Bits&Chips 4 | 4 September 2020 | Trends in software development by Bits&Chips

4 SEPTEMBER 2020 | 23 OCTOBER 2020

Has Intel lost its mojo?

COMMA INTERFACES CONQUER HIGH TECH

BITS-CHIPS.NL

Holst betting big on the foil battery

PROGRAM

Chair Aad Vredenbregt (Valoli)

23 SEPTEMBER 2020 VERKADEFABRIEK â&#x20AC;&#x2122;S-HERTOGENBOSCH

09:30

Registration

Keynote 09:45

AI on the edge: interacting swiftly and surely in a dynamic world Nicolas Lehment (NXP)

10:30

Break

11:10

Less is more: do 20 times more machine learning with 95 percent less compute Menno Lindwer (Grai Matter Labs)

11:50

Building custom AI hardware for edge computing Nick Destrycker (Edgise)

12:30

Lunch

13:30

Targeting energy waste in homes using AI and IoT data Stephen Galsworthy (Quby)

14:10

Deep learning in healthcare: cuff-based blood pressure estimation is a thing of the past Vincent Janssen (Verhaert AI Lab)

14:50

Break

15:10

Hybrid machine learning and domain engineering knowledge for real physical machines Guillaume Crevecoeur (Ghent University)

15:50

Quickly finding performance drivers in the high-dimensional data streams of advanced lithography systems Vincent Aarts (ASML)

16:30

Break

16:50

Barriers to adopting AI for engineering Albert van Breemen (TUE)

17:30

Drinks

Silver sponsor

Sponsors

Partners

MLCON.NL #BCML

4 Subject to change

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

Penny wise, pound foolish

n 2017, a group of experts chaired by former WTO chief Pascal Lamy recommended more than doubling the EU research budget to 160 billion euros to “allow funding of a reasonable proportion of proposals.” A year later, the European Commission proposed 100 billion euros – still a substantial increase over the 77 billion allocated to the running research program, Horizon 2020. The European Parliament, however, deemed 120 billion euros a better number. But as it stands now, the next EU research program, called Horizon Europe, has been allocated a core budget of 75.9 billion euros, boosted by a one-off 5 billion euros from the pandemic recovery fund. Without a doubt, European leaders had some really big fish to fry at the heated budget summit last July. The budgetary consequences of the UK leaving the Union and the preferred way to soften the economic blow caused by the corona outbreak opened some deep rifts across the continent. Still, it’s disappointing to see research and innovation used as a balancing item. Painfully, the Dutch government, along with three frugal allies, carries more than a little responsibility for slashing the EU’s research ambitions. These countries had but one priority: pay as little as possible, pressing issues such as the international tech race and climate policy be damned. In the case of the Netherlands, this shouldn’t have come as a surprise. Gert-Jan Koopman, director-general of the European Commission’s Budget department predicted as much last year. As a guest in the podcast “Betrouwbare bronnen,” Koopman characterized the Dutch govern-

ment’s approach to the EU budget as “a little schizophrenic.” Beforehand, it’s all about getting Europe ready for the 21st century: cutting spending for agriculture and cohesion policy in favor of more strategic activities, such as research and innovation and industrial manufacturing. For example, take Finance

Pressing issues such as the international tech race and climate policy be damned Secretary Wopke Hoekstra’s Humboldt Speech on Europe, delivered in May last year. “Our continent’s economic engine is stuck in second gear, and we’re investing very little in the economy of the future. Artificial intelligence, big data, nanotechnology and biotech – we’re hardly doing anything in these fields,” Hoekstra warned, arguing for an increase in spending on research and technology development. (Author’s note: I used this quote before as a sign of hope. Please forgive me, I didn’t know what I know now.) Once at the negotiating table, however, these lofty ambitions get tossed out of the window. This is what happened last time (2014-2020), and it’s what transpired last July. Prime minister Mark Rutte returned to The Hague having secured a lean EU budget and an additional 450 million euro per year discount on Dutch EU

membership. Meanwhile, Dutch universities stand to lose 400 million in research grants, according to the president of the Association of Universities in the Netherlands (VSNU). I mean, I get it. The discount is an easier sell at home. Next year is an election year and the populists are breathing down the necks of the center-right coalition parties. And other European powers aren’t exactly championing modernization either, addicted as they are to the traditional subsidies from Brussels. Still, Rutte & Friends could have been a little less myopic about the budget size. Europe is already being outspent left and right on R&D. Amid major tectonic shifts in the geopolitical world order, this represents a serious long-term threat to our prosperity. On a total budget of 1.8 trillion euros, another 80 billion to double R&D spending is peanuts. Fortunately, the numbers aren’t yet final: the European Parliament, which needs to approve the budget, may be able to force some changes to the agreement. Let’s hope European leaders, the frugal ones, in particular, realize the strategic mistake they’ve made in the heat of battle and come to their senses.

CONTENTS IN THIS ISSUE OF BITS&CHIPS

Analysis

News

Has Intel lost its mojo?

Holst betting big on the foil battery

Intel has forfeited technological leadership deﬁnitively. Is the king of the semiconductor hill about to lose its crown?

Can billions of micropillars make for a better battery than we currently have? Yes, Holst Centre believes.

Dutch silicon anodes inch closer to the gigafactory

News 7 8 11 12 13 16 18 22 25

Surveilling aging infrastructure using ﬁber-optics

Noise Has Intel lost its mojo? US-Chinese trade war leaves semiconductor industry in limbo Dutch silicon anodes inch closer to the gigafactory Holst betting big on the foil battery Surveilling aging infrastructure using ﬁber-optics Pink RF and America’s Odyssey link to bring RF repair to Nijmegen Pandemic accelerates ASML’s adoption of AR Performing an ultrasound on a wafer to align it

Cracking the code to craftsmanship

Opinion 3 15 19 27 35 47

Penny wise, pound foolish – Paul van Gerven The headhunter – Anton van Rossum Wi-Fi thermal challenges for RF front-end designers – Cees Links Engineering the corona response – Wim Hendriksen What’s this thing called software engineering? – Han Schaminée Mastering complexity – Marcel Pelgrom

Theme Software engineering 28 32 36 40 42 44

Comma interfaces open the door to reliable high-tech systems No systems engineering without digital engineering Cracking the code to craftsmanship A pressure cooker for software talent Optimizing your high-tech development and machine performance High-level FPGA programming for nanosecond timing in terabit communication

2020

UPDA TED

EVENT CALENDAR

23 SEPTEMBER 2020, ’S-HERTOGENBOSCH

24 SEPTEMBER 2020, ’S-HERTOGENBOSCH

Theme Software engineering

Comma interfaces open the door to reliable high-tech systems

7 OCTOBER 2020, EINDHOVEN

Once a research project, initiated by ESI and Philips, the Comma framework is now conquering Dutch high tech.

4 NOVEMBER 2020, EINDHOVEN

“High up in an organization, you’re busy with keeping management at ease”

Theme System architecting

48 “High up in an organization, you’re busy with keeping management at ease” 52 True architects understand the art of omission and know where to dig deeper

Interview

56 Innovation and character light the path to IMS success

5 NOVEMBER 2020, EINDHOVEN

BITS&CHIPS

BENELUX R F CONFERENCE

12 NOVEMBER 2020, NIJMEGEN

BITS&CHIPS

INDUSTRIAL 5G CONFERENCE

DATE TO BE ANNOUNCED, EINDHOVEN

bits-chips.nl/events

7 OCTOBER 2020 IGLUU EINDHOVEN

Platinum sponsor

Bronze sponsor

Sponsor

Partner

IDEA2INDUSTRY.COM

#BCI2I

NEWS

NOISE

Top semiconductor sales leaders H1 2020, in million dollars Rank 1 Rank 1 H20 H19

Company

1H20

1H19

Change (%)

38,951

32,038

Intel

Samsung

29,750

26,671

TSMC

20,717

14,845

SK Hynix

13,099

11,558

Micron

10,624

10,175

Broadcom

8,109

8,346

-3

Qualcomm

7,857

7,289

Nvidia

6,525

4,674

6,241

6,884

-9

Hisilicon Top-10 total

5,220

3,500

147,093

125,980

Innovation

Corona prompts meteoric rise of health passport in Hype Cycle

Technologies rarely enter the Gartner Hype Cycle for Emerging Technologies with a 5-20 percent market penetration, but this year, one did: health passports. This is mostly because of widespread adoption in populous countries such as China and India, where large numbers of citizens and visitors are required to use an app that indicates Covid-19 status if they want to use public transportation, for ex-

Compared to the first half of last year, the sales of the ten biggest semiconductor companies surged by 17 percent in the first half of 2020 – more than three times the industry-wide 5 percent increase. Three top-10 companies saw their sales jump by a whopping 40 percent or even more. TSMC, ranked third behind Intel and Samsung, mainly profited from a surge in sales of 7nm application processors to Apple and Hisilicon. Nvidia, at rank 8, strongly recovered from a ‘crypto hangover’: after seeing demand for graphics cards explode during the cryptocurrency mining boom, Nvidia’s sales were in the doghouse for quite a while when the boom ended. And finally, a 49 percent sales increase catapulted Huawei’s chip subsidiary Hisilicon into to the top 10, though its time in there may be short in light of US sanctions. PvG

Artiﬁcial intelligence

David Attenborough narrates Reddit threads

Garett MacGown decided that the world cannot not do without Sir David Attenborough’s narration. Using audio samples obtained from Youtube, the software programmer employed machine learning to mimic the sonorous voice of the British wildlife documentary broadcaster, now aged 94, and had him read out loud content on the community website Reddit, including threads on relationship advice. It’s not exactly the most riveting material, but the result, most people will agree, isn’t bad at all. Though the somewhat tinny sound gives away the fact that the audio is computer generated and there are some strange mispronunciations, there’s definitely an Attenborough-esque quality to it. The deepfake was created using Google’s text-to-speech software, improved by employing a software-generated voiced model trained on Attenborough’s real speech. PvG

Credit: Gartner

Source: IC Insights

ample. Gartner currently classifies health passports as being on the “peak of inflated expectations” but expects them to reach the “plateau of productivity” in less than two years. Other emerging trends to watch, according to the market researcher, include beyond-silicon electronics (such as carbon-based transistors and DNA computing and storage), the ‘digital me’ (digital representations of people) and algorithmic trust (models that ensure privacy and security of data). PvG

Semicon

A death sentence for Huawei

The global chip and smartphone industries are bracing for a major disruption after the American government yet again tightened restrictions for Huawei. US sanctions already made it impossible for the telecom company to design its own chips – because that requires software from American companies – or to have them manufactured at TSMC – because the foundry uses American semiconductor equipment for that. That still left Huawei with one option: buying off-the-shelf chips. This escape route has been all but cut off now that the US also requires a license for the sale of any chip developed or produced using US technology to the battered Chinese company. “Huawei is probably finished as a maker of 5G network equipment and smartphones once its inventories run out early next year,” stated Gavekal Research in a report. This, in turn, will hit Huawei suppliers, though the impact may be (partially) offset by competitors regaining market share. PvG 4

ANALYSIS CHIPS

Has Intel lost its mojo? Having failed to impress in new markets and being circled by the competition on its home turf, Intel now also has forfeited technological leadership definitively. Is the king of the semiconductor hill about to lose its crown? Paul van Gerven

s great an achievement as it was, the release of the Intel 4004 was merely a very small step towards the company’s rise to global dominance in the semiconductor industry. As Tim Jackson notes in “Inside Intel” (1997), the commercial introduction of the world’s first microprocessor “did not yet make the company stand out far from other startups trying to make money from the uncertain new technology of integrated electronics.” More than anything, what drove Intel’s success was a ruthless commitment to flawless execution, Jackson argues. “It became a company whose slogan was to deliver – to make sure its good ideas were turned into practical products that customers could use, that arrived on schedule and at prices that fell consistently year by year. This transformation was no mean feat. It forced Intel to become rigorously organized and focused, and to find a balance that allowed it to keep firm control over its operations without jeopardizing the creativity of the scientists who were its greatest asset.” The path to success was by no means smooth, but by the early 90s, Intel had become the world’s leader in semiconductor sales and it didn’t give up the top spot for nearly a quarter of a century. Only during a particularly strong boom in memory, in 2017 and 2018, did Intel have to settle for second place, behind Samsung. Yet all that glitters isn’t gold. Over the years, Intel’s efforts to diversify have amounted to miserable failures. Starting in the late 90s, for example, Intel ended up spending over 10 billion dollars to force entry in the digital signal processing and optical communication markets but ended up with little to show for it. 8

Such stumbles, however, were kept covered under a blanket of excellent financial performance in the microprocessor market. Even though for the moment that blanket seems as comfortable as ever, judging from recent quarterly results, industry observers nonetheless believe the once unshakable behemoth now is in danger of taking a nasty fall that may be difficult to recover from.

