. Only with the help of tech companies themselves, enough children will be inspired to choose for a future in their business area. Is your company one of those? Send an email to info@ontdekhal.nl and discover how this great initiative for all children will be beneficial for you as well!

The Inventors programs are there to inspire youngsters for a future in design and technology. Projects are supported by tech companies such as ASML, Brainport Industries, DAF Trucks, Frencken Europe, Hager, High Tech Campus Eindhoven, NTS Group, Philips, Prodrive, Stam en De Koning and VDL Group, and by Bits&Chips as the media partner.

discoveryfactory.nl

B a c kg r o u n d

Testing

Dutch Meteoriet con sortium tackles MEMS testing bottleneck Currently, MEMS testing basically comes in two ﬂavors: highly automated but tailored to a very speciﬁc device that’s manufactured in high volumes, and, essentially, manual inspection for everything else. A consortium of ‘neighboring’ companies and research institutes aims to bridge that gap by developing an all-electric, universal MEMS testing solution suitable for all volumes. Paul van Gerven

no mistake, though: large manufacturers stand to benefit, too, as their testing time per unit can be reduced. All in all, the new testing methods will remove a major bottleneck for the adoption and application of MEMS technology.

Great value

Most chips only need to have electrical signals run through them to find out whether they work as intended. MEMS, on the other hand, require another type of physical input for testing, such as a temperature Credit: University of Twente

magine having to discard a product at the final stages of its production process because it turns out a chip is malfunctioning. This doesn’t happen very often in the electronics industry because chips are rigorously tested before they’re mounted. But not all companies, particularly manufacturers of specialized MEMS chips, have the luxury of low-cost and highly automated testing procedures. Because of their mechanical nature, MEMS are more complex to test than electrical ICs. Large MEMS manufacturers like Analog Devices and Bosch can afford to develop highly specialized testing equipment, but for companies with lower volumes, sometimes as low as a few thousand devices per year, that’s never going to be costeffective. Nor are there any specific ‘MEMS testing houses’ to outsource to. These companies have to resort to labor-intensive ‘manual’ testing, which often isn’t enough to prevent that a percentage of the final assembled product won’t work. Three companies and two research institutes, all located in the east of the Netherlands, have partnered up to solve this problem. Co-financed by the European investment fund OP-Oost, the Meteoriet consortium is looking to develop a universal MEMS testing technology. Make

An early version of the MEMS chip Meteoriet will be using for calibration.

increase, a movement or a changing magnetic field. A MEMS airbag sensor chip, for example, is tested by subjecting it to a massive deceleration. And a MEMS microphone device is exposed to a range of sound frequencies. Each type of device requiring a different physical stimulus makes MEMS testing hard to standardize. But perhaps it doesn’t need to be. The Meteoriet partnership is convinced that it’s possible to test MEMS chips the same way regular ICs are tested: by electrical means only. “The movement in a MEMS chip is associated with electrical signals, which can be measured and used to obtain information about the displacement that has occurred. It’s also possible to induce the desired movement electrically. Taking together, you can test a MEMS chip electrically,” explains Paul van Ulsen, CEO of test technology company Salland Engineering, which spearheads Meteoriet. An all-electrical MEMS test could look like this: an electrical stimulus is applied to induce a displacement of a particular mechanical part of the chip. Next, an appropriate electrical measurement is performed to verify whether the displacement did indeed take place. Other functionalities, such as temperature sensors or heaters, can be tested in a similar fashion.

C re

dit:

dE llan

ngi

nee

ring

Machine learning A Salland Enginneering 8-channel PXI instrument to measure extremely low capacitance, which is required for electrical MEMS testing.

The advantages of such a relatively simple procedure are myriad. It can be performed on the wafer, allowing for the elimination of malfunctioning devices at the early stages of the production process, reducing waste and saving costs. Wafer-level testing intrinsically also speeds up testing because many chips can be tested at once, and can also be automated quite easily. “Electrical testing of standard electrical parameters is already part of the MEMS testing toolbox, but a generic, universally applicable test solution for the physical parameters doesn’t yet exist,” says Van Ulsen.

Clearly, it would be very lucrative for his company if the Meteoriet partners succeed in developing one, especially given the rosy growth prospects of the MEMS market. “MEMS devices are often sensors, and sensors are an integral part of the IoT and Industry 4.0. Thus, speeding up and simplifying MEMS testing would help drive maturation of these technologies by reducing production cost.” For Salland, the deliverable of the Meteoriet project is a testing solution, along with the associated IP. The Zwolle-based company will deploy that in a variety of ways. “We’ll sell stand-alone test instruments, but we’ll also offer add-ons for existing equipment, so they can be made suitable for MEMS testing. Lots of existing IC testing equipment can be upgraded to be able to handle MEMS testing, which we believe is a great value proposition. On top of that, we can start offering test services for small and medium-sized companies that don’t have the critical mass to do testing in-house.”

One of the project’s main challenges is to figure out which stimuli and measurements are effective for testing different MEMS chips and their components. This is why the University of Twente (UT) and the Saxion University of Applied Sciences are involved in Meteoriet. These institutes are tasked with developing models that relate the non-electric physical quantities (movement, temperature, gas flow, and so on) to electrical characteristics to test the MEMS. “Our models are mathematical representations of devices that, given the input, reveal what the measured output says about the physical quantity that sits in between. From these models, Salland can derive the requirements for its testing instruments,” explains Dennis Alveringh, assistant professor at UT’s Integrated Devices and Systems group and part-time research scientist at Salland Engineering. “The model of a die can be viewed as a black box that takes the mechanics or other quantities like temperature changes out of the equations. The input and output are electrical, yet they yield information about something non-electrical happening inside the device,” adds Aleksandar Andreski, associate professor in nanophysics at Saxion. The goal of the project, however, isn’t just to find out whether a MEMS chip (or a 3 45

B a c kg r o u n d

Testing MEMS component) works or doesn’t work. Alveringh: “We also want to know how well it works. Because of random variations in the production process, some chips may have an above-average sensitivity in a certain measuring range, for example. Others may have an excellent response time. We can find out which test results are associated with such characteristics. This will allow manufacturers to select chips and match them with different product lines that are based on the same chip. This so-called product binning is a common practice in the semiconductor industry. And, finally, we might even be able to predict what the right calibration will be once the final product has been assembled.” In addition, Saxion will be validating the results. “At the end of the day, Salland needs to prove its claims about purely electrical MEMS testing to customers. At Saxion, we’ll collect the extensive data set to do so. Of course, as research institutes, we want to publish our research results, for which proof is required,” says Andreski. Finally, the consortium wants to know more about the nature of the defects and other anomalies encountered, and what effect these have on the test results. This is where a fourth consortium partner comes in, Enschede-based Maser Engineering, a failure analysis specialist (among other things). “There will be defects that won’t show up with electrical testing, and we’d like to know what type of defects these are. Maser has the equipment and expertise to paint a detailed picture. We’re also working on complementary automated testing methods to identify and characterize these defects. We’re developing machine learning software that can do just that from pictures of the wafer. Optical inspection already is more or less standard in MEMS testing, and we can easily upgrade Salland’s testing equipment to handle both electrical and optical testing – adding even more value,” Andreski continues.

Dear to my heart

Now, there’s just one piece of the puzzle missing: MEMS chips to test. This need will be partially satisfied by UT, which will design a state-of-the-art MEMS chip, taking into account requirements for testability. In parallel, a company itching for more effective testing procedures will do the same: mass flow sensor manufacturer Bronkhorst High-Tech from Ruurlo. 46

C re

: dit

B ro

H rst

igh

ch -Te

The IQ+Flow is one of Bronkhorst’s product series that’s already based on MEMS technology.