A field day

Intel currently faces existential threats on two fronts. First, leveraging its manufacturing prowess, combined with smart Intel Inside marketing, no one could keep up with the company in the PC market for a long time. But new high-volume semiconductor markets emerged, many of them not necessarily CPU centric, and Intel has yet to impress in any of these. As EE

CEO Bob Swan speaking at the 2019 Intel Investor Meeting. Credit: Walden Kirsch/Intel

Times’ Bolaji Ojo observed, the king of the PC “has been outmatched by lesser beings.” Even worse, the competition is already circling Intel’s home turf. Apple has ditched X86 and Arm and GPU chips are lining up to move into PCs and data centers. Infinitely more painful, however, is Intel’s loss of technological leadership. Its 10nm chips arrived a stunning five years late: originally scheduled for volume production in 2015, only this year have they started to ship in meaningful numbers. By its own admission, the company’s technological goals for 10nm chips had been set too ambitious. Aggressive scaling goals combined with the implementation of several new technologies, such as cobalt interconnects and self-aligned quadpatterning lithography, proved extremely problematic to realize.

In January this year, at CES 2020, Intel previewed the much-delayed 10nm Tiger Lake mobile PC processors. Credit: Intel

Intel’s promises of catching up went down the drain when the company recently divulged that development of its 7nm process is delayed, too. With manufacturing yields of its 7nm process one year behind schedule, TSMC will soon be in the lead by as much as a full process node. AMD, Intel’s only X86 rival and customer at the Taiwanese foundry, will have a field day. Already the fabless is making competitive products: it recently reached an all-time high market share of nearly 20 percent in PCs and notebooks, according to a market analysis from Mercury Research.

Manufacturing momentum

Intel is now all but waving the white flag. CEO Bob Swan recently spent almost an hour discussing an idea that not long ago would get him burned at the stake: outsourcing production for core products. “To the extent that we need to use somebody else’s process technology, we’ll be prepared to do that,” Swan told analysts on a conference call. “That gives us much more optionality and flexibility. So in the event there is a process slip, we can try something rather than make it all ourselves.” The flexibility is “not a sign of weakness,” Swan insisted, stressing that Intel would keep investing in its own manufacturing capabilities. But many industry observers

believe that he de facto announced the end of the IDM era. Not immediately, but in five to ten years. This is the harsh reality of going fablite, after all. What’s the point of sinking large sums of money in your own fabs, if you need a foundry to manufacture your most advanced chips? That would only make sense if there was any chance of regaining technological leadership. For Intel, that would require a miracle. Not only is leading-edge process node development at a regular pace already astronomically expensive, the company will also have to accept reduced margins if it outsources production. This will hurt R&D expenditure and capex. Meanwhile, the foundry industry will strengthen its position, both financially and in terms of manufacturing momentum, making it even harder to catch up.

Only the paranoid

How could Intel have fallen so deeply? Over the past few years, several former employees have stepped forward, blaming a drastic change in corporate culture. Recently, they were joined by François Piednoël, who left Intel in 2017 after serving as a principal engineer and performance architect for twenty years. On 4 August, Piednoël posted a Youtube video called “How to fix Intel,” in which he blames a management culture that favors MBAs over technical expertise

for making bad technical decisions over the past few years. The next question, then, is how a company that built its success on flawless execution could drop the ball so carelessly? To understand this, we need to travel back to the nineties, when Intel started a symbiosis with Microsoft that was so successful that it got its own name: Wintel. Consumers only wanted Microsoft’s Windows operating system on their new PCs, so they could keep running the software they already had, and Windows PCs only ran on X86 processors. This allowed Intel to fab a seemingly endless number of ever more powerful chips for ever more powerful Windows PCs. Without any competition to speak of, both companies had a lot of pricing power. Intel, in particular, did everything it could to hold on to its half of the duopoly. Dictating the terms in the computer industry, it squeezed PC builders to fatten its own margins. This also had a strategic side effect, namely weakening potential competitors like Nvidia or Ati, which ended up making up the difference at the PC OEM. Other potential threats were attacked in court by the dozens. “At one point, the general counsel who headed Intel’s legal department was told that one of the targets he would have to meet in order to get a good performance appraisal was a fixed number of new law4

ANALYSIS CHIPS

suits to start each quarter,” Jackson wrote in “Inside Intel.” Thus, Intel became a company that enjoyed the unusual combination of high volumes and high margins. Such a business proved highly addictive, gradually turning the behemoth into a spreadsheetmanaged one-trick pony. Ironically, this is exactly what Andy Grove (1936-2016), Intel CEO from 1987 to 2004, had always been most afraid of. “Success breeds complacency. Complacency breeds failure. Only the paranoid survive,” stated the man who more than anyone built Intel’s culture of drive, focus, execution and innovation.

Missteps

The Wintel machine was still burning at full speed when Steve Jobs approached Intel to manufacture the processor for the first Iphone. Intel, unsurprisingly, refused. Concerns about the unproven nature of the product aside, it wasn’t interested in the low-margin business of manufacturing lowly phone chips. Around the same time, in 2006, Intel also sold its Arm-based Xscale handheld and cell phone chip business to Marvell. Intel, of course, later did try its luck again in mobile with the Atom processor. The attempt was doomed to fail. After all, how could Intel release a low-margin, low-power and high-performance product without

cannibalizing its own cherished high-end parts? And why would any smartphone manufacturer, very much aware of Intel’s instincts to subjugate entire ecosystems, be inclined to buy chips when there’s a competitive Arm ecosystem to take advantage of? What’s true for mobile, is true for many other markets: the Wintel legacy has plagued – and keeps on plaguing – Intel. Either a new product would cannibalize on existing business or it would require out-of-the-box thinking, both in terms of business and of technology. Intel hasn’t been capable of either for a long time, as is illustrated by a remarkable long string of failures in new markets over the years. DSP, communication and mobile, but also server farms avant-la-lettre, smart toys, TVs, wearables and, most recently, 5G modems – all spectacular, often very expensive, missteps. Perhaps Intel can eventually find another high-margin cash cow, but having dropped the ball in process technology, it will never again be able to dictate terms. It would be foolish, however, to write off a company that has stared into the abyss before: in the 80s, Intel successfully pivoted from being essentially a memory maker to a microprocessor company within a few years. To flourish again, it will similarly have to radically rethink how to maneuver a drastically changed semiconductor landscape.

Intel smartphone reference design Credit: Intel Free Press

Vlakbodem 10

3247 CP Dirksland

the Netherlands

+31 187 602 744

www.tbp.nl

info@tbp.nl

tbp also supplies the smart farming industry electronics manufacturing services

ANALYSIS SEMICON

US-Chinese trade war leaves semiconductor industry in limbo As the US government implements one export restriction after another, the semiconductor industry is left wondering where it will end. Paul van Gerven

“T

he health and vitality of the US semiconductor industry are essential to America’s future competitiveness. We cannot allow it to be jeopardized by unfair trading practices,” said President Ronald Reagan when announcing import tariffs for Japanese semiconductors. The move was about more than unfair trade, though, as the New York Times explained: “The tiny slivers of silicon that are the essence of computers and other electronic products, are considered vital to national security.” That was 1987, and history is repeating itself. Deeply concerned about China’s rapid technological advance, as well as its own incomplete semiconductor manufacturing base, the US has engaged in another semiconductor trade war. But this time it doesn’t look like it will peter out in a few years.

Watertight

The US has been particularly wary of two major Chinese tech firms, Huawei and ZTE, for a long time. Already in 2012, the House Intelligence Committee deemed the companies a national security threat. An investigation into ZTE for dealings with Iran and North Korea, initiated by the Obama administration, resulted in the telecom company being blacklisted in 2018. Surprisingly, however, the sanctions were lifted swiftly, in exchange for a hefty fine and putting in place mechanisms to assure compliance with US trade restrictions. But from then on, the US has worked steadily to tighten the net around Huawei without any backpedaling. The Commerce Department placed the telecom giant and its affiliates on the “Entity List” in 2019, which, among other things, made it illegal to supply US technology and software to Huawei. This doesn’t affect American companies only: any

company using US technology in their products needs to comply or risk repercussions. To the annoyance of the Trump administration, however, this barely slowed down Huawei. Due to the fact that the rule exempted foreign-produced goods from US export controls if the goods contained less than 25 percent US-origin exportcontrolled technology, many companies could keep supplying Huawei. Chief among them: TSMC, on which Huawei depends to manufacture its advanced chip designs. And so the US moved ahead recently and created a special rule scrapping the 25 percent condition for Huawei and other blacklisted entities. Since TSMC depends on US semiconductor equipment and software for its manufacturing operations, this means the US government can bar TSMC from supplying Huawei. Or at least, that was the intention – it turned out that the updated regulations still weren’t watertight. TSMC has indicated it won’t exploit remaining loopholes, however.

Unending succession

Lined up next, starting 29 June, is an expansion of existing export restrictions to cover all military end-users in China, which, according to US definitions, includes private firms. The new rule requires companies with US technology to screen their customers for ties with the Chinese military and present their case to US regulators. The latter have provided little to no guidance on what will and what won’t be allowed, however, leaving companies guessing on how to deal with it. More of this is sure to come: the US is undoubtedly adamant about thwarting China’s technological ambitions. The way it’s going about it, implementing one (often hazy) restriction after the other, makes industry very nervous. What’s allowed today, could be illegal tomorrow. So far, effects have started to ripple through the ecosystem, but on a systemic level, the damage has been limited. Few will be reassured that it will remain that way. 4 11

NEWS BATTERIES

Dutch silicon anodes inch closer to the gigafactory Nanotextured silicon anodes perform even better than expected, boosting battery energy density to world-record levels, startup Leydenjar found. Now the Dutch company has set out to develop highthroughput deposition equipment to take its superior anodes into gigafactories. Paul van Gerven

Credit: Leydenjar

eydenjar Technologies has developed a lithium-ion battery anode that’s made entirely from silicon. The spinout of applied research institute TNO has shown that swapping carbon for its silicon cousin results in batteries with up to 70 percent higher energy density. Preparations are underway to massively scale up the production of the anode. Silicon is an excellent host for lithium ions but in itself, not a good anode material because of its propensity to crack under the stress of repeatedly taking in and letting go of guests. Leydenjar adopted a plasma-enhanced chemical vapor deposition (PECVD) process originally developed for thin-film solar cells to create nanotextured silicon that can comfortably accommodate the volume changes associated with lithium loading and unloading. Last year, the Leiden-headquartered company proved that its process could make commercially relevant silicon anodes. Since then, it has been exploring the potential of its invention by making real batteries in various configurations. “The performance is well over what we thought was possible,” says Leydenjar CEO Christian Rood. “We assumed that swapping graphite for our silicon would result in a 50 percent increase in energy density, but it turns out we can reach 70 percent, achieving the highest energy for lithium-ion batteries in the world. Using the very best cathode materials available right now, we believe we can go even higher.” Another advantage of the silicon anodes is that their single-step production is more efficient than making traditional graphite-based anodes. It requires less material and mining and processing graphite is rather energy intensive, as is the heating step that follows a slurry-based coating of

the material. Leydenjar estimates that all things being equal, silicon anodes result in 62 percent less CO2 emissions.