“We have MEMS-based products on the market for about 15 years now. Currently, about 5-10 percent of our offerings are MEMS-based. Recently, we’ve identified some very interesting new opportunities in the gas flow sensor markets that we’d like to capitalize on. However, scaling up our MEMS activities will require faster and more effective testing methods,” says Joost Lötters, science officer at Bronkhorst and professor in microfluidic handling systems at UT. Bronkhorst manufactures its own MEMS chip in the cleanroom of the MESA+ Institute for Nanotechnology, a shared production facility for micro and nanotechnology located on the UT campus. After production, these chips are subjected to a number of manually performed tests, among which an optical inspection and an electrical test of certain components, but that’s not enough to ensure the product works as intended. During final testing and calibration of the assembled module, some chips turn out to be malfunctioning after all. At Bronkhorst’s current needs of several thousands of MEMS devices per year, this situation is manageable. Expecting volumes to grow to tens of thousands of chips per year or more, however, the company’s labor-intensive testing procedures clearly

won’t cut it anymore. “These volumes are still not high enough for foundries. We really need access to comprehensive, automated testing,” states Lötters. That’s why his company gladly participates in Meteoriet to help Salland develop the testing solutions. Evidently, the chip that’s being designed as a part of the project will be used in next-generation mass flow meters. And so everything neatly falls into place within Meteoriet. Amazingly, all partners are located no more than 50 kilometers from one another, as the crow flies. “There’s a lot of knowledge and expertise in this part of the country, but we find it difficult to come together. I get it, there are always plenty of other priorities, right? Still, I think that cooperation between Dutch tech companies is very important to be able to compete in the global marketplace,” says Van Ulsen. “The Meteoriet project is particularly dear to my heart because it involves automation. Thanks to automation, costs are lowered, meaning manufacturing operations – and the associated jobs – that have been outsourced to Asia or Eastern Europe could return to the Netherlands,” Van Ulsen concludes. “That’s why I hope Meteoriet will inspire companies across the country to put in the effort and start working together more often.”

B a c kg r o u n d

TESTING

Hot topics in IC and electronic system testing – from all angles For already a quarter of a century, the last week of May is reserved for the IEEE European Test Symposium (ETS). The 26th edition has a balanced program where, besides professors and students, all key suppliers and industrial users of test equipment and design-for-test (DfT) software are present. Under pandemic rule, ETS-2021 will take place completely online, be it with several live events.

High-performance parallel parametric testing.

Erik Jan Marinissen Georges Gielen Michele Stucchi Elena-Ioana Vătăjelu

n its first 25 years, the IEEE European Test Symposium has set up its camps in almost every European country at least once. For example, the Netherlands has hosted it twice: in Maastricht in 2003 by Erik Jan Marinissen (at that time working at Philips Research) and in Amsterdam in 2016 by professor Said Hamdioui of Delft University of Technology. Belgium, on the other hand, was still one of the very few blank spots for ETS on the European map. When a team from Imec and KU Leuven in 2019 proposed to the ETS Steering Committee to host the 2021 edition in the historic city of Bruges, their bid was quickly accepted. Then, in early 2020, the Covid-19 virus outbreak struck and forced ETS-2020, originally planned to take place in Estonia’s capital Tallinn, to move to an online conference format. Although at that time, many were convinced that we would be freed from the pandemic and its restrictions on our daily lives well before May 2021, the Belgian organizers of ETS-2021 soon realized that there was no guarantee for a Covid-19-free world when their event would take place. International meetings such as ETS have a special responsibility with respect

Credit: Imec

to Covid-19 since their participants come from many different countries from all over the globe and have the potential to spread an infection at a global level. The only way to mitigate the associated risk was to turn ETS-2021 into a fully online conference. In the summer of 2020, the Leuven-based organization team canceled their reservations in Bruges, to fully concentrate on the virtual edition of ETS-2021.

Quiz

Online conferences offer many benefits over conventional in-person meetings. No need to travel, since one can participate from the comfort of one’s own home or office; obviously, that saves time, money, and

also contributes to ‘saving the planet.’ Of course, participants do lack the networking and interactions with peers. For organizers, conference budget items such as meeting rooms, AV equipment and banquets aren’t applicable in a virtual setting. Consequently, the registration rates for the 2021 virtual ETS edition could be reduced drastically. All recorded lectures will also remain online for a full month after the event. Most likely, everybody has already attended several online conferences and/or other events during the ongoing pandemic. Despite the clear advantages, many of us sense a feeling of remoteness and a lack of interactivity associated with such virtual events. The ETS-2021 organizers have 3 47

B a c kg r o u n d

TESTING

made extra efforts to address this issue and to maximally stimulate interactivity during the conference. The keynotes will be broadcasted live (and recorded for later viewing). Although all regular paper sessions, as well as the vendor sessions, will be pre-recorded, the speakers will be present during the entire session, so that the audience can directly interact with them via chat – this is a level of interactivity even beyond what’s possible at a regular (non-virtual) conference. To avoid the fatigue of sitting too many hours behind a computer screen, the program is restricted to only five hours per day (14-19h CET) and sessions are smeared out over four (instead of the usual three) days while keeping the usual amount of technical content. ETS also features a big online social event, where interactivity is of course paramount: the very first Global Test Community Quiz.

26th IEEE European Test Symposium

• live on 24-27 May 2021 • one more month on-demand online • reduced rates until 1 May • discounts: IEEE/CS members, speakers, students • 4 keynote addresses • 10 regular technical paper sessions • 3 industry sessions • 5 vendor sessions • 4 special sessions • 3 embedded tutorials • 1 panel session • poster sessions • European semi-finals of the Ed J. McCluskey Best Doctoral Thesis Award • virtual exhibition with 10 exhibitors • social event: Global Test Community Quiz 2021 ets2021.eu

DfT standard

In the ETS-2021 opening keynote, Oliver Dial of IBM’s TJ Watson Research Center will talk about quantum computers, which hold the promise to be able to crack problems that are simply too hard for today’s computers. While on the application side, the objectives and benefits are clear, the hardware implementation still has quite a number of tough hurdles to take. And, 48

of course, these new computers will bring their own set of unique test challenges – operation at cryogenic temperatures (just above 0 kelvin) is just one of them. ICs are also increasingly used in safetycritical applications such as automotive and healthcare; as these new domains have very little tolerance for test escapes, test quality needs to be improved. The ETS program has at least four presentations related to automotive quality, while the keynote by Chris Van Hoof, VP R&D at Imec and general manager of the Oneplanet Research Center in Wageningen, the Netherlands, will address disruptive new health devices. ETS also features papers on cell-aware test generation, a new approach that has proven itself with respect to test quality improvement. It explicitly targets realistic open and short defects inside the individual basic digital library cells. ETS-2021 offers a quite large number of contributions devoted to memory testing: next to classical charge-based memories such as SRAMs and DRAMs, new memories like resistive (RRAM) and magnetic (MRAM) memories are on the verge of a breakthrough in the market. Another new kid on the block is optical data propagation: silicon photonics. The circuit level addressed at ETS ranges from individual digital library cells, via full ICs, to printed circuit boards containing multiple ICs. Over time, papers on board testing have made way for papers on system-level testing and test and design-for-test solutions for 3D-ICs containing multiple dies stacked on top of each other in a single package. As the test infrastructure of stacked dies needs to work in concert to get test stimuli up and test responses down in the stack, in 2020, a DfT standard was released for such 3D-ICs (IEEE 1838) and several papers address this topic now at ETS-2021.

Revolution

ETS has expanded into domains neighboring to pure manufacturing tests, such as reliability and security. While the main question addressed in test papers is “whether the chip is okay at the start of its lifetime (t = 0),” reliability asks a similar question for t > 0. Hence, it’s no surprise to encounter reliability papers at a test conference. Having security papers at a test event is less obvious, as testability and security seem fundamental adversaries: testability

ETS-2021 collocated events 17-23 May 2021 (prior to ETS) Test Spring School: TSS 2021 “Robustness in new computing paradigms and technologies” 27-28 May 2021 (after ETS) Three focused workshops: • SURREALIST – Security, Reliability, Test, Privacy, Safety and Trust of Future Devices • AI-TREATS Workshop on AI hardware – Test, Reliability and Security • TAAA – International Workshop on Test Access, Automation and Adoption requires access to the internal components of an IC to be able to control and observe their state, while security typically wants to prevent such access. But the technical skills necessary to address both issues have a large overlap and an increasing number of products simply require to be both testable and secure. Ever since ETS has added security to its topics list, it has become one of the most popular topics, and the ETS-2021 edition is no exception. The keynote of KU Leuven professor Ingrid Verbauwhede will be on secure hardware design. The fourth keynote, delivered by professor Subhasish Mitra from Stanford University, will talk about a Cambrian revolution in system testing that grows beyond testing manufacturing defects to address robustness issues that end users really care about (like design bugs, reliability and security) in applications from (self-driving) cars to the cloud. These factors create golden opportunities for new system-driven test approaches that deal with the seemingly diverse problems at seemingly diverse scales in the emerging 21st-century designs. Erik Jan Marinissen (Imec, Leuven, Belgium) is the ETS-2021 industrial-relations chair. Georges Gielen (KU Leuven, Belgium) and Michele Stucchi (Imec, Leuven, Belgium) are the general co-chairs. Elena-Ioana Vătăjelu (CNRS/TIMA, Grenoble, France) is the program chair. Edited by Nieke Roos

d f t ou r o r a 1 Pr nso bito ean 02 o i op 2 sp exh ur ium d E s an EE mpo IE S y st Te

— Select ion 2 02 1 —

Salland No. 3

Salland No. 2

Applicos No. 9

Classical custom designed blend – enriched with almost 30 years of test expertise.

European testhouse for small and medium volume production test.

A unique characterization that perfectly blends with your sample testing.

A flavor that perfectly enhances the taste of your existing ATE.

Salland No. 5

Salland No. 42

Applicos No. 9

Applicos No. 12

Get the taste of tomorrow with our delicate High-Speed SerDes signal blends.

Our revolutionary sparkling and brilliant combination of silicon and light.

A unique texture that perfectly blends into your production process.

A wild and strong combination of generators, digitizers, and power supplies. Mix it!

Applicos No. 76

Salland No. 101

Salland No. 5

Salland No. 1

ATE-flavored analog characterization lab covered by a small and delicate box.

We extract the best flavor from every single atom. A marvelous micro-blend.

Test Applications

High-Speed

ATX7006A

Testhouse

Photonics

MEMS

Analog

Custom OEM

Supply Chain

Based on well-tolerated ingredients, this bite is a revelation for the ATE palate.

Salland No. 8

Instruments

Modules

Your Project

Once you try the crunch, you won’t want to miss our unique and flavorful boost.

B a c kg r o u n d

Photonics

Photonﬁrst ready to claim world leadership in integrated photonic sensing

Photonﬁrst CEO Daan Kersten.

Photonﬁrst is looking to become the world’s innovation leader in integrated photonic sensing. It’s just a matter of reaching the stepping stones that will take its technology to high-volume sales. With a few dozen of potential applications in six different markets, the Alkmaar-based company is Credit: Photonﬁrst quickly closing in on a pivot point. Paul van Gerven

n 1 January, the integrated photonic sensing business unit of instrumentation specialist Technobis was spun out as an independent company called Photonfirst. After fifteen years of lovingly nursing the technology to maturity, the time had come for the bird to leave the nest. However, it’s not quite ready to stand on its own two feet yet: it’s going to take another few years to scale the business in scope and size. “Our technology to measure temperature, pressure, acceleration, strain or shape deformation has already proven itself in some eye-catching applications, such as the overheat detection system for Airbus 350 planes. Or the sensors that measure the sway of the Taipei 101 skyscraper, which

some time ago was the world’s tallest building,” says Photonfirst CEO Daan Kersten. “Now, we need to move into killer applications that generate high-volume sales. But engineering a technology into costeffective end-to-end solutions takes time.” Interestingly, Photonfirst hasn’t identified one but several of such potential killer applications and is exploring them in parallel. Testament to the versatility of its lightbased sensor technology, the company is targeting no less than six markets: aerospace, energy, high-tech systems, infrastructure, medical and mobility. “We’d love to focus on a single application, but since we don’t know which one will be the first to take off, we can’t afford to put all our eggs in one basket,” Kersten explains. All companies looking to introduce a new technology have to go through this difficult phase of finding traction in a market,

but it may be a little harder for Photonfirst. After all, integrated photonics has only recently grown beyond the demonstration level, and few companies are aware of its potential, let alone possess the in-house knowledge to apply it. Fortunately, Photonfirst can count on the help of Photondelta, a Dutch organization tasked to accelerate the integrated photonics industry. Photondelta actively promotes the adoption of this disruptive technology and has successfully created an ecosystem that’s able to design, develop and produce innovative solutions based on integrated photonic technology. Photonfirst is a front-runner within this ecosystem, but we’ll know more in 2025, the year the Alkmaar-based company aims to generate sales exceeding 100 million euros, go public and become the global innovation leader in integrated photonic sensing.

Inert

Photonfirst’s sensor systems are based on optical fibers inscribed with microstructures called fiber Bragg gratings (FBGs), which partially reflect light traveling through the fiber. As the fiber is subjected to strain, a temperature change or another external influence, the wavelength of the reflected light changes proportionally. By measuring this shift, the strain or temperature change can be quantified. The devices that detect the wavelength change, so-called interrogators, are usually based on optical spectrometry. Spectrophotometers measure light intensity as a function of wavelength over a certain range. Normally, these instruments are rather bulky, expensive and power-hungry. That’s why Photonfirst has ‘shrunk’ a spectrophotometer to fit on an optical chip – or photonic integrated circuit (PIC), as they’re formally called. This massproducible PIC is not only much cheaper, smaller and more energy-efficient than regular spectrophotometers, it also features better performance in terms of resolution and accuracy. Taken together, FBG optical fibers and PIC interrogators make for a powerful duo. “You can fit a lot of FBGs in a fiber, allowing the system to cover large areas with ease. Optic fibers are durable, maintenance-free, lightweight, chemically inert, and don’t experience electromagnetic interference. The system as a whole scales very well to high-volume production, and when you do, costs drop dramatically,” lists Kersten.

Heartbeat

Given these features, it’s no wonder Photonfirst is eyeing so many different markets. In aerospace, the system can be employed for load sensing of landing gear, adding virtually no weight or additional safety risks. By keeping track of the strain experienced by the landing gear over time, it can be assessed when maintenance is required. In extreme cases, a catastrophic failure may be prevented. Similarly, a mesh of FBG measuring points can monitor the structural integrity of the airplane’s wings. Or of wind turbine blades. Or of bridges. In electric cars, the fibers could be used to extend the driving range by measuring the temperature distribution in the battery pack, allowing for optimization – batteries operate most efficiently at a certain temperature. Car manufacturers will be

Credit: Photonﬁrst

hard-pressed to find a solution that would take up less space and use less power. In the medical domain, Photonfirst’s technology can be used to keep track of a patient’s temperature while being subjected to an MRI scan. Obviously, any electrical temperature sensor would fail in this environment of high electromagnetic radiation. And since the heart generates a lot of electric interference too, the fiber-based sensors are ideal to use in cardiac surgery tools, for example as force transducers. In minimally invasive surgeries, the instruments are delivered to the affected area via hollow tubes inserted through a small incision in the skin. While limiting damage and thus speeding up recovery, a disadvantage of this technique is that the surgeon isn’t in direct contact with the tissue he’s manipulating. Photonfirst’s fibers, embedded in pincher instruments, can be used to measure the tightness of the grip and translated into haptic feedback on the handle of the tool. “We found that doctors could actually feel the heartbeat in a blood vessel,” Kersten says.