The right formula

The batteries, however, still feature a relatively short life span. “Currently, our cycle life is only appropriate for some specialty applications, such as e-flight, medical devices or robotics. In these markets, a life span of multiple years isn’t important. We recently started working with industry partners to work towards commercial solutions in these domains,” says Rood. The comparatively limited life span isn’t surprising, as Leydenjar hasn’t spent much effort on improving it yet. Key will be identifying the right electrolyte composition – a delicate process that involves a bit of trial and error. Leydenjar recruited the help of German specialists to find the perfect formula.

Gigafactories

Meanwhile, Leydenjar has also been working on the manufacturing aspects of the

anodes. This is actually the startup’s core business, as it won’t be selling anodes or batteries but deposition equipment along with the know-how. The in-house battery research is nonetheless a necessity because not a single sale would be made without proof that the anode enhances performance. Another hard requirement will be to seamlessly fit into current battery production processes, without adding cost. Progress in the manufacturing department has been good too, Rood reveals. “We started building our pilot production line in Eindhoven last year. After installing and fine-tuning the machine originally used by TNO for the solar cell research, we now have a stable roll-to-roll deposition process (the silicon is deposited on copper foil, PvG). We’re getting sample orders from companies that want to test our technology. We’re also getting paid to make commercial prototypes.” The next step will be to construct a modular deposition machine optimized for production. “We have the expertise to do that ourselves, though obviously, we’ll work with partners, both Dutch and German. Only when we’re ready to ramp up, will we likely outsource production.” Though battery manufacturing is currently predominantly an Asian affair, Leydenjar’s strategy is mainly focused on Europe – and not just because the company received funding from the European Union. Determined not to miss out on another key technology, the EU has set ambitious battery plans in motion, which line up well with Leydenjar’s roadmap. Rood: “There are around 15 gigafactories planned in Europe over the next years, all of which will need a competitive edge. We’re convinced we can provide that, so this is an enormous opportunity for us.”

NEWS BATTERIES

Holst betting big on the foil battery

it:

ls Ho

en tC

Can billions of micropillars make for a better battery than we currently have? Yes, Holst Centre believes, after the ﬁnal piece of the puzzle fell into place in the lab recently. Now, the institute is looking to scale up and commercialize the technology. Paul van Gerven

battery can always do better. It can be cheaper, it can hold more energy, it can charge faster or it can have a longer life span. So, naturally, battery makers are constantly looking for ways to improve their product and gain an edge over the competition. No doubt, a lot more performance can be squeezed from the familiar electrochemical cell. Still, engineers at Holst Centre feel that in the long run, a transition to a radically different design will get us much closer to the perfect battery. Having made promising strides in the lab recently, the research institute found funding to prove its concept works on a larger scale and can be mass-produced. The futuristic battery Holst Centre has been working on for about five years already is called 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, consider each pillar to be a tiny battery in itself. The main challenge is to make enough pillars – billions and billions of them – so that together, they make up a full-fledged battery. This design has several advantages. First, compared to regular batteries, the lithium ions need to travel only very short distances, translating into faster charging and de-charging times. Secondly, there’s no liquid electrolyte involved, meaning a longer life span and little to no danger of

fires or explosions. And thirdly, the design is inherently lightweight. The Dutch Ministry of Economic Affairs & Climate Policy and the province of North Brabant, home of Eindhoven-based Holst Centre, are sold. They’re providing the research institute with 3 and 1.5 million euros, respectively, to set up a small-scale production line to demonstrate the viability and commercial potential of the technology. Not waiting for the results to come in, Holst Centre will go ahead and start a company to commercialize the technology. “The battery world is highly competitive. We simply can’t afford to go one step at a time if we want to bring this to the market, eventually,” explains Holst Centre director Ton van Mol.

Troublemaking

Since the patents haven’t been filed yet, Van Mol can’t go into the details of how the battery-on-foil is made. But, in broad strokes, first, an array of micropillars, each measuring a few microns in diameter and 100-200 micron in height, is grown on the foil. Next, they’re coated with the layers of battery materials using atomic layer deposition (ALD). This deposition technique is excellent at growing well-defined (multi-) layers on complex structures. However, in its traditional form, it’s batch based and hence not ideal for high-throughput production. Fortunately, this is one area of specialty of the Eindhoven region.

ALD, in general, involves sequentially exposing a substrate to different gases, which react with the surface in a self-limiting fashion, thus ‘coating’ it, literally, atomic layer by atomic layer. In traditional (temporal) ALD, a stationary substrate is placed in a reactor vessel, which is repetitively filled and emptied. Over a decade ago, however, Holst Centre’s parent organization TNO (and, separately, the company Levitech) developed a continuous ALD process, in which the substrate moves past the different gasses. This so-called spatial ALD (SALD) achieves much higher deposition rates and is compatible with high-throughput techniques like roll-to-roll processing. TNO originally developed SALD for solar cell manufacturing, but over the past few years, it has been working on expanding its range of applications. TNO spin-off Saldtech is now working on OLED displays and SALD (a company with roots in TNO as well) is exploring a number of applications, including batteries – ‘traditional’ batteries, that is. Holst Centre previously upscaled SALD to large-area roll-to-roll processing for a variety of materials. Still, it had to find out if it could also deposit the necessary thinfilm battery materials using the technique. The final piece of the puzzle recently fell into place when SALD deposition of lithium phosphorus oxynitride (LiPON) was demonstrated. This material serves as the solid-state electrolyte in the 3D solid-state 4 13

t re

NEWS BATTERIES

Credit: Holst Centre

thin-film batteries but could also be applied in traditional batteries. After all, the battery industry would gladly get rid of the troublemaking liquid electrolyte as well.

Heavily subsidized

Holst Centre is currently in the process of drawing up the specifications of the production equipment. After the tools have been manufactured and installed, the search for a viable production approach for thin-film batteries can begin. All in all, Van Mol estimates Holst Centre will need one and a half to two years to deliver the first working batteries. Soon afterward, Van Mol hopes to start working with companies to show the technology can deliver on their requirements. “We consider wearables to be a good entry market, but our calculations show our technology can be applied in almost every battery market, even electric vehicles,” says Van Mol.

It will be a rocky road, though, Van Mol admits. “We’re working at the cutting edge of what’s technologically possible, so success is by no means guaranteed. But, in my opinion, betting on a next-generation

technology like this is a much better way for Europe to catch up in battery manufacturing, rather than heavily subsidizing factories that are practically the same as those already operating in Asia.”

The future is closer than you think You want your company to stand out, but how? You’ll need specialists who can share their thoughts at every level. With creativity, dedication, and lots of knowhow, NTS helps its customers find the ultimate solution to every problem. We specialize in developing, manufacturing, and assembling (opto-)mechatronic systems, mechanical modules, and critical components. Our expertise? Precision and maneuverability. From the initial design to the prototype, all the way to assembly — our comprehensive support helps customers realize their ideal product faster. NTS excels in solving complex issues. Leave it to us to design and create an application that’s exactly what our customer was looking for. With our versatile knowledge and wealth of experience, we help accelerate technological innovations. The technology of the future? With NTS, it’s at your fingertips.

nts-group.nl/career

Our (opto-)mechatronic systems and mechanical modules contribute to future technologies

pinion

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

Ask the headhunter V.C. asks: After an international sales career of 20 years, of which 15 years in the semiconductor industry, I settled again in my hometown in the Netherlands. There’s no place like home! We moved there at the beginning of this year, just before the Covid-19 pandemic broke out. Since then, I’ve been working remotely for my employer. No one goes to the office anyway and traveling to customers is very limited. As I don’t think this will be sustainable much longer, I’ve agreed with my CEO to look for a position outside the company. Following my MBA 20 years ago, I switched from a sales job in the Netherlands to a position as a sales and marketing executive at an international distributor of electronic components in Italy. After a few years, I came into contact with my current employer, a medium-sized semiconductor company with (at the time) a less thriving position in a few niche markets. I started there as an account manager and have grown into a European sales director and member of the board of directors. To continue my growth in this company, I would have to move to the US, but – as said – I prefer to work in the Netherlands. Problem is that my business network here has become less strong due to my long stay abroad. I also know very little about the local chip scene. Recently, I saw a vacancy for a

marketing manager at a large semiconductor company in the Netherlands. I’m thinking of applying but I’m hesitant: the position is below my level and I cannot leave confiden-

Try to build a network first tial information about my current employer in the hiring company’s general application system. Yet, I’d like to get in touch. How do you compare the vacancy to my current position? Could it help me progress my career? Do you happen to have a contact there or can you mediate in a first step?

The headhunter answers: When you’ve been away from the Netherlands for so long and hardly have any contacts in the semiconductor industry here, it doesn’t hurt to start with scrutinizing the field. You can of course apply for this position, which is indeed far below your level, but that might harm your interests in the long run. Each position is to some extent career furthering, but they want to fill the vacancy now and not start again in 6 months. My advice to you is to try to build a network first. There are various ‘platforms’, such as fairs, organizations and lectures, you can use to

make valuable contacts. You can also choose to directly reach out to the management of potential employers to explore opportunities. You’ll understand that this takes time and perhaps some perseverance. It isn’t clear to me what confidential information about your employer you’re referring to. When you want to get a managerial commercial position elsewhere, you inevitably need to show something of your market knowledge, know-how and added value. Evidently, you don’t share turnover figures and details from commercial agreements, but you don’t have to be secretive about what can be gathered from public sources, such as products, markets and distribution channels. It’s good to show your loyalty to your present employer and your sensitivity to the subject, but too much restraint will hinder your opportunities.

4 15

NEWS SENSORS

Surveilling aging infrastructure using fiber-optics Somni Solutions, a small startup from The Hague, provided fiber-optic tilt sensors for the new Genoa bridge, which recently reopened, two years after the tragic collapse. CTO Remco Nieuwland explains how his company acquired this prestigious assignment and why we should start monitoring our own aging infrastructure. Jessica Vermeer

Credit: Somni

n 14 August 2018, the Morandi bridge near the Italian city of Genoa collapsed during a rainstorm. When it was opened in 1967, the number of vehicles and their average weight was much lower than it is today. Bridges were typically designed for a 50-year lifespan. The Morandi bridge failed just under 51 years after its opening. After the collapse, the leftovers were demolished and a new bridge was built. The construction of the replacement was completed in April of this year. Somni Solutions, a small startup from The Hague, provided glass fiber sensors that will monitor the misalignment of the pillars. These sensors will give much-needed insight into the condition of the new bridge. “Monitoring with glass fiber sensors is pretty standard in China and Japan,” says Remco Nieuwland, CTO of Somni. “But in Europe, we’re lagging.” He has seen some pilot projects but no permanent setups for bridge monitoring. “In Europe, we seem to need an actual disaster before we’re willing to prevent further damage.” When Somni started initial discussions with the customer, the sensor design wasn’t yet finished. Acquisition of the assignment took about a year and the startup spent that time sketching, developing and gaining the customer’s trust. After one year of hard work, the long-awaited purchase order finally fell in favor of Somni. Nieuwland: “It hasn’t been easy, but we were highly motivated to get this assignment and show the world our capabilities.”

The two sensors within each set are placed at a 90-degree angle.

Credit: Pergenova

The new Genoa bridge has eighteen pillars upon which the road surface lies. Each pillar has four tilt sensors, two up top and two at the bottom.

Fiber Bragg grating

Somni Solutions was founded in late 2017 by Nieuwland and Jac Gofers. Both recognized the enormous potential of fiber-optic sensors for Europe’s aging infrastructure. With a team of highly skilled professionals, they shaped Somni, a startup that designs, develops and manufactures fiber-optic sensors that monitor the structural integrity of large assets. The company name expresses the goal to solve sensor problems everywhere – the “s” standing for “sensors” and “omni” being Latin for “everywhere.” Somni uses standard telecommunication glass fiber for its sensors. “But we modify them,” explains Nieuwland. “Using lithography, we write a pattern on the inner fiber core over a few millimeters length.” He compares the pattern to a bar code. “When we slightly compress or stretch the fiber, the reflected light will have a slightly different color.” The microstructure Somni applies is called a fiber Bragg grating (FBG). It allows for the reflection of a specific wavelength depending on the periodicity of the grating. “We built a mechanism around the FBG – a transducer – that converts the parameter we want to measure into stretch or compression of the fiber.” For the Genoa bridge, that parameter is the tilt of the pillars. The stretch or compression is measured from the wavelength shift that results from the grating.