Coherent

This is just a small selection of applications that Photonfirst, together with potential customers, has explored already. “It’s amazing how many applications have been thought of over the past fifteen years. At least dozens. We’ve worked with many companies to demonstrate our capabilities, but it takes quite some time before such demonstration projects move up from the research domain onto product roadmaps. Basically, we’re driving for that to happen. At our end, everything is ready to make that next step.” “In a way, the disruptive nature of our technology is a disadvantage. For example, the telecom sector is already very familiar

with light-based technology. Introducing integrated photonics there can be as easy as swapping one component for another. Our technology, based on a principle that hasn’t been commercially applied yet, requires more effort to integrate in existing products.” Kersten is glad his company doesn’t stand alone in getting the integratedphotonics market off the ground. Photonfirst has partnered up with other companies, which together cover the entire integrated-photonics value chain, to form the Photondelta ecosystem. The Photondelta office in Eindhoven fosters collaboration, promotes the technology and provides financing, among other things. “Photondelta has been doing a great job increasing awareness about integrated photonics in various end-markets. It has also been advantageous for our company to be able to present ourselves as a part of a world-leading ecosystem. This inspires confidence with potential customers because they’re aware the success of a company depends on the supply chain it’s part of. In Photonfirst’s case: we do almost everything in-house, ranging from R&D to test & assembly, except for manufacturing the PIC. Among others, this is produced by Smart Photonics in Eindhoven, which is obviously a very strategic partner for us.” “If Photonfirst scales up, then Smart Photonics needs to be able to accommodate that. There are many interdependencies like this in the ecosystem. That’s why it’s very important for the value chain to develop coherently. Photondelta guides this process, for example by making sure all companies receive adequate funding. If only certain companies would succeed in finding funding, the roll-out of integrated-photonics technology would be severely hampered. Ultimately, we need each other to succeed.” 3 51

B a c kg r o u n d

WIRELESS

Creating digital superhighways with optical wireless communication By developing free-space optical communication technology and deploying it on Earth as well as in space, a recently formed Dutch consortium is going to give RF communication – and the optical ﬁber, for that matter – a run for its money. Paul van Gerven

i-Fi and other wireless communications techniques are great, but they have their limits: there’s only so much data you can squeeze into the radiofrequency spectrum. Even with advanced signal modulation and spatial multiplexing techniques, it’s a struggle to keep up with the exponentially growing data demand. Tellingly, 5G has to annex another part of the electromagnetic spectrum – the millimeter-wave domain – to open up additional capacity.

Further up the spectrum, in the visible and infrared region, there’s even more bandwidth available. We’re quite familiar with that part, of course, because packets of light have been zipping through optical fibers for decades already, at high speed and with low latency and excellent energy efficiency. Efforts to release these photons from their glass prisons have been started years ago, but there’s still some way to go before radio waves get serious competition from optical wireless communication.

Credit: TU Delft

A Dutch public-private partnership consisting of five universities, two research institutes and fourteen companies (see inset) have set out to push things forward. Having been awarded 4.1 million euros by science financier NWO, the Free consortium is aiming to develop free-space optical communi-

RFIC & analog/mixed-signal ASIC Solutions

www.bruco-ic.com 52

2020063 • ADV_BIC_190x75.indd 1

19-03-2020 13:32

cation at all length scales. In other words, whether it’s for connecting your laptop at home or for inter-satellite communication, the collaboration wants to build a wireless optical link for it. Apart from providing high bandwidth, free-space optical communication (FSO) has a number of intrinsic advantages, says Free program leader Eberhard Gill, professor of Space Systems Engineering at Delft University of Technology as well as scientific director of the TU Delft Space Institute. “FSO links are directional, so, unlike radio waves, they don’t spread out. This has two important consequences: they’re inherently more energy-efficient. Additionally, there’s no need for spectrum licenses. On top of that, FSO could be the most practical way to obtain the ultimate data security: quantum encryption.” The ambitions of the partners go beyond developing the FSO links over various distances, however. They envision using them to set up the ultimate ‘network of networks,’ in which a globe-spanning network of (mini)satellites is seamlessly integrated with ground-based networks. The ability to route connections through space opens up new opportunities for a wide range of applications, particularly those involving moving objects in need of high-bandwidth, low-latency connections. “From a network point of view, it would no longer be relevant whether a node is a drone, a satellite or a car,” says Gill.

Practical

Unsurprisingly, the space networking aspect of the project enthuses space engineer Gill in particular. “By linking the space and terrestrial domain, we’re moving away from the idea that these domains are separate. Space research and technology is often seen as a field of its own, having little overlap with the equivalent terrestrial applications. I’m really excited that we’re going to bring them together.” But in case of the Free program, it also makes a lot of sense to bring them together, Gill explains, as there are many commonalities between optical wireless communication technologies for space and terrestrial settings. “Receivers are a good example. We’ll be looking at developing receivers that are much more efficient than they cur-

Credit: TU Delft

rently are. It goes without saying that these devices will be equally useful for space and terrestrial applications.” “Developing actual applications will be up to the participating knowledge institutes and companies, of course, but the goal of the program is to push the technology readiness levels considerably. We want industry to be able to pick this up as early as possible, and we’ve taken a lot of steps to make sure the technology we develop will be manufacturable and practical in the real world.” “So, for example, all research questions of the fifteen PhD students that will work on this project will be verified and ap-

Partners of Free consortium Universities: Delft University of Technology, Eindhoven University of Technology, Leiden University, University of Twente, Vrije Universiteit Amsterdam Knowledge institutes: TNO, NLR Companies: Airbus DS, Aircision, Demcon, Effect Photonics, Flexible, Hyperion Technologies, ISIS, Lionix, Phix, Quix, Signify, Single Quantum, S&T, VTEC

proved by industry. We’ve defined sixteen use cases, which dictate the requirements for all technologies to be developed. And to smooth the transition from research to industry, we have an Architecture and Integration work package to spot manufacturability and system integration issues at an early stage.”

Competitive advantage

The big question, then, is when the world will start building the optical wireless superhighways that the Free consortium is envisaging. “That’s really hard to tell. All I can say is that I expect FSO technology to be adopted gradually. Initially, it might be in niche applications such as beaming down data from earth observation satellites. As the technology develops and matures, it will enter into more and more mainstream applications. Eventually, it will compete with RF in a lot of applications, though I don’t expect FSO to make RF obsolete entirely.” “It won’t take fifty years to reach that stage, nor will it take five. Let’s just say that I expect to see quite a move in FSO within the next ten years. After all, there’s a need for it in our increasingly digital society – the demand for bandwidth keeps growing. It will be a big market and I’m convinced our consortium will give the Dutch industry a significant competitive advantage.” 3 53

B a c kg r o u n d

Wireless

Building the internet of the future The demand for data has been exploding over the past years and will continue to do so for this decade to come. How will our telecom networks cope? KPN has partnered up with Eindhoven University of Technology to look at that challenge from every angle. Paul van Gerven

f there’s one sector that really shined during the corona crisis, it’s telecom. Reliable and high-speed internet connections kept large parts of the economy going and the quarantined entertained. This is only the beginning, however: many more opportunities are waiting to be unlocked by the beautiful marriage of telecom and digital services, ranging from self-driving cars to smart cities. This, however, places heavy demands on our wired and wireless networks, which will need to shuttle ever higher amounts of data back and forth, efficiently and affordably. How will telecommunications companies make this happen, without letting energy demands get completely out of hand? In a nutshell, this is what the publicprivate Smart collaboration is all about. Starting in 2017 with Smart One and recently expanded with Smart Two and Smart Two+, Eindhoven University of Technology (TUE) and telecom company KPN are elaborating what the internet of the future will look like, and what you can do with it.