In total, Somni has delivered 72 sensors to Genoa. The new bridge has eighteen pillars upon which the road surface lies. Each pillar has four tilt sensors, two up top and two at the bottom. The two sensors within each set are placed at a 90-degree angle. Somni writes several FBG patterns in the glass fiber, all with a slightly different periodicity. Every grating reflects a slightly different color. “That means we can accommodate several sensors using one glass fiber cable, without the need for a return cable,” points Nieuwland out. “Depending on the spectral bandwidth of the light source and the dynamic range needed for one sensor, up to a hundred sensors can be built using just one cable.” “The best thing about these sensors is that they don’t work on electricity as many conventional sensors do,” says Nieuwland. “They work solely based on light.” Conventional electric sensors need at least two electric wires to read the information from one sensor. If such sensors are used to monitor a bridge, the data needs to be transported over a kilometer distance. “That’s quite complicated. Usually, the analog signal is converted into a digital signal, which requires a bunch of extra electrical components inside the sensor. And every component has a risk of failure. If one component fails, the entire sensor fails.”

Photonic sensors are simpler. They have several advantages, such as low maintenance and high reliability. Also, glass fiber is completely insensitive to electromagnetic radiation. Nieuwland: “It could even be used in an MRI scanner. And the data quality over long distances is excellent.” An added advantage – the light is sent into and reflected by the same cable. “That means we have built-in redundancy,” explains Nieuwland. “All we need to do is lay down a return cable. If the cable breaks somewhere, we can read the sensors from the other side.”

Merwedebrug

Soon after the Genoa sensors were successfully delivered, Somni received a new challenge. ITER, the nuclear fusion reactor in France, asked to supply sensors. This high-profile project required sensors that are robust and reliable, compatible with vacuum conditions and resistant to neutrons and gamma radiation. In addition, they need to operate at temperatures above 300 °C. The Somni team has taken up this challenge and is currently in the development phase. Installation is slated for the first months of 2021. Nieuwland hopes the success of the Genoa project will open up new possibilities. In the future, Somni aspires to collaborate with Rijkswaterstaat to monitor roads and bridges in the Netherlands, to prevent cases like the Merwedebrug near Gorinchem, which was closed for heavy freight traffic in October 2016 after measurements by TNO revealed hairline cracks in the construction. “Trucks are becoming heavier and no one really knows which arteries are worn out,” says Nieuwland. “If we monitor our infrastructure properly, we won’t be forced to close down an important thoroughfare like the Merwedebrug ever again.” 4 17

NEWS RF

Pink RF and America’s Odyssey link to bring RF repair to Nijmegen Dutch radio frequency specialist Pink RF is expanding operations. In collaboration with the American RF repair company Odyssey, European RF and semiconductor customers will have a new option located in Nijmegen.

he city of Nijmegen has a new hightech repair shop. The new business, which started operations in March of this year, is a collaboration between the Netherlands’ Pink RF and its new American collaborator, a world leader in the repair and service of radio frequency (RF), direct current (DC) and microwave devices, Odyssey Technical Solutions. Together, the companies will look to work with European wafer fabs, big and small, offering repair and maintenance services for RF and microwave equipment in the Dutch and European high-tech industry.

The right fit

Headquartered in Round Rock, Texas, Odyssey has a long-standing reputation in the radio frequency and microwave repair domain. The company has been repairing RF equipment used in semiconductor production for the last twenty years, having successfully repaired more than 60,000 devices for its US and Asian customers. With this success, Odyssey has had a keen interest in better serving the European market. For quite some time, the Texas-based company has been looking for a first cooperation partner in Europe. As the search for the right fit led them to the Netherlands, the decision was clear – Pink RF was its choice. “The new company creates a lot of synergies,” describes Klaus Werner, director and founder of Pink RF. “For Odyssey, we offer relevant expertise in the mindset of RF and semiconductor manufacturing. Odyssey, on the other hand, contributes its resources and an established business model within the domain. Furthermore, the collaboration will give Pink RF access to a large network of experts that can repair systems, including our own. This will allow us to really put our focus on our core strengths.” 18

Credit: Fotowerkt.nl

Collin Arocho

Place to be

The duo has opened its doors on the Novio Tech Campus (NTC) in Nijmegen, adjacent to the current offices of Pink RF. “Not only were our offices already established here, but the availability of and proximity to RF and microwave expertise and suppliers like NXP, Ampleon, Macom and Minicircuits is fundamental to Pink RF’s success,” says Werner. “The reality is that Nijmegen is the place to be for RF and semiconductors. Being located amid semiconductor manufacturing heritage, we have unrivaled access to contacts, customers, vendors and experienced talent.” According to Odyssey CEO Jim Plourde, the company was already performing repairs for European customers, but taking a foothold in Europe was badly needed. “Odyssey has been doing repairs for European customers for years, but logistically speaking, a location in Europe made sense as the company grew,” Plourde explains. “We’re essentially aligning our service locations with our global customers by establishing the new Nijmegen service center.” “Oost NL informed us about the semiconductor sector in the region,

helped settle administrative obligations and informed about the possibilities surrounding the Semicon fair in Munich. Then we were guided from start to finish by Kadans Science Partner and Novio Tech Campus in the campus location,” states Plourde. “This was the deciding factor for us to become part of the Nijmegen health and high-tech network.”

Ramping up

While the first few weeks were spent building out the new offices and setting up operational structures, the RF company is ready to hit the ground running. The shop has already received its first few orders for repair. Currently, operations consist of just three employees, however, the ambition is to grow to 15-25 workers, at the Nijmegen location, over the next three years. “As the newest state-of-the-art repair facility in Europe, we’re looking forward to ramping up our business,” highlights Werner. “We intend to forge further cooperation with companies close to semiconductor manufacturing businesses to enlarge our reach within Europe.”

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 thermal challenges for RF front-end designers

here are two main design challenges when it comes to thermal management in the Wi-Fi front-end. The first is the increased demand for smaller, sleeker routers, access points and wireless speakers that must be aesthetically pleasing. Consumers are migrating away from one wireless router per home and toward a mesh-networked home, driving the need for a smaller and less obtrusive product. Additionally, Wi-Fi is designed into set-top boxes, speakers and voice assistant devices that are also becoming smaller and sleeker. Smaller and more pleasing to the eye is good for the consumer, but it creates additional pressure on the design as the devices inside have less space to properly dissipate the heat they create. The second challenge is in the enterprise market, where the power source for Wi-Fi products is in the Ethernet connection. This Power over Ethernet (POE) connection is limited, so the challenge for product designers is to maximize the RF output power of the RF front-end with a limited POE. Efficient power dissipation is key. Power dissipation is defined as the amount of power consumed and converted to heat. Electronic devices produce heat, as an unwanted byproduct, which is a waste of energy. The desired state of any RF circuit is to reduce this wasteful heat and provide extensive RF output power, system efficiency and RF signal range. This means RF front-end designers must create products that function using a low power source but produce high RF output power. And they also need to ensure their products effi-

ciently remove the unavoidable excess heat created, to maximize RF output and reduce unwanted cooling fans or bulky heat sinks. These fans and heat sinks hinder the manufacturer’s ability to meet the sleek, small, aesthetically pleasing product criteria. To solve the thermal heat/dissipation challenge, an innovative plan of attack is needed, one that focuses on efficiency, low current consumption and maximum power output. This de-

To solve the thermal challenge, an innovative plan of attack is needed sign approach can help reduce thermal dissipation by 25-50 percent per RF stream while maintaining output power requirements and RF range. Wi-Fi 5 and Wi-Fi 6 routers and access points with multiple RF multiple-input multiple-output (MIMO) streams are typically subjected to average temperatures of 60 °C or higher, even when the room is a moderate 25 °C. Additionally, more functionality and more data throughput in a smaller product footprint all contribute to the same problem: more heat. Fortunately, there are some innovative 2.4 and 5 GHz products in the market that address both Wi-Fi 5 and Wi-Fi 6 solutions with multiple RF streams and voltage requirements. These RF front-end (RFFE) products address the thermal and power dissi-

pation challenge, but they also meet the small form factor requirement. They help enable a broad range of end-products and market sectors, such as retail, operator, enterprise and consumer, in a form factor that’s acceptable to customer requirements. Heat can degrade overall system performance, impacting throughput, range and the ability to prevent interference. So, when designing WiFi systems, it’s important to choose RFFE components that mitigate heat-related problems. Product designers should also consider using fully optimized, integrated front-end modules (FEMs) instead of discrete front-end components. These modules reduce line lengths and the need for tedious tuning of the design, which contributes to insertion loss and system heat. By choosing RFFE modules that provide a complete solution – that meet the stringent thermal and RF range requirements and integrate filter, PA, LNA, switch, detector and coupling in one package – designers can remove the expense and tedious task of piecing these individual components together. Instead, the design process can be streamlined, costs reduced, products certified faster, and most importantly: time to market is reduced and stringent schedules can be met.

4 19

YOUNG INVENTORS IMPROVE THE WORLD, EVEN IN A PANDEMIC To inspire youth for a future in tech, The Inventors Foundation every year organizes the Inventors Competition for all children in primary schools. This year’s final event was a special Covid-19-proof edition and showed that youngsters can inspire grownup inventors as well!

nventors Competition In the Inventors Competition, children can post their ideas, solutions and inventions on a special website (deuitvinders.com). They’re encouraged to recruit online votes, which addresses their talent for entrepreneurship. The best inventors are invited to the finals, where they can pitch their ideas for a large audience and a jury. Due to the Covid-19 pandemic, this year’s finals at the Discovery Factory in Eindhoven could only host 30 persons. All young inventors were able to do a ‘normal’ presentation on a real stage, facing the crowd, with a large projection on the wall. Through a live connection, their schoolmates could still cheer for them. Tech rules! Guitar fire extinguisher Although all finalists deserved a prize, the jury had the harsh task of selecting the winners. The award for the most original invention went to 11-year-old Willemijn Heijman. She developed a “guitar fire extinguisher” for firefighters in love. With this invention, they can bring their sweetheart a serenade and – paradoxically – at the same time water squirts out of the guitar neck, extinguishing all starting fires. Jury member Nicol van Hoof, owner of metal company Van Hoof Group, praised this love-spreading idea when handing over the award.

Anti-rain traffic light According to the jury, the best inventor was 10-year-old Teun Roux. He invented the “antirain traffic light”: in case of rain, a sensor on the traffic light makes sure that cyclists are prioritized over cars, preventing them from becoming soaking wet. Teun had built a working prototype, which he demonstrated on stage, greatly impressing the jury and audience. Jury member and alderman Stijn Steenbakkers (City of Eindhoven) promised to have implementation of this solution seriously researched by the responsible department of the municipality. Inventor virus The Inventors Competition and the finals demonstrated that using creativity, technical solutions really can make the world a better place. Thanks to the young inventors of course. But also thanks to the partners of The Inventors Foundation and Discovery Factory, without whom the competition wouldn’t exist. If your company wants to participate as well or even set up an Inventors project of its own to inspire youngsters, contact The Discovery Factory in Eindhoven. In a partnership anything is possible. The inventor virus is stronger than any other virus!

The Discovery Factory is 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, NTS Group, Philips, Stam en De Koning and VDL Group, and by Bits&Chips as the media partner.

discoveryfactory.nl

NEWS SEMICON

Pandemic accelerates ASML’s adoption of AR Travel restrictions due to the coronavirus pandemic opened a door for ASML that had been ﬁrmly closed: using augmented reality to troubleshoot complex issues at customers’ fabs remotely. Now the company sees plenty more opportunities to take advantage of the emerging technology across the entire organization. Paul van Gerven

bout three months ago, an ASML team installing an EUV scanner in a Taiwanese fab ran into trouble. There was a pressure problem they couldn’t wrap their heads around. Normally, this means an expert is flown in posthaste. However, with corona quarantine measures in place, it would take him at least two weeks before he could get to work, seriously jeopardizing the customer’s production planning. To save the day, a local engineer put on a pair of hastily couriered Microsoft Hololens 2 mixedreality smart glasses (see sidebar), allowing the expert in San Diego to get a view on what the engineer was seeing and guide him through the right steps to fix the problem. It wasn’t as easy as that, of course. Though ASML had been experimenting with virtual and augmented reality technology for a while, it wasn’t quite prepared to perform a remote assistance operation like that. Why would it be? The thought had been entertained more than once, but the 22

time was never invested because there was no way customers would allow it. Taking a camera to the heart of an IC manufacturing operation? Unthinkable. Until the coronavirus reared its ugly head, that is. Faced with installation delays and idle scanners, chipmakers quickly set aside their objections. In fact, one of them suggested the option, says Peter Peusens, director of ASML’s DUV Customer Support operations. “Not long after travel restrictions were put in place, someone working at a major customer of ours sent a Youtube video

The Hololens 2

about the Hololens to our service team. That e-mail ended up on my desk, with the request to see if I could look into it.” And so started a frantic operation across the organization to add augmented reality and related technologies to the toolkit of ASML Customer Support. After several successful service actions, hopes are high that they will be a permanent addition. “Over the past few months, we’ve been making more progress in this area than we have over the past few years. We’ll work very hard to expand on that, even when corona restrictions

Microsoft’s Hololens 2 is a mixed-reality headset developed for industrial applications. Its base functionality is live streaming of whatever the wearer is looking at. Viewers can add information to the wearer’s view, ranging from drawing a simple arrow to draw attention to a specific item, to projecting animations that show how to perform a certain action. In the future, real-time data may also be visible in the lens, for example showing the pressure value of a subsystem.

Credit: ASML

are lifted,” explains Michiel Haverkorn, director of Customer Support at ASML.

Bridge the gap

ASML scanners are well taken care of to optimize their output. Day-to-day operations and maintenance are handled by teams working 24/7 in shifts. Should a problem arise that transcends their expertise, a call for help is placed to the local support office. The vast majority of issues are taken care of by these two support tiers. But once or twice a day, something pops up that requires the attention of ASML’s Development & Engineering (D&E) department, staffed with people that know the systems inside out. These issues that escalate all the way to D&E are the prime use cases for augmented reality (AR) support, mainly because of the drastically reduced response time. After all, the right man for the job could be involved within the hour. But beyond the business perspective, the implementation of AR technology improves the work-life balance for D&E engineers. Even without quarantine requirements, not having to rush to a customer site on the other side of the world will be very much appreciated. At the moment, another excellent use case is EUV scanner installation, notes Peusens. “This is a relatively new technology, and we haven’t finished documenting all the procedures involved. Nonetheless, our customers are pushing to get their systems online. In such a situation, even though it’s standard procedure to have D&E engineers present at EUV installations, complex issues are bound to arise. D&E has been getting a lot more requests for assistance in the EUV domain recently.” Haverkorn adds: “As these systems and their installation haven’t yet been fully industrialized, it’s simply impossible to anticipate what kind of expertise might be needed. AR can help to bridge that gap.” Peusens and Haverkorn stress that despite the great potential, AR will always remain a tool that cannot replace the expertise and skill of people. “Especially in the field, AR only works in conjunction with well-trained staff with good hands. You can’t simply have an inexperienced engineer put on a Hololens and expect him to do what normally takes years of training,” says Peusens.

Years of training

ASML has started using AR internally as well. For example, whenever CS updates work instructions, these are verified by physically entering the factory in Veldhoven and do a test run. Currently, only essential personnel is allowed in there, however. The Hololens proved to be an excellent alternative. Haverkorn: “I’d say it worked even better than our old procedure. Even if our CS engineer stands right next to the person carrying out the instructions, he can’t see through his eyes. With AR, he can do this from the comfort of home without having to deal with the procedures to enter and exit cleanrooms.”

“Now that people have heard about the first examples, we’re getting daily inquiries from all corners of the company,” continues Haverkorn. “And we do see a lot more potential, of course: using virtual meetings of design teams located in different parts of the world instead of organizing review sessions in person. Training new engineers, streamlining our collaboration with suppliers, or remote customer acceptance releases of new systems – the possibilities are endless, especially as the technology evolves, but we need to prioritize right now. We can’t do it all.”

4 23

ORDER NOW

“ASML’s Architects is an impressive book, a curious book and a book for the curious. (…) Clearly a labour of love by Raaijmakers but nonetheless an easy read.” Peter Clarke, eeNews, February 1, 2019 “Rene Raaijmakers’ book on the history of ASML is a monumental work in its depth and breadth from ASML’s beginning through 1996. (…) No tech company’s history has ever been covered to such a degree.” Dan Hutcheson, The Chips Insider, February 1, 2019

techwatchbooks.nl/architects

NEWS IMAGING

Performing an ultrasound on a wafer to align it The markers used as ‘beacons’ to precisely position wafers during semiconductor manufacturing may become hard to find after several chip layers have been deposited on them. Researchers at ARCNL may have found a way to spot them: using acoustic waves. Paul van Gerven

Credit: ARCNL

sing very high-frequency sound waves, researchers at ARCNL have found a way to detect nanostructures buried under many layers of opaque material. Their findings may be useful to spot grating lines used for wafer alignment during semiconductor manufacturing. As these indispensable markers get buried deeper and deeper under layers of materials during the manufacturing process, they become harder and harder to spot with the technique that’s normally used: light. Fortunately, many materials that are opaque to light do pass on sound waves. So similar to performing an ultrasound, the ARCNL researchers sent sound waves into layers of materials you might find on a wafer, stacked on top of a grating. Actually, they shot short laser pulses at it, knowing this would induce high-frequency sound waves in the opaque material. The frequency of the waves is much higher than those used in medical echoes – the higher the frequency, the smaller the features

that can be discerned. Obviously, for finding nanostructures, very high-frequency waves are necessary. The big question was whether the waves would reach the grating in the first place. “I was a bit skeptical in the beginning because the sound waves have to travel through so many layers of dielectric material before they reach the grating buried inside. If they reflect at all these interfaces, we would have ended up with a complete mess of sound waves. But it turned out that the stack of thin dielectric layers acts as one thick layer because the individual layers are thinner than the wavelength of the sound wave. So the sound waves travel straight to the buried grating lines that we want to see,” says ARCNL group leader Paul Planken.

Limits

Having traveled through the opaque material, the sound reflects at the grating. Since the grating is not a flat surface but has periodic valleys and peaks, the sound

from the valleys reaches the surface slightly later than the sound from the peaks. The sound wave causes a very small displacement of the atoms when it reaches the surface, causing a copy of the grating to appear there. This pattern can be detected using a second laser pulse. Now that they’ve shown that it’s possible to detect nanostructures buried under opaque material, the researchers are going to further investigate their method. Planken: “Our results not only reveal interesting features in photo-acoustics that haven’t been investigated before, but also offer a promising solution for practical issues in nanolithography. For industrial applications, we should optimize the system to get signals that are stronger, faster and more robust. But we also want to increase our understanding of all the effects that we see in the signal, and find the limits of our method, for example by trying to discern a grating with lines that are very close to each other.” A femtosecond pump laser shot at the opaque material (1) causes a high-frequency acoustic wave to travel through the layers (2) until it reaches the buried grating lines. The acoustic waves are reflected at the grating and travel back (3) as a grating-shaped wave. When this wave hits the surface (4), the grating-shaped deformation can be detected from the diffraction signal of a femtosecond probe laser.

4 25

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-dept technical texts, from blogs and newsletter texts to background articles and case studies. We write in both Dutch and English (native) and we charge per hour. You will receive full ownership of the text and you can choose how to distribute it.

bits-chips.nl/content-services

For publication in Bits&Chips, we offer a reduced publication fee. Contact us via sales@techwatch.nl for more information.

pinion

HEALTHCARE Wim Hendriksen is retired.

Engineering the corona response

n the first half of 2020, our complex society came to a complete worldwide halt by a simple string of RNA called the coronavirus. How could we have let this happen? Perhaps we lost track of some fundamental engineering principles. For sure, we could use them to try to manage the aftermath. Engineering Principle #1: define what you want to achieve. The first thing to decide when starting a project is to define what the goals are: what parameter or cost function do you want to minimize? Most countries saw what happened in Wuhan, where intensive care units risked being overrun by too many patients. Therefore, the Dutch government decided to try and minimize the number of people who need hospitalization. This was a clear goal, which was communicated crisply to all citizens. People understood what the problem was and what they could do to reach this goal: stay at home and wash your hands. The approach worked and the ICUs were not overrun: goal achieved. Of course, five thousand Dutch who died a horrible death and a tanked economy prevent us from calling this approach a success. When the number of people in hospitals went down after the peak, our government didn’t define a single new parameter to minimize, so everybody invented their own criteria, resulting in a lot of noise and sloppy behavior of the citizens. Bartenders wanted the same social distancing rules as airliners, people wanted to socialize more, protest marches were organized – we wanted to restart the economy too soon. This all will result in a revival of the virus and more un-

necessary grieve in the second half of 2020. This will not be a success. Engineering Principle #2: make a model of the process you want to control, based on what you see in the real world. Lots of scientists started to evaluate the sparse and noisy data, tried to find correlations and under-

Statistics is a blind spot for most people stand the nature of the virus, the way it spreads and what it does to the human body. Meanwhile, most of us stayed at home and tried to keep the reproduction factor below 1. But after a few months, several nitwits surfaced, claiming to know all there is to know. They started to add white noise. For me as an engineer, it was a relief to see that scientists kept at it, while the nitwits were effectively silenced by the press. The signal-to-noise ratio remained at a healthy 50 dB. Engineering Principle #3: understand the math. The virus spreads from person to person. When one person infects more than one, the number of infected people grows exponentially – the infections double every N days. The human brain isn’t good at handling exponential growth; we’re used to more or less linear extrapolation, so it feels counterintuitive – things don’t look very dangerous when the numbers are low. Fortunately, Maarten Steinbuch has taught us in several Bits&Chips

keynote talks how exponential systems behave for artificial intelligence. Check “The AI revolution: the road to superintelligence” for a mind-changing introduction. The math also works for coronaviruses, and the pitfalls are the same! Statistics is a blind spot for most people. The concept of mean is roughly understood, but few people grasp the concept of probability distribution and standard deviation. When the number of deaths decreased three days in a row, a certain president shouted that from now on, the virus will disappear by itself. Engineering principle #4: what you don’t measure, you can’t control. When you’ve devised a model, you can build a control system of your process by feeding back relevant measurements into your control loop. But there’s a pitfall: lag. When you drive a car with a blinded front window, using a video of the road from last week, everybody understands that you have to be very careful. Currently, we face a virus with an incubation period of six days and a test system that yields results after a few days. So the lag in the control system for the government is about two weeks: it takes two weeks before you measure the response of your control system. Thus, the corona dashboard shows the results of previous control actions. Good luck trying to develop an optimal control strategy from these figures!

4 27

THEME SOFTWARE ENGINEERING

COMMA INTERFACES OPEN THE DOOR TO RELIABLE HIGH-TECH SYSTEMS Once a research project, initiated by ESI (TNO) and Philips, the Comma framework is developing into a mature product for creating and managing software interfaces. Now, Thales is also looking to use it to streamline its software engineering, as are Thermo Fisher Scientiﬁc and Kulicke & Soffa. “Comma is the place where you express everything you want and from there, you generate everything you need, like documentation, monitoring, simulation, visualization and, as of recently, test cases.” Nieke Roos

“O

ur medical devices are growing bigger and bigger,” observes Daan van der Munnik, software manager at Philips Healthcare in Best. “We have to chop them up in smaller subsystems to keep their development manageable, but also for validation purposes. Up to a year ago, we validated a complete device in one go – a huge effort. By chopping it up in smaller subsystems, we can focus our validation efforts on the parts of the system we actually touch for a particular feature. We do need to show that when we put everything together, it still does what it’s supposed to do. Both the disassembling and reassembling call for good interface management.” Subsystems are also increasingly being farmed out to subcontractors. “We’re really moving to a system-of-systems development, where we make some parts ourselves and some parts come from outside,” notes Van der Munnik. “For instance, in one of our image-guided therapy systems, we have three types of patient tables. One is developed by us, two are made by other companies. From a user perspective, however, they all have to appear to be an integral part of the system – the user experience, for instance when moving or tilting, has to be exactly the same.