Eavesdropping

Smart is an umbrella program, encompassing a wide range of research projects. “An eclectic approach is required because differ54

ent aspects across communication systems tie into one another. The overarching goal of our systems is to deliver ever more relevant data to users, economically and energy efficient. This depends as much on the characteristics of the network infrastructure as it does on the various higher-level aspects,” says Ton Koonen, who leads the Electrooptical Communications research group at TUE and oversees several Smart projects. One example would be how network technology affects security. Wireless connections are susceptible to eavesdropping because radio waves aren’t confined to a particular place or space. It would be great if the practicality of a wireless connection could be combined with the inherent security of a wired one. This is the focus of one Smart project: extending optical signals from optical fibers all the way to the living room (or any other room). Currently, optical signals tend to stop at the doorstep: as soon as they reach the building, they’re converted into elec-

High tech highlights

A series of public-private success stories by Bits&Chips

Credit: TU Eindhoven/Bart van Overbeeke

trical and radio signals for wired and Wi-Fi connections, respectively. Koonen’s group is working on ‘optical Wi-Fi,’ thus eliminating the need for signal conversions and enabling end-to-end optical connections. In this system, signals are distributed by optical fibers to ceiling-mounted antennas in different rooms. From there, they’re wirelessly transmitted to various devices using narrow infrared beams. Because these beams, unlike radio waves, are highly directional, eavesdropping is much more difficult. “In cases where very high security is needed, the windows could be coated with an infrared-blocking coating,” says Koonen. Other advantages of his optical wireless system include much higher data rates, lower latency and lower energy consumption, compared to regular Wi-Fi. Outside our houses and offices, radio waves networks will also face competition from other parts of the electromagnetic spectrum. In dense urban areas and crowded places like sports stadiums, 5G networks will harness the power of millimeter waves (mm-waves) to provide high-bandwidth wireless connections to a large number of people simultaneously. Wedged between microwave and infrared waves, mm-waves offer very fast data transfer, but at a cost: they don’t reach very far, they don’t bend around corners and they’re easily blocked by objects and even people. This last downside is being looked at by TUE’s Electromagnetics group. By studying the effects of human movement on the quality of mm-wave transmission, researchers expect to obtain clues about how to build mm-wave base stations that minimize this type of blocking. They’ll also gain

insight into where to best place the base stations.

Semantic data

A perhaps surprising element that enters into the equation when considering the communication systems of the future is architecture. The way our houses and buildings are constructed affects the kind of communication systems we can set up in there and hence what services can or cannot be provided. That’s why there are also Smart research projects performed together with TUE’s Department of the Built Environment. In terms of networking ICT, houses and buildings are the most heterogeneous environments, adds Nico Baken, strategist at KPN, part-time professor in Koonen’s research group and initiator of the Smart collaborations. “That makes them the most challenging environments for a company like KPN. The problem is: we don’t know much about them. It would enable great new possibilities if we had a kind of fingerprint, a kind of genome, of all our houses and offices. That would considerably help with designing the ideal network for them.” Particularly in smart buildings, in which heating, ventilation, air conditioning and lighting are controlled down to individual rooms and user preferences, it’s not just about the network and the sensors collecting the data, but also about data handling. Bringing together streams of different types of data is by no means trivial – cur-

rent methods are not yet fully capable of doing that. That’s why researchers at TUE’s Department of the Built Environment are studying digital representations of buildings that can integrate all the information. These digital twins (or, in this case, building twins) link different data models to improve monitoring, decision-making, automation, optimization and individualized control of buildings. This will improve user comfort while reducing energy consumption. Another project at the Department of the Built Environment is about self-driving cars, and, more specifically, about how they can safely find their ways around cities and highways. This, too, isn’t just a matter of hardware (network infrastructure and processors at various levels), it also requires new ways to handle data. The Smart researchers are looking into using so-called semantic data models for handling mobility-related data in the future. Essentially, these models, constructed using promising new techniques that are also used for building the next generation of the internet (Web 3.0), add context-related meaning to the data and the relationships between them. This allows for seamlessly One of the Smart projects is working on ‘optical Wi-Fi.’ In this system, signals are distributed by optical ﬁbers to ceilingmounted antennas in different rooms. From there, they’re wirelessly transmitted to various devices using narrow infrared beams.

linking and integrating different data sources, which will contribute to a safer and more efficient traffic ecosystem.

Very happy

Though future oriented, some Smart research is already paying off for KPN. “Especially in artificial intelligence, we enjoy the fruits of the program. Thanks to AI, we’re able to spot issues before they become big problems, so we can fix them before the customer even notices,” says Baken. Even if most research results need a couple of years of maturation before KPN can take advantage of them, they’re very valuable for his company, stresses Baken. “There’s profit and there’s value. We do not work with universities for an immediate profit, we’re in it for the value. It allows us to see more of the future, which helps us to determine which directions to take. Recently at KPN, we made a systematic inventory of our needs. These turn out to align beautifully with the research projects we’re involved in, so we’re very happy.” For his part, Koonen considers working with industry to be a natural phenomenon. “Before joining TUE, I worked in industrial research environments for over twenty years. So, like many of my colleagues here, I have many connections in industry. And that’s how it should be: we want our students to understand how to connect knowledge and skills with applications. After all, the technology that we’re working on here is meant to be applied.”

Credit: TU Eindhoven

3 55

Online

SOFT SKILLS & LEADERSHIP

Experimental techniques in mechatronics

Neurodiversity @work: coping with autistic characteristics in the technical world

Advanced motion control

Starts 5 July 2021 (3 consecutive days)

Starts 11 October 2021 (5 consecutive days)

20 May 2021 (1 half day)

Time management in innovation

Mechatronics system design – part 1

Presentation and body language skills

Metrology and calibration of mechatronic systems

Leadership skills for architects and other technical leaders

Mechatronics system design – part 2

Starts 31 May 2021 (2 times 2 days + evening sessions)

Starts 8 November 2021 (5 consecutive days)

Effective communication skills for technology professionals – part 1

Passive damping for high tech systems

Starts 23 November 2021 (3 consecutive days)

Presentation and body language skills

Starts 30 November 2021 (3 consecutive days)

How to be successful in the Dutch high tech work culture

OPTICS

Starts 11 October 2021 (5 consecutive days)

Starts 25 May 2021 (1,5 day)

Online

Starts 2 November 2021 (3 consecutive days)

Starts 25 May 2021 (3 morning sessions)

Dynamics and modelling

Starts 7 June 2021 (3 days + 1 evening)

Online

New

Online

Starts 22 June 2021 (3 sessions of 2,5 hours + e-learning) 1 July 2021 (1 day)

Modern optics for optical designers – Part 2

Effective communication skills for technology professionals – part 2

Starts 10 September 2021 (15 weekly morning sessions)

Applied optics

Starts 7 July 2021 (3 days + 1 evening)

Starts 1 November 2021 (15 weekly afternoons)

System architecting masters

Modern optics for optical designers – Part 1

Consultative selling for technology professionals Starts 4 November 2021 (2 days + 1 evening session)

SOFTWARE

ELECTRONICS

Starts 25 May 2021 (4 consecutive days)

Expected February 2022 (15 weekly morning sessions)

Starts 4 October 2021 (nine-month program)

Object-oriented analysis & design – fast track

Electronics cooling thermal design

Multicore programming in C++

Starts 31 May 2021 (5 consecutive mornings)

Starts 7 June 2021 (3 consecutive days)

EMC for motion systems

Design patterns and emergent architecture

Starts 23 June 2021 (3 consecutive days)

Starts 21 June 2021 (4 consecutive days)

Advanced thermal management of electronics

Speed, data and ecosystems

28 June – 1 July 2021 (4 consecutive afternoons)

Starts 13 October 2021 (2 consecutive days)

Test and design-for-test for digital integrated circuits

Good software architecture

Starts 7 September 2021 (3 consecutive days)

Starts 9 November 2021 (2 times 2 days)

EMC for mechatronic engineers

Software engineering for non-software engineers

9 September 2021 (1 day)

Starts 23 November 2021 (2 evening sessions)

EMC design techniques

Secure coding in C and C++

Signal integrity of a PCB

SYSTEMS

Microelectromechanical systems

Value-cost ratio improvement by value engineering

Solid State generated RF and applications

System requirements engineering improvement

Switch-mode power supplies

System architect(ing) in Eindhoven

Thermal design and cooling of electronics

Introduction to SysML

Starts 29 November 2021 (3 consecutive days)

Starts 20 September 2021 (4,5 consecutive days) Starts 20 September 2021 (2,5 consecutive days)

Starts 20 May 2021 (2 consecutive days)

Starts 28 September 2021 (3 consecutive days)

Starts 27 May 2021 (2 consecutive days)

29 September – 1 October (3 consecutive days)

Starts 31 May 2021 (5 consecutive days)

Starts 5 October 2021 (2 times 3 consecutive days)

28 June 2021 (1 day)

Starts 3 November 2021 (3 consecutive days)

MECHATRONICS

New

Starts 30 September 2021 (3 days in 3 weeks)

Starts 14 June 2021 (5 consecutive days) Starts 15 June 2021 (3 consecutive days)

2 + 9 September 2021 (2 days in 2 weeks)

Design for manufacturing

Motion control tuning

Thermal effects in mechatronic systems

Design for sustainability

New

System architecting masters

Starts 4 October 2021 (nine-month program)

Design principles for precision engineering

System architect(ing) in Zwolle

Starts 21 June 2021 (5 consecutive days)

Starts 11 October 2021 (5 consecutive days)

Advanced feedforward and learning control

Introduction to deep learning

Starts 23 June 2021 (3 consecutive days)

Starts 9 November 2021 (1 day)

Actuation and power electronics

System architect(ing) in Eindhoven

Starts 29 June 2021 (3 consecutive days)

Starts 15 November 2021 (5 consecutive days)

hightechinstitute.nl

pinion

WIRELESS Cees Links is a Wi-Fi pioneer and the founder and CEO of Greenpeak Technologies and currently General Manager of Qorvo’s Wireless Connectivity business unit.