Comma

Comma (Component Modeling and Analysis) is an ecosystem supporting model-based component engineering. It’s a combination of domain-specific languages (DSLs) in which the interface between a server and its clients can be specified by three main ingredients: the interface signature, the allowed client-server interactions and the time and data constraints. The interface signature consists of groups of commands, signals and asynchronous notifications. Commands are synchronous: the caller is blocked until a reply is received, whereas signals are asynchronous: they do not block the caller and do not require a reply. State machines are used to describe the allowed client-server interactions, such as the allowed order of client calls and the allowed notifications from the server in any state. Finally, Comma enables the definition of constraints such as the allowed response time, notification periodicity and data relationships between the parameters of subsequent calls. An example Comma model.

Camera.interface import "Camera.signature" interface ICamera version "3.14" machine camera { initial state Off {

}

transition trigger: PowerOn do: reply(Status::OK) next state: On OR do: reply(Status::Failed) next state: Off

state On { transition trigger: TakePicture(Time timestamp) do: reply(*) Click next state: On

}

transition trigger: PowerOff next state: Off

timing constraints TC1 in state On command TakePicture - [ 10.0 ms .. 75.0 ms ] -> notification Click TC2 in state Off command PowerOn and reply(Status::OK) -> [ .. 100.0 ms ] between events data constraints variables int nr1 int nr2 DC1 reply(nr1) to command TakePicture;reply(nr2) to command TakePicture where nr1 < nr2

Camera.signature import "Basic.types" signature ICamera commands Status PowerOn int TakePicture(Time timestamp) signals PowerOff notifications Click

Page 1 Basic.types type Time enum Status { OK Failed }

Credit: Philips

Thanks to Comma, it’s easier for Philips to enhance its medical systems.

This means that, for our subcontractors, the interfaces need to be clearly defined, both on a low technical level and a high subsystem level.” Last but certainly not least, good interface management is key for system evolvability. Van der Munnik: “Our medical devices have very long lifetimes. We need to ensure that, over their lifetime, they’re expandable and suitable for form/fit/ function replacements.”

Evolving interfaces

Fellow high-tech company Thales faces similar challenges. “Traditionally, we developed, built and qualified our combat management and radar systems, delivered them to the customer and mostly touched them to replace obsolete components – to avoid unnecessary risks, functional changes were rather limited and implemented at long intervals,” explains Pepijn Noltes, software architect at the Hengelo-based company. The last ten years, however, the operational scene is changing more rapidly at extended operational lifecycles, with customers increasingly demanding new features. Thales is adapting to this need by looking for ways to implement software updates more frequently, including incremental enhancements.

But that’s easier said than done. “For complex software-centric systems like ours, it’s very expensive to change something, integrate and test it – especially so in the military domain that we’re in, where you may have to do live firing trials to really validate the system,” says Noltes. “Also, even the tiniest update may cause an avalanche of changes. It then boils down to the question: how well can you revise part of your system without touching the rest?” Noltes has learned that to be able to continuously update a complex system, you need to keep the changes local and to do that, you need to focus on the interfacing. “We tend to touch the interfaces as little as possible because they’re expensive to change. Bigger problems that you can’t work around will eventually get fixed, but small issues will remain, as a result of which the code quality will slowly deteriorate. We’re now looking at evolving interfaces to facilitate the need for change.”

Single source of truth

Enter Comma (Component Modeling and Analysis), an ecosystem supporting model-based component engineering. “It started about six years ago as a research project between ESI and Philips,” recalls Jozef Hooman, senior

scientist at ESI, the high-tech embedded systems joint innovation center of the Netherlands Organization for Applied Scientific Research (TNO). “We began using domain-specific languages for all kinds of purposes, generating code, analysis tools and much more. While doing this, we noticed that a lot of issues Philips had with its software were due to interface problems and, gradually, the insight came to us that these DSLs were especially useful for describing the interfaces. So, in small steps, we moved from general-purpose languages to a domain-specific language, Comma, which we reused for many different interfaces.” Although a research project, the development of Comma wasn’t driven by research considerations, Hooman points out. “We really looked at what the engineers at Philips needed and adapted the language accordingly. We started with a state machine describing the interface protocol, ie the interaction between client and server. Based on user feedback, we modified it to make it more user friendly and include things like timing and data constraints. The patient table, for instance, is very sensitive to both timing and data – when the controlling joystick stops, the table should stop too within a certain amount of time and without moving too much.” 4 29

THEME SOFTWARE ENGINEERING Step by step, Comma developed into what it is now: “the single source of truth,” as Hooman calls it. “This DSL is the place where you express everything you want and from there, you generate everything you need, like documentation, monitoring, simulation, visualization and, as of recently, test cases. Monitoring, especially, is very important. You can use that to see if your implementation satisfies the specification by running the system, collecting traces and check whether the execution conforms to the interface. If an interface changes, you can re-generate everything, and if your developers, or your third-party suppliers for that matter, introduce a software update, you can check that for conformity – all with the push of a button, continuously, as an integral part of your test process.”

Forming an ecosystem

At Philips, Comma is now firmly embedded in the company’s software engineering practice. Van der Munnik:

“We use the DSL to write the interface specs and generate documentation and code. As part of our continuous integration pipeline, we check interface conformance against the Comma specs when executing our automated test scenarios. We’ve created a maturity matrix, which sets off our interfaces against these development stages, and we’re now raising the bar for all of them. Thanks to the unambiguous definition of interfaces and the subsequent automatic validation, Comma brings us a huge amount of business value as we find interface issues early, well before integration.” Two years ago, Thales started the Dynamics project to research dynamic system updates in collaboration with ESI. “We’re looking into evolvable interfaces and so-called adapters to keep the old and the new working together,” clarifies Noltes. “So when you introduce a client with an updated interface, you also generate an adapter that connects it to your existing server and provably ensures that nothing breaks

down. ESI did a small technology survey on interface specifications and Comma came out as the solution that best fits our needs. Although Dynamics is still ongoing research, Comma is already useable out of the box and we’re busy to include it in our component development framework. By doing more at design time, we hope to eliminate much of the risk in projects.” Slowly but surely, Comma is conquering the Dutch high tech. “We’re also working with Thermo Fisher Scientific in Eindhoven, for instance,” illustrates ESI’s Hooman. “For some critical interfaces, a model has been made and a monitor has been built into the nightly smoke tests, which automatically checks the log files. In the morning, they can see what properties have failed. And Kulicke & Soffa, also from Eindhoven, is looking into making a generator for its middleware layer.” Senior research fellow Benny Akesson, ESI’s liaison to the Dynamics project, adds: “It’s interesting to see this ecosystem starting to form.”

5 NOVEMBER 2020 • EINDHOVEN CALL FOR PRESENTATIONS We’re looking for high-quality presentations and we welcome development cases via softwarecentricsystems.com/call-for-presentations Powered by

SOFTWARECENTRICSYSTEMS.COM #BCSC2 30

Credit: Thales

Thales is looking to use Comma to make the interfaces in its software-centric systems evolvable.

Backward compatible

According to Akesson, there are basically three ways of using Comma. “This monitoring facility has already been there for years. If you have an interface and all the traces you run through the monitor are compliant, you know you’re in a good place. When you update your interface, you can automatically generate a new monitor and feed it the same traces to see whether they still work. As you don’t have to do a complete impact analysis of the change manually, that saves you time. The problem with this is that you’re now no better than your traces. You need to have traces that are representative of all the desired system behavior.” Together with its industrial partners, ESI is working on a solution based on so-called Petri nets, a formal method using state-transition models to study concurrent and distributed systems. “By generating Petri nets for an interface, you can see the possible state transitions that can occur in the protocol,” explains Akesson. “You can then produce tests that cover those possible transitions and thus systematically explore the state space.” Philips is now using Petri nets to do exactly that: to create test cases from the Comma specifications.

A third approach is to play by a slightly more restricted playbook, continues Akesson. “This is what we’re doing in the Dynamics project with Thales. By not using certain constructions in Comma, and using Petri nets in a different way, it’s possible to build tooling that can statically tell you whether your new interface is backward compatible and if not, automatically generate an adapter – if one exists. We’re now lifting this from a proof-of-concept command-line tool into the Eclipsebased Comma environment, providing immediate developer feedback on why a change is or is not backward compatible and whether an adapter can be generated.”

Open source

This static checking is high up on Philips’ wish list as well, divulges Van der Munnik. “The main benefit for us at the moment is still the dynamic conformance checking while running the test cases, but maybe some of that can also be done statically. Furthermore, we want to extend the Comma framework with the ability to create smart stubs and simulators for clients and servers. And we’re looking into reverse-engineering interfaces by automatically constructing Comma models from execution traces –

but this is still more in the research phase.” At Thales, Noltes is hoping to get Comma out of that research phase and into the modeling practice. “As part of our work with Philips and Thermo Fisher Scientific, we’re extending Comma with the concept of components, ie objects with multiple interfaces,” states Hooman. “These interfaces are often inter-related, which means that if you do an action on one, the state of another changes as well. We’re developing a component that lets you express the relations and possibly the timing constraints between the interfaces. We’re also looking into testing multiple interfaces.” To further the spread, the partners are working on open-sourcing the framework. Hooman: “We’re defining a kind of Comma core in the form of an Eclipse plugin, which others can extend, for instance with their own generators.” Van der Munnik underlines the importance of this development: “It adds to the maturity of Comma. What started as a research project is now a product that can actually be used by developers, in terms of UI, speed, ease of installation and so on. By making it open source, we’re hoping that others will contribute back into Comma, thereby extending and improving the framework even further.” 4 31

THEME SOFTWARE ENGINEERING

NO SYSTEMS ENGINEERING WITHOUT DIGITAL ENGINEERING Digitalization makes it possible to automate large parts of the engineering process. Much can be gained by better harmonizing information from various disciplines, stimulating reuse, formalization and simulation, digital twinning, smart feedback and the application of artificial intelligence. However, the implementation is so comprehensive that collaboration between academia and industry is the only logical step. The High Tech Systems Center is building a consortium around digital engineering to work together on continuous improvement of development processes without losing sight of the business aspects. Alexander Pil

“M

any companies struggle to keep up with technological progress. More and more, techniques and tools are coming onto the market that could be beneficial. Engineers have to be trained to master new cloud technology or simulation methods, for example. This is costly, time consuming and only a small piece of the puzzle. The big question is how to embed the acquired knowledge and potentially crucial methods in their organization and systems – a major problem for the industry. As a result, companies regularly get stuck in the choices they initially made when setting up the product line. And if you stick with the old, y ou quickly fall behind.” These are the words of Marc Hamilton, a fellow at Eindhoven University of Technology’s High Tech Systems Center. He notes that these problems not only affect SMEs. “It’s also a hell of a job for large OEMs with a lot of knowledge and capacity to integrate new tools and techniques into their engineering process. It’s complicated and risky for them to adapt or change their existing workflow. Smaller companies have an advantage in this regard but often lack capacity. For all categories, it puts a brake on how quickly they can integrate new technologies, methodologies and tools.”

Another group for which this problem frustrates progress are the companies that traditionally have a mechatronic origin but increasingly come to the conclusion that they have to retrain themselves as a software company. “Building all kinds of apps, adding connectivity and monitoring in the cloud is becoming a normal requirement,” Hamilton sums up. “Even for something as simple as a toothbrush, there are now apps that track how well you’ve brushed. Companies are forced to transform into becoming software experts, and that’s a struggle. After all, it was never their core business and they simply don’t have the knowledge on board. Good software experts are hard to find. How do you ensure that you keep up to date with all the rapid developments? Tomorrow, a new technology will emerge and the same struggle starts all over again.”

Context

For the further evolution of systems engineering, an impulse is needed for digital engineering. “That involves the far-reaching digital support of the design process, especially in places where there’s still a lot of human intervention,” explains Hamilton. “System developers already use many tools for digital engineering. Each of

these instruments works fine, but it’s still difficult to connect them in a flow. It takes a lot of human intervention to interpret how to transfer the outcome from one tool to the next. This often goes wrong, especially between the various disciplines in a development organization. It would be very nice if we could automatically transfer all that data.” That’s easier said than done. “It’s difficult in two ways,” clarifies Hamilton. “First, each tool uses its own format to specify and save its input and output. Moreover, this is done based on its own paradigms or principles. Second, the context in which you apply them is crucial. For instance, you can easily make a state machine of the behavior of the weather. You could also generate code from such a model, but that doesn’t mean you can make it rain. Engineers must decide when to use which tool and how to use the results in the other steps. We as humans must manage that interpretation of what a tool contributes to the process.” In many areas, there are already new standards and frameworks around digital engineering, or they’re emerging. In simulations, for example, models such as FMUs can be linked together based on the FMI standard. And the CAD world

maintain and expand that position by using the knowledge to make valuable products. That means we have to engineer faster and better. We need to build an ecosystem where we can continuously and quickly explore improvements and new engineering tools, get the results to the market and get businesses to apply them.”

Credit: Eindhoven University of Technology

Three main themes

has various widely supported data formats. Major industries such as defense, aerospace and automotive have a long tradition of engineering standards and frameworks. Hamilton: “However, modules that tie the models together in a simulation environment aren’t suitable for controlling a system. We’re flooded by all kinds of standards that solve partial problems. But the real trick is integrating all these types of technologies and reducing the knowledge required to apply them. We’re trying to get this under control with digital engineering so that we can take further steps to make engineering more efficient and to expand applications of artificial intelligence.”

“How do you keep up to date with all the lightning-fast developments? Tomorrow, another new technology will arrive and the same struggle starts all over again,” says Marc Hamilton, a fellow at the High Tech Systems Center.

Falling behind

Hamilton also looks at the competitiveness of the Netherlands against the rest of the world. “At the moment, we’re not behind in system development, but we’re certainly not ahead. Take artificial intelligence, which plays an increasingly important role in engineering. Many of the examples come from the US or China, where investments are huge. Europe is lagging. We may not be doing badly at the moment, but at this rate, we’re going to fall further and further behind.” “We have a lot of specialized knowledge in the region and we’re really good at building high-tech systems,” Hamilton continues. “You want to

The call for better digital engineering isn’t new. “I notice that many companies have been feeling the pain for some time,” says Hamilton. “Some of them have taken the first steps themselves. However, almost all of them lack the capacity to encompass the complex complete picture.” A few years ago, the High Tech Systems Center (HTSC) took the initiative to set up a consortium and bring parties together to work on better systems engineering. “During the first meetings, topics and challenges continued to come up. It only diverged. There turned out to be so many aspects of the engineering process that needed to be addressed in one way or another that it was difficult to pick and focus,” explains Hamilton how comprehensive the issue is. HTSC has put its back into it and brought the problem down to three main themes. The first is platform engineering. “Many companies are working towards a platform for their product lines so that they can reuse parts and knowledge more easily,” Hamilton points out. “In high tech, with its high degree of optimization, it’s rarely the case that you can transfer subsystems from one machine to another. However, it’s certainly possible to reuse building blocks and modules at higher design levels and to refine them for the new system.” The second theme is the large number of available tools. “A new package or another license sounds expensive, but they can speed up your process tremendously. However, that’s difficult to quantify, which means it will encounter a lot of resistance in the higher management layers. So you 4 33

THEME SOFTWARE ENGINEERING have to link them to the company KPIs. Naturally, this differs per sector. Sometimes, time to market is crucial. If you can gain a month of development time with a new tool, you have a clear business case. In another branch, it may be about reliability and wanting the product to be one hundred percent right. Then it makes sense to invest in verification tools.” Data and feedback are the third main theme. Today’s systems collect a wealth of data. With that input, you can make very valuable improvements for an update or in the next system. Of course, this only works with meaningful data, which can also be expensive to collect.

Data lake

Many relevant research proposals have been collected around these themes, which have subsequently been clustered. This clustering has now led to a clear starting point for further research in the other clusters. Therefore, the consortium’s first concrete step is presented. “We’re going to work on a foundation that we call systems engineering process orchestration,” says Hamilton. “In doing so, we link models, data and tools to a data lake so that the possibilities for tackling the challenges are expanded with modern data analysis and AI technologies. Essential is the context

of system design decisions, application of digital twins, the design of the data generated by the system and processing of that data to design improvements.

Credit: Eindhoven University of Technology

Fast spinoff

HTSC is building a consortium on digital engineering from its new location in the Meulensteen House of Robotics on the TUE campus.

of the models and data. After all, data has no value if you don’t know where it came from in the engineering process and how it relates to the systems being developed. In addition, you must be able to value the steps of that engineering process in terms of relevant company KPIs to be able to optimize. Such engineering knowledge must be injected into the data lake.” While this basis offers opportunities for embedding and tackling digital developments in engineering processes, other clusters emphasize improvements in process support. This includes process optimization, automation of synthesis, support

Do you have an unstructured design flow and no time or knowledge to improve it? www.dizain-sync.com PCB, FPGA, Cabling and Chip design environments PB 34

The High Tech Systems Center is now building a consortium on digital engineering. “We’re currently in the phase of pitching these new ideas to companies. We want to build a proof of concept. That would make it a lot more tangible. We’re not starting blindly; we rely on existing technology and we work to expand it. There are already many options and parts. Aside from the necessary fundamental research, the consortium also wants to give full attention to the coupling of applied research, and rapid spinoff and feedback of the results to practice.” The sooner Hamilton can discuss his plans with partners, the better. “Hopefully, then we can make everything much more concrete based on use cases that participants contribute themselves. When will we have the first meeting? I hope this fall. That may be ambitious, but I think we have a good roadmap, a good campaign plan.” Interested parties are welcome to participate. Please contact Marc Hamilton via m.a.m.hamilton@tue.nl.

pinion

SOFTWARE ENGINEERING Han Schaminée is a product innovation consultant.

What’s this thing called software engineering?

n September 2018, Corinne Vigreux started CODAM, a training program in programming. Corinne is one of the founders of Tomtom and she devotes part of the capital she earned there to make this world a better place. And since this world needs software, it needs talent that can create that software. Claiming everybody can learn to program in 3.5 years, CODAM believes there’s still a lot of untapped talent around, especially among people who can’t afford to take a course or to attend university. The program is for free and in this way, Corinne believes she cannot only help these people but also help industry find the software talent that it so desperately needs. Checking out CODAM’s website, I noticed it talks about programming. Is that the same as software engineering, I wondered. Where does software engineering fit in, actually? Many engineering disciplines make use of coding, like mechatronics, automotive system design and healthcare system design. Is software engineering just a capability within these engineering disciplines, like thermal analysis for hardware design, construction calculations for an architect or solving a differential equation for a motion control engineer? Or is software engineering a discipline in itself? And if it is, will it benefit from CODAM? In my daily work, I notice many different definitions of software engineering. People, especially without a background in computer science, often see software engineering as coding or programming. That doesn’t feel like an engineering discipline, does it? To me, an engineering discipline includes understanding the

problem and applying technology to find a solution for it. In fact, I feel coding is the smallest and easiest part of software engineering. And as a consequence, I believe, efforts to just optimize the coding part fail to address our biggest problem. Wikipedia defines software engineering as the systematic application of engineering approaches to the development of software. It also says

Software engineering has a lot to do with system design software engineering is a subfield of computer science, management science and systems engineering. The latter is quite interesting. Indeed, the field of system design is part of different departments at different universities, be it computer sciences, mechatronics or management sciences. These all approach things differently. I’ve seen the same in companies – in some, system architects preferably do not have a software engineering background! To me, software engineering has a lot to do with system design. But especially the design of systems in which the complexity isn’t so much in the components itself, but in the interactions between them. I have yet to hear other system design disciplines consider association and inheritance relations to

better understand the interactions between. The complexity is not in fighting physical constraints, but in managing all these interactions and all the possible states caused by the many interactions. Software engineering is a rather new discipline. And like many other new disciplines, people believe it’s easy, just common sense, something everybody can do and not as difficult as real disciplines like physics or chemistry. The same is true about user interface design or even leadership. Of course, we need algorithms and programs. But that’s not unique for software engineering; the other disciplines need that as well. It would be good when we as professional software engineers would do a better job communicating what our discipline is all about. I believe 3.5 years of programming training in CODAM doesn’t make you a software engineer, like a soldering training doesn’t make you an electronic engineer. But it’s still a very good basis and a fantastic initiative. Thank you, Corinne.

4 35

THEME SOFTWARE ENGINEERING

CRACKING THE CODE TO CRAFTSMANSHIP

With the growing reliance on software in an increasingly high-tech world, it’s more important than ever to master the art of software engineering. Trainers Robert Deckers and Bart Vanderbeke have taken it upon themselves to turn developers into craftsmen. Nieke Roos

“A

colleague once told me about one of his former project managers, who, upon realizing that the estimates didn’t align with his timeline, just cut them in half to make them fit. I find it unheard of, not only that you’d do such a thing as a project manager but also that people stand for that kind of behavior. You don’t have to scold him, but you can open your mouth. Instead, at the end of the project, when everything has gone haywire, everyone complains about how this has happened to them.” Inspired by Google executive Fred Kofman and his book “Conscious business,” Bart Vanderbeke calls on software engineers to stop playing the victim. “It’s unacceptable and unhealthy,” he claims. “You’re the craftsman. When someone tells you that you need to do something in half the time, or skip the design, or refrain from reviews, you say no – constructively. Software engineers are scarce, so you’re in a comfort-

“The real complexity is in the non-functional” able position, certainly no position to self-victimize. Don’t hide behind ‘management.’ As a software craftsman, using a term coined by Kofman, you’re ‘unconditionally responsible’ for everything you do or don’t do.” 36

At NXP in Leuven, Vanderbeke leads a team of fifteen software engineers, working on 2.5 GHz radio applications for personal health – think hearing aids, headphones and earplugs. “Tiny systems containing tiny software stacks,” he notes. “But even if you have a codebase of 100k or 200k, like us, software craftsmanship is of paramount importance. Building the hardware takes about a year, followed by maybe five years of software enhancement. I’ve developed a series of lectures to help my colleagues bring out their inner craftsman.”

The non-functional

A kindred spirit, Robert Deckers, too, aims to increase software craftsmanship, but with a focus on architecture – “the most difficult trick of the trade,” as he calls it. “It already starts with the question: what is software architecture? You can find hundreds of books that try to give an answer. While some are bad, terrible even, most of them are meaningful, but they all tell a different story.” This was one of two triggers that led him to dive into the subject, develop his own view, write his own book and bestow his insights upon others. The second trigger was the realization that in traditional methodologies, there’s too much focus on the functional requirements, whereas the non-functionals are the hardest to get to grips with and therefore take up the most time. “Way back when I was an OOTI trainee at Eindhoven University of Tech-

Descending down the management funnel, the focus narrows and the risk for conflict grows.

nology, I was the software architect for a copier,” reminisces Deckers. “After two months of design, the anomalous system behavior started to rear its ugly head and I realized that we had to do error handling as well – while obvious to someone with 20 years of experience, it hadn’t crossed my newbie mind. When we were finished, to my big surprise, no less than 85 percent of all our code

Credit: Bart Vanderbeke

“You need to take unconditional responsibility, which literally means that you need to have the ability to respond – in a meaningful way,” says Bart Vanderbeke.

be blown away by unsubstantiated claims. Someone once asked me if we could speed things up by taking shortcuts, upon which I replied: ‘The only shortcut you can take is skimp on the specs’ – and that was the end of the discussion.” “You need to take unconditional responsibility, which means that you need to have the ability to respond – in a meaningful way,” continues Vanderbeke. “In my workshops, I use several small examples, taken from my everyday work, to get my point across. For instance, someone escaping responsibility would say: ‘I cut my estimation in half because my project manager told me to,’ as opposed to someone keeping ownership, a ‘player’ in Fred Kofman’s terms, who would say: ‘I wanted to avoid a fight with my project manager, so I gave in.’ Likewise, a victim would say: ‘I make estimations because our process demands it,’ whereas a player would say: ‘I want to stay with the company, so I use the established process.’ By making them aware of these little things, people are more inclined to correct their behavior.” Credit: Robert Deckers

turned out to be for error handling, so only 15 percent of our efforts had been focused on the functionality. That’s when I first experienced that the real challenge, the real complexity, is in the non-functional.” After the OOTI traineeship – the PDEng Software Technology, as it’s known today – Deckers sharpened his views at several companies, including Philips Research and Sogeti. Since 2013, he’s running Atom Free IT, coaching organizations and their architects, helping them create architectures, set up the architecture process and embed it. The last five years, he’s combining this with a PhD project at the Vrije Universiteit Amsterdam, researching the cognitive aspects of systems engineering.