Wi-Fi 6E: the new kid on the block

he Wi-Fi 6 standard runs in the so-called unlicensed, low-power 2.4 GHz and 5 GHz frequency bands. These bands have been in place since the beginning of this century following quite a long process, as each country has the autonomy to assign band usage within its own territory. In the years running up to 2020, the Wi-Fi industry has worked with about 80 MHz in the 2.4 GHz space and 580 MHz in the 5 GHz space – 660 MHz in total. There are two possibilities to get more data at higher speeds through the air: more bandwidth (more frequency bands) or more sophisticated radio communication (more megabits per MHz). Over the last 20 years and several generations of Wi-Fi, the data rate has increased by using more advanced radio technologies. The original 11 Mb/s of 802.11b was followed by 802.11a, 802.11g, 802.11n, and so on, with the data rate going up into the multi-gigabits per second. At the same time, the calls grew louder for the governments to provide more unlicensed spectrum. Around 2010, a new frequency band was made available in the 60 GHz space with a lot of spectrum and a lot of bandwidth – almost 13 GHz. The idea was to be able to go to very high data rates – and indeed, Gb/s were achieved. A standard was even developed, called Wi-Gig or IEEE 802.11ad. But the range, about 3 meters, was disappointing. As the number of applications requiring very high speeds over very short distances (‘Bluetooth on steroids’) was limited, the usage of the 60 GHz space hasn’t gotten much commercial traction.

In 2019, there was worldwide momentum to make 6 GHz available as a low-power unlicensed band because several applications using 6 GHz had started to fade away. This coincided with reports showing the economic value of Wi-Fi returning in taxed profits, taking away the concerns about ‘giving away’ spectrum to the industry. Governments prefer to license spectrum, so they can immediately monetize the usage in license

Wi-Fi has accelerated from 11 Mb/s to almost 10 Gb/s fees, while unlicensed spectrum is only indirectly monetized through taxes on economic activity generated through the usage of that spectrum. In the US, there was strong momentum to match the amount of spectrum licensed for 5G with a comparable amount of unlicensed spectrum for low-power, indoor applications, of which Wi-Fi is one of the best known. This led in April 2020 to the FCC announcement to ‘un-license’ the 6 GHz band for low-power usage. And thus, Wi-Fi 6E was born – the new Wi-Fi standard that extends WiFi 6 into the 6 GHz space. The 1,200 MHz of available bandwidth is a sizable extension compared to the combined 660 MHz that was available previously. Initially, we’ll only see Wi-Fi 6E as an increase in capacity – the number of Wi-Fi connections

that can be used at the same time (eg more users in the home using demanding networking applications, including video and gaming). The real next steps in higher data rate, together with exploiting the higher capacity, will be made when Wi-Fi 7 comes into fruition. Other countries have picked up the challenge. The end of 2020 saw a growing list of locations where WiFi 6E technology can be applied. In Europe, the UK took the lead and via the ETSI, the rest of the continent is expected to follow. (In Europe, at least for now, only the lower part of the 6 GHz space will be made unlicensed – about 500 MHz.) South Korea and several South American countries (Brazil, Chile, Peru) have decided to follow America’s 6E expansion. Furthermore, the FCC has proposed to add a 45 MHz band to the 5 GHz unlicensed spectrum. With 1,200 MHz available in the 6 GHz space, why is this important? Because the available unlicensed 580 MHz in the 5 GHz spectrum was just falling short of being able to accommodate three simultaneous high-speed 5 GHz channels. The additional 45 MHz solves that problem, allowing the development of more advanced products within the 5 GHz space. Over the last 20 years, Wi-Fi has accelerated from 11 Mb/s to almost 10 Gb/s – close to a factor of 1000! The 6 GHz addition will pave the way for engineers to figure out whether another factor of 1000 is achievable over the coming two decades.

3 57

B a c kg r o u n d

RF PCB designs – challenges, solutions and tips Today, RF circuitry is crammed into a large variety of commercial products. Most of these are handheld wireless devices for medical, industrial and communications applications. NCAB’s Harry Kennedy explains what to keep an eye on when designing the circuit boards. Harry Kennedy Credit: NCAB

he RF frequency range is usually from 300 kHz to 300 GHz, with microwave being anything above 300 MHz. There’s a considerable difference between RF and microwave circuitry versus typical digital and analog circuits. In essence, RF signals are very high-frequency analog signals. Therefore, unlike digital, at any point in time, an RF signal can be at any voltage and current level between the minimum and maximum limits. Moreover, a single band of signal can be very narrow or very wide and carried upon a very high-frequency carrier wave. RF PCB design is also very much different and difficult, compared to high-speed digital-signal board design. When handling RF boards, there are many new challenges for PCB designers.

Challenges

First, RF is far more sensitive to noise, incurring ringing and reflections, which must be treated with great care. The noise can be dealt with by properly terminating the signal, thereby solving the ringing and reflec58

tion issues. Another method is to optimize the return path with proper ground. Second, impedance matching is extremely critical for RF. The higher the frequency, the smaller the tolerance. Practically, if the total length of the trace from the driver to the receiver is greater than 1/16th of the wavelength of the signal, impedance control of that trace is required – 1/16th of the wavelength is called the signal’s critical length. If you’re routing a 1 GHz signal and its total length is greater than 425 mils, that trace needs to be impedance controlled. For example, 50 ohms out from the driver, 50 ohms during transmission and 50 ohms into the receiver. Third, the return loss must be minimized. At very high microwave frequencies, the return signal takes the path of least inductance. As a result, without good PCB design, it will go through power planes, through the PCB’s multi-layers or take some other route, and it will no longer be impedance controlled. Ground planes underneath the signals are good at providing an impedance-controlled path. Therefore, there should be no discon-

tinuities in the plane underneath the signal all the way from the driver to the receiver. Ground planes help minimize not only ground loop currents but also RF leakage into circuit elements. Fourth, crosstalk is a major issue in high-frequency designs. This is because crosstalk is directly proportional to the edge rates of the active line. In this case, the coupled energy from the active line will be superimposed on the victim line. When the board densities increase, the problem of crosstalk becomes more critical. As a solution, always leave adequate space around the signal trace for smooth bends and isolation of the RF signal. Keep all the traces coming out of or going into the transmitter and receiver modules as small as possible. The high-speed signals should be routed as far apart as possible. The length that lines run parallel to each other should also be kept to a minimum. All these measures will reduce crosstalk. Other solutions include reducing the dielectric spacing between the line and its reference plane and introducing a co-planar

structure, where a ground plane is inserted between the traces. Terminating the line on its characteristic impedance can also reduce the crosstalk by as much as 50 percent.

Other losses

There are other signal losses as well. There’s the skin effect loss, more specifically the skin effect loss on the trace of a signal. Dielectric loss is a companion since both can be created at extremely high frequencies, when electrons flowing through a conductor bounce back and forth with the electrons on the FR4 PCB substrate, for example. During this interaction, some of the energy from the electrons flowing through the conductor is transferred to the electrons on the FR4. Consequently, that energy is converted to heat and subsequently lost. In such instances, for extremely highfrequency microwave circuits, it’s best to use polytetrafluoroethylene Teflon (PTFE). These laminates have a dissipation factor of around 0.001 – compared to 0.02 for FR4. Using gold body on RF circuits can also greatly reduce skin losses. When using RF circuits, PCB designers need to consider the laminate properties, such as the dissipation factor and dielectric constant value and its variation. FR4 has a higher dissipation factor than highfrequency laminates like Rogers and Nelco. This means that insertion losses are much higher when using FR4. These losses are also a function of frequency and will increase as the frequency rises.

Credit: NCAB

The dielectric constant value of FR4 can vary as much as 10 percent. This in turn causes the impedance to vary. Highfrequency laminates have more stable frequency properties. Then there’s the dielectric constant value itself. When it comes to microwave circuits, this value is tied to the size of the circuit elements, so designers may be able to decrease the size of the circuit by choosing a laminate with a higher dielectric constant value.

Tips

To create better designs and improve the anti-interference for high-frequency PCBs,

there are a couple of tips engineers can take to heart. First, use inner layers as power ground layers. This will have the effect of shielding and even decreasing spurious inductance while shortening the length of the signal wire reduces cross-interference between signals. Turning the circuit layout 45 degrees will help reduce high-frequency signal emission and coupling. The shorter the layout length, the better, and the less the better for through holes. The layout between layers should be in a vertical direction, on the top layer in a horizontal direction and on the bottom layer in a vertical direction. This will help reduce signal interference. Increasing copper on the ground layer is helpful, too. Packaging important signal traces can obviously improve the signal’s antiinterference ability. Of course, we can also package interference sources to avoid interference on other signals. For the layout of signal traces, avoid loops and use a chrysanthemum link instead. In the power section of integrated circuits, it’s important to bridge the decoupling capacitor. Make the RF signal 50 ohms and always lay out the RF first. Last but definitely not least, isolation is key. Harry Kennedy is a field application engineer at NCAB Group. Edited by Nieke Roos

Credit: NCAB

3 59

INTERVIEW PETER VERMEULEN

“SOMETIMES WE FORGET, SOFT SKILLS CAN BE THE HARDEST SKILLS OF THEM ALL” High-tech development relies heavily on diverse teams of multidisciplined experts, each with a specific role. But in today’s cutting edge, it’s not just the expertise that’s diverse. According to High Tech Institute trainer and autism expert Peter Vermeulen, neurodivergent brains are also playing an essential role, and if utilized correctly, can lead to great benefits. Collin Arocho

“I

was a stay-at-home father with my one-year-old son when my former master’s thesis mentor called me about a new position with the National Autistic Society in Belgium,” recalls Peter Vermeulen. He was a trained clinical educationalist with an emphasis on education and child-rearing, with a particular focus on people with disabilities. “I think that in all of my education, I could refer to maybe one page on the topic of autism. But I went to the interview anyway – this was in the 80s and there weren’t a whole lot of job prospects in my field,” he explains. “I sat there answering a long list of questions, all about autism, and all of my responses were completely wrong. But somehow, they still offered me the job,” Vermeulen remembers. “It turns out, giving the wrong answers was a good thing. It showed them that I didn’t have to unlearn the many stereotypes and misunderstandings surrounding autism.” Instead, he could immediately get to work with the foundation, which was looking for a specialist that could aid and train parents of autistic children. “Now, nearly 40 years later, I still say that I never chose autism. Rather, autism really chose me.”

Credit: Peter Vermeulen

Neuroharmony

After some 4 decades in the field, not only has Vermeulen learned better answers to those interview questions, but he has literally written the book on autism. Actually, he’s written 15 of them and now he spends his

time speaking passionately about autism and neurodiversity in an effort to expel the myths and spread understanding about neurodivergent brains within society and the workplace. His goal is to reshape conventional thinking and replace the focus on the negative connotations of autism by highlighting the positive aspects of different minds working together. This is what he refers to as neuroharmony, a topic that lies at the heart of his numerous lectures and training sessions, including High Tech Institute’s “Neurodiversity @ work: coping with autistic characteristics in the technical.” “For decades already, people have created a false narrative that somehow puts all autistic people into a box, labeled as intellectually disabled, unable to hold meaningful jobs, get married or function in society – among many other stereotypes. But that’s simply not the case,” describes Vermeulen. “The reality is, no two brains are the same and some are just a little more different than others. In fact, many of the behaviors regarded as autistic are expressed similarly in ADHD, dyslexia and super intelligent giftedness.” According to Vermeulen, while neurodiversity can certainly present

barriers or challenges in the workplace, the push for neuroharmony, aimed at recognizing and welcoming the various neuro types, can actually serve as a great asset for companies, particularly in the high-tech domain. In fact, he points out that that’s why some of the biggest tech companies in the world, the likes of Microsoft and HP, are actively employing more autistic and neurodivergent people. “Many companies have the idea that people with autism are ‘weirdos.’ But one thing they’re learning now is that when you need creative, out-ofthe-box ideas, there’s no one better at coming up with a unique approach than these ‘weirdos,’” illustrates Vermeulen. “It’s not very easy to be original these days, and companies are pouring millions of euros into efforts to differentiate themselves with team-building exercises and brainstorming for fresh ideas. But when everyone in the room thinks exactly the same, you keep getting the same ideas over and over.” For autistic people, thinking out of the box and being creative isn’t a problem. Not because they necessarily possess incredibly high levels of creativity. Rather, it’s simply in their nature. Vermeulen: “It’s not a choice

they’re making. Their brains simply work differently to process information and offer them a different view. While to us, it may seem out of the ordinary, to them, it’s just their normal perspective. A perspective that can be a real benefit to a company willing to recognize the value.”

Context blindness

However, even with the benefits that neurodiversity has to offer the workplace, Vermeulen also admits some challenges can arise. One of the most difficult lies in social interaction and communication. “The human mind is an extraordinary vessel for information and conveying messages. However, we don’t say everything we mean. A large part of the way we communicate is through inference and building context. This is something that remains very difficult for people – particularly those with autistic brains, and is what we refer to as context blindness,” explains Vermeulen. “If I’m your boss and tell you something is due by the end of the week, most people have no problem understanding when it’s expected. For some, however, the end of the week can cause some distress. Does

3 61

Credit: Imec

From neurodiversity to neuroharmony

that mean Friday morning, Friday at 4:59 PM, or even Sunday at 11:59?” depicts Vermeulen. “Yes, for some, this may seem trivial and can even be an annoyance, but not everyone processes information and makes inferences the same way. That’s why I put a lot of effort and focus into concrete communication – designed to be so clear that you can maximally avoid any misunderstandings.” By paying extra attention to providing clear and direct feedback and explanation, some of the social difficulties in communication can be wholly avoided, believes Vermeulen. That’s why it’s one of the main themes of his one-day neurodiversity training. In fact, in his training, to illustrate this point, he asks participants to draw pictures to express their message. The only rule is to be so clear and concise that you don’t need to use words. “We have beautiful ways to express ourselves in all the different languages, but sometimes that means we talk around things, rather than being explicit and direct. That’s difficult for many people, not just those with autism,” Vermeulen points out. “But what I’ve found is that if this style of communication helps you more effectively speak with autistic colleagues, then it will certainly help you be more clear with those who are nonautistic. It works for everyone.” “Especially within the context of the high-tech industry, there’s a real separation between hard and soft skills, but I think sometimes we forget that for some, these social or soft skills can be the hardest skills of all. Good and clear communication is one of several keys to success, no matter what the job. And I must say, in all my years in lecturing and training, I’ve never met a person who says at the end, ‘Peter, thank you very much, but this was way too clear to me and I understood everything. So please, next time could you make it a little bit more confusing,’” jokes Vermeulen. “It’s something that can be helpful for everyone.”

pinion

SOFT SKILLS Hans Odenthal is the HR manager at Sioux Technologies.

The power of the why

doubt if Anita Meyer’s song “Why, tell me why” would have made it to the top of the charts if it had been called “How, tell me how.” And still, we mostly talk about the ‘how’ in our meetings. In meeting notes, we write down the tasks to be performed as SMART as possible. In principle, there’s nothing wrong with describing how we want things to be done. But is it really the smartest thing to do? When having a meeting, the subject often shifts quickly from the things we want to achieve to how to get there. And that’s where the fireworks begin. Many discussions really ignite on small details, even though we already agreed on the proposed result or direction. On the other hand, we rarely speak about the ‘why,’ the reasoning or context behind the question. Weird, isn’t it? Especially if a company encourages you to take more responsibility. And you can only do that if you understand the context. To me, taking on actual responsibility is very important. In my daily work, I try to focus as much as possible on the ‘why’ and leave the ‘how’ up to others. When I train a new colleague, I typically start by telling him what tasks he needs to do. But I also give the reasoning behind the task and explain the context and what result he should achieve. If he has other or better ways than I would have employed to achieve the same, that’s fine. Sometimes, we get into nice discussions about the context and whether it’s still valid. Sometimes we conclude that the suggested tasks aren’t even needed anymore. That’s the power of the why.

The why’s side effects are also very positive. Describing the why is much more inspiring than talking about the details of the implementation. Somebody who understands why he’s doing what he’s doing is much more involved in a project. And it’s not just a task that’s being trans-

You can only take more responsibility if you understand the context ferred, but implicitly also the responsibility to think for oneself. This adds to work satisfaction. So, a win for all involved. It’s not straightforward to move from how to why, however. You need task-mature people. Not only on the executing side but especially on the delegating side. The latter is probably the most difficult since it’s so much easier to tell somebody what you want him to do. I think that this is the main reason why it’s so difficult to achieve actual freedom and responsibility in one’s work. The people who should delegate often got their leading position because they excelled at performing those tasks in the past, making it too tempting to transfer the ‘how to do it’ to the next generation and forgetting to transfer the more important ‘why.’ We’ve been taught that you need to grasp the real problem behind

the original customer demand. For example, you can use the 5-whys methodology for this. We all embrace this way of working when we try to build a system. I strongly suggest doing this for building an organization as well. I recommend adding a “Why” column to the action list in meeting notes. And to answer Anita Meyer’s question in your next meeting and tell everybody why, tell them why.

UPCOMING ISSUES

Bits&Chips in 2021 Bits&Chips is transitioning toward the future of digitalization. As of this year, we’re sending out a monthly digital edition, filled with exclusive content available to our readers with a registered membership. It offers a recent overview of the important news and background stories from Dutch and Belgian high-tech, as well as interviews, analysis and opinions from our host of domain experts and innovators – all with the simple click of your mouse.

Interested in contributing? nieke@techwatch.nl

The digital magazine issues 8 times a year, of which 3 specials will be published in print too. Edition

Date of publication

Published

1+2

31 March

Online

14 May

Online + print

25 June

Online

3 September

Online + print

8 October

Online

19 November

Online + print

10 December

Online

Interested in advertising? sales@techwatch.nl

About Bits&Chips

Bits&Chips is an independent news magazine for people who work to make smart products and machines. Bits&Chips is a publication of Techwatch BV in Nijmegen, the Netherlands.

Transistorweg 7H – 6534 AT Nijmegen tel +31 24 3503532 info@techwatch.nl techwatch.nl

Training courses Linda van Hoeij – course manager tel +31 85 4013600 – linda.van.hoeij@hightechinstitute.nl Petry Jansen – marketing and sales employee tel +31 85 4013600 – petry.jansen@hightechinstitute.nl Mariska van Hoeven – marketing and sales employee

Editorial Nieke Roos – editor-in-chief tel +31 24 3503534 – nieke@techwatch.nl René Raaijmakers – editor tel +31 24 3503065 – rene@techwatch.nl Paul van Gerven – editor tel +31 24 3505028 – paul@techwatch.nl Collin Arocho – editor tel +31 24 3503533 – collin@techwatch.nl Alexander Pil – editor tel +31 24 3504580 – alexander@techwatch.nl Design Justin López – graphic designer and illustrator tel +31 24 3503532 – justin@techwatch.nl Sales, marketing and events Kim Huijing – head of marketing and sales tel +31 24 3505195 – kim@techwatch.nl Marjolein Vissers – event manager tel +31 24 3505544 – marjolein@techwatch.nl Mariska van Hoeven – marketing and sales employee tel +31 24 3505544 – mariska@techwatch.nl Bo van Gaal – marketing and sales employee tel +31 24 3505195 – bo@techwatch.nl

Administrative Mathilde van Hulzen – finance tel +31 24 3503532 – invoices@techwatch.nl Advisor Maarten Verboom External staff Femke Veldhuis, Babette van Wichen Contributing writers Maarten Buijs, Georges Gielen, Robert Howe, Angelo Hulshout, Harry Kennedy, Cees Links, Erik Jan Marinissen, Hans Odenthal, Anton van Rossum, Michele Stucchi, Elena-Ioana Vătăjelu Publisher René Raaijmakers tel +31 24 3503065 – rene@techwatch.nl ISSN 1879-6443 Publisher in Belgium René Raaijmakers Biesheuvelstraat 1 2370 Arendonk, Belgium Printer Vellendrukkerij BDU

Subject to change

Bits&Chips membership The print editions of Bits&Chips are free of charge for delivery in the Netherlands. The digital magazines are available to registered members. You can join as a member for 9 euros per month or 99 euros per year. That gives you (in addition to the digital and print magazine) access to the exclusive content on bits-chips.nl. For companies, we have a business membership costing 150 euros per year. That includes two copies of every print edition, sent and billed to the company address, and package deals for multiple event tickets. To request a membership, please visit bits-chips.nl/subscribe or write to info@techwatch.nl. Back issues: write to info@techwatch.nl. Delivery complaints Did Bits&Chips arrive late or not at all? Other comments related to delivery? Please let us know by sending an email to info@techwatch.nl. Advertising Advertising rates are listed on our website (bits-chips.nl). If you would like to be informed of our upcoming themes and specials or to reserve advertising space, please contact our sales department, tel +31 24 3505195 – sales@techwatch.nl. Publication dates 31 March, 14 May, 25 June, 3 September, 8 October, 19 November, 10 December Copyright All rights reserved. (c) 2021 Techwatch BV. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher.

Disclaimer Though the publisher and editors take the utmost care in creating, compiling and distributing the information in Bits&Chips, they cannot guarantee the information is correct or complete. The publisher and editors are in no way liable for damages that may arise in connection with the publication of information in Bits&Chips. Columnists and external staff write in a personal capacity. Reader responses fall outside the publisher’s and editors’ scope of responsibility. The publisher and editors are in no way liable for the content and signature of reader responses. The editors reserve the right to edit responses and to publish them in part or not at all. Cover Main photo: Nexperia

bits-chips.nl

CONTENT SERVICES Content is key • Do you want to stand out? • Are you too busy to generate content? • Do you want to share your technical success stories? Our tech writers can help you cut costs and save time with in-depth technical texts, from blogs and newsletter texts to background articles and case studies. We write in both Dutch and English (native). You will receive full ownership of the text and you can choose how to distribute it. We’re already creating content for a number of customers, including High Tech Software Cluster, Holland High Tech, NLR, Photondelta, Salland Engineering and Unit040. For publication in Bits&Chips, we offer a reduced publication fee. Contact us via sales@techwatch.nl for more information.

bits-chips.nl/content-services

Bits&Chips 3 | 14 May 2021

Articles inside

Sometimes we forget, soft skills can be the hardest skills of them all

Building the internet of the future

Creating digital superhighways with optical wireless communication

Photonfi rst ready to claim world leadership

Dutch Meteoriet consortium tackles MEMS testing bottleneck

How to measure a planet

Industry 4.0 is a means to an end

Prodrive streamlines software development with its own platform

Machine learning isn’t hard with OpenML

Creating software to keep naval systems always-on

Clearing the critical software path

Prepare to be frustrated and try to remember, it gets better

Seven Dutch hopefuls in battery technology

ASML reduces DUV overlay error to 1 nanometer

ASML’s sales to surge 30 percent this year

Partial de-globalization appears inevitable – Maarten Buijs

Nexperia is growing into a new set of clothes

Avular and Sorama team up to soothe the buzz of drones