software craftsman, you know how to organize your work and you have the assertiveness not to accept compromises on the way of working. Instead, you go for the optimum, taking into account the influencing variables and conditions. You don’t do things because someone tells you to, but as you understand the need, you autonomously decide to do so,” summarizes Vanderbeke the values he intends to convey in his workshops. Learning to say no in a constructive way is a key topic in Vanderbeke’s teachings. “That requires you to come prepared. When you’re asked to plot a course in a project, you need to have a couple of options readily available, not down to the minute detail but to such an extent that you can weigh them and make an informed choice. When someone steps up to you and says something can be done in half the time you estimated and you don’t have your facts straight, he may well be right – you have no way of telling. If you know what you’re talking about, that will not happen. You can have a constructive conversation and you might be challenged, persuaded even, but you won’t let yourself

Come prepared

Vanderbeke and Deckers are the newest additions to the software and systems portfolio of High Tech Institute. Both want to help software engineers be better at their work – become real craftsmen. “As a

A good architecture is correct, consistent and communicated. 4 37

Brazilian view on Dutch mechatronics In 2018, Vinicius Licks, professor and associate dean of mechatronics at Brazil’s Insper College, made his first of three long treks from South America straight to the Netherlands. He didn’t travel across the globe to enjoy a vacation; he came to get a feel for the Dutch high-tech environment, specifically through the mechatronics training cluster provided by High Tech Institute. Licks’ observation: “The curricula are completely designed for someone who wants to have a complete view of the field of mechatronics design. The sequence of courses is built in such a way that some frameworks will be dealt with continuously, but from different perspectives and with increasing complexity.” insper.edu.br hightechinstitute.nl/mechatronics

THEME SOFTWARE ENGINEERING

Fish nor fowl

In his training courses, Deckers relays his ideas about good architectures. “The role of architecture is to offer a solution approach for the key system properties that are the hardest to realize. As an architect, you always have to make sure you’re working towards a solution, providing guidance and serving your stakeholders’ needs, while also keeping

“Learning to say no requires you to come prepared” an eye out for things that could go wrong if not addressed in the architecture. If you’re not doing this, you’re probably not architecting. Also, I want engineers to understand that an architecture needs to offer business value and that it’s feasible to build the system within the organization at hand. You can only be an effective architect when you’re pre-

Good software architecture

The training “Good software architecture” is now available at High Tech Institute. You’ll learn to bridge between customer needs, technological constraints and the development process in order to deliver the best software architecture.

pared to step out of your technology comfort zone.” According to Deckers, a good architecture is correct, consistent and communicated. “A system has to be correct in that it has to adhere to the stakeholder concerns and the technical environment. The development process has to be consistent. At Philips in Bruges, I once witnessed an engineer testing all preconditions of all the functions he programmed because he wanted his code to be robust. Meanwhile, in the cubicle right next to his, a colleague was using pointers without testing anything because he wanted his code to be fast. Combined in one system, that gives you neither fish nor fowl. You need to be clear on the key properties – my advice: hang the top five on the wall. Finally, an architecture should be described in such a way that you can discuss it with the different stakeholders, which means using different views for different aspects.” Deckers stresses the importance of focusing on the non-functional properties, aka the quality attributes. He acknowledges that this seems to be at odds with the popular Agile principle of delivering working software as quickly as possible. “People often ask me: how do you match Agile and architecture? My answer to them: you don’t. They’re two different mindsets. Architecture is about looking before you leap, whereas with Agile, you just go and adjust based on the feedback you get. That’s perfectly fine for some businesses, but not for a copier or a medical scanner, where aspects like reliability and safety are known beforehand. The closest way to match Agile and architecture is to bend the rules and dedicate the first few sprints to the key concerns.”

Better decisions

With collaboration, there’s bound to be friction. A software craftsman, therefore, also nee ds tools for conflict resolution. In his workshops, Vanderbeke presents a management

Robert Deckers stresses the importance of focusing on the non-functional properties, aka the quality attributes.

funnel doubling as an inverse conflict pyramid. It goes from wide to narrow in three levels: strategic, tactical and operational – from the what to the how. Descending down the funnel, the focus narrows and the risk for conflict grows. When a conflict actually arises, going back up the funnel to try and find a shared goal or principle helps to smooth things over. “Engineers who are in disagreement about the way to tackle a problem often are at the bottom, stuck in their own solution. Taking them up and discussing the problem criteria usually ends the stalemate as they establish common ground,” illustrates Vanderbeke. “It’s a very useful instrument in process management as well. When you’re in a meeting that’s going nowhere, revisit the reasons why it was set up in the first place and a way out will present itself almost automatically.” Stocked toolbox in hand, software engineers are well equipped for craftsmanship. With Deckers, Vanderbeke concludes: “It would be great to see them make better decisions. To see them operate more autonomously. And, at the same time, to see them have more fun in what they do.” 4 39

THEME SOFTWARE ENGINEERING

A PRESSURE COOKER FOR SOFTWARE TALENT Still eager to learn, even after an extended bachelor and two masters, Tom Vrancken signed up for the PDEng Software Technology program. In two years of different projects with different companies, he’s gaining the experience that otherwise would have taken him 10-15 years to get in industry. Nieke Roos

Credit: Rien Meulman

“A

s a university graduate in computer science, you usually start your professional career as a junior developer. Only after at least 5-10 years of experience, you get to be a software architect. You need to grow into that role, really get to know the company and its products. Win your spurs, so to speak,” explains 31-year-old Tom Vrancken. “But as I like to solve technical challenges on a higher level, I wanted to start as an architect. The PDEng Software Technology program offers that opportunity.” The PDEng Software Technology (ST) is a two-year salaried post-master technological designer program on a doctorate level for top MSc graduates with a degree in computer science or a related field. It prepares the trainees for a career in industry by strengthening their theoretical basis and confronting them with challenging problems from industrial partners. With a variety of clients offering complex system/software architecture and design-related challenges, you learn to develop innovative solutions meeting industry standards while mastering all the aspects of teamwork, different roles and professional skills. Already at the start of the program, Tom got what he wished for: the software architect role, in a project at CERN, the European Organization for Nuclear Research. “We developed a database system and the associated API to store data for a new experiment slated for 2026, called the Search for Hidden Particles. We built on a feasibility study made by the 2018 group of PDEng ST

trainees, turning their concept into production-ready software. As the architect, I was responsible for creating the specifications and making sure the team adhered to them.” Technically, it wasn’t very difficult; for Tom, the big challenge was getting everybody on the same page – both the team and the customer. “I thought I’d described it all very clearly. Yet, I got solutions from the team that were not at all what we’d agreed upon. It was also very interesting to deal with the client having a completely different view. I’ve learned that there’s more to being a software architect than just solving a technical puzzle – you’re working with people from different backgrounds, so communication is key.”

Cybersecurity

Having tinkered with computers in high school, computer science was the logical choice for Tom, and Eindhoven University of Technology (TUE) in his hometown the logical place of study. He enrolled in 2006 at the age of 17. Six years later, he obtained his bachelor’s degree. “It took me longer than average,” he admits, smiling. “But that’s because I combined studying with kayaking for the Dutch national team. Each year, the university allowed me to move one or two subjects to the next year, to free up 20-30 hours a week for me to train and race.” While studying for his bachelor’s, Tom also did an extra minor and even got a jump on a master’s degree. “My first minor was part of the regular curriculum. I chose mechanical engineering because I’m also interested in machine control and robotics – an interest that was spoon-fed to me by my father, who studied mechanical engineering here in Eindhoven. For my second minor, I followed a couple of more in-depth computer science classes. At the end of my bachelor’s, I had six months to spare, which I decided to fill by starting the educational master. I really like to educate people – I used to be a student assistant and a member of TUE’s student PR team, and I’m a certified kayak and snowboard instructor.” In 2012, Tom continued his computer science education with a technical master in information security technology. “This was a special cybersecurity program organized by TUE, together with the Radboud Univer-

For CERN, Tom and his team developed a database system and the associated API to store data for a new experiment slated for 2026, called the Search for Hidden Particles.

sity in Nijmegen and the University of Twente. As part of the program, I had to attend classes at all three participating universities, leaving me with no room for other major activities. So I quit the national kayaking team and joined the local canoeing club instead, and I put the educational master on hold – to be picked up again after my graduation.” For his graduation project, Tom connected with Arpa2.net, a non-profit organization he came across at one of the computer science conferences he attended. “Their goal is to make the internet safer and more privacy friendly, taking control away from the tech giants and giving it back to the user. My graduation project with them was about integrating the authentication protocol Kerberos with the cryptographic web protocol TLS to create a more secure way of communicating over the internet. I improved the design and built a prototype. In my free time, I’m still involved in this project, called TLS-KDH.” After his graduation, Tom worked part-time as a software and security engineer at Thermosmart and for his own company, V-Studios.

Expert coaching

Still eager to learn, even after an extended bachelor and two masters, Tom signed up for the PDEng ST program in 2019. “I was looking to deepen and broaden my computer science knowledge. Despite all the courses I’d followed, I didn’t feel ready yet to take on the final responsibility for a product or system. I wasn’t

confident enough about my gut feelings. I wanted to learn from experts whether my technical instincts were right. When I was a student counselor during my bachelor, I once attended a presentation about the ST program, by then director Harold Weffers, and I had always kept that in the back of my mind as a viable option in case I wasn’t fed up with school after my graduation.” One of the main benefits offered by the PDEng ST program, according to Tom, is that it combines the best of the academic and the industrial world. “On the one hand, you follow state-ofthe-art university courses, for example about software design and validation, taught by top scholars. On the other hand, you receive training and coaching from experts who have worked for 20+ years at companies like ASML and Philips. Who know how to play the game and who can tell you what to do and what not to do.” By the end of his first year, Tom will have had three different jobs at three different clients, each for 8-9 weeks. After being a software architect for CERN, he recently worked with the automotive supplier Valeo, as a test and configuration manager for an autonomous driving applica-

PDEng ST 2020 edition The 2020 edition of the PDEng Software Technology program is scheduled to start on 26 October. MSc graduates or software engineers interested in participating are invited to contact ooti@tue.nl.

tion. Up next is a machine learning project at the European Space Agency ESA, where he’s going to have the team leader role. His second year will be largely taken up by a 10-month graduation project. “Together with Philips, I’m looking to combine software engineering and security.” “The ST program is a pressure cooker,” summarizes Tom. “In two years, you get to do all kinds of different projects and see a host of different companies, both publicly and privately funded, thus allowing you to find out relatively quickly what suits you best. In industry, it would take 10-15 years to gain the same experience.”

Ready for business

After he graduates, Tom will finally put his educational career to rest – at least for now. “A PhD isn’t completely off the table. With my PDEng degree, I could shave off some time, but it would still take me two or three years. That’s not on top of my mind right now. A teaching job at the university would be equally interesting – for later.” Almost halfway the ST program, Tom is starting to feel ready for business. “My role as a software architect in the CERN project has affirmed my belief that that’s the way to go for me. The positive feedback I’m receiving from the coaches has taken away that little bit of insecurity I had about my technical instincts. The program has already boosted my confidence to such a point that I can’t wait to go to work as an architect in industry.” 4 41

THEME SOFTWARE ENGINEERING

OPTIMIZING YOUR HIGH-TECH DEVELOPMENT AND MACHINE PERFORMANCE An increasing proportion of high-tech equipment consists of software. At the same time, the importance of utilizing software in optimizing development processes is growing. That doesn’t take away the fact that significant progress is always made in the interplay with mechatronics and electronics. “Getting the best results calls for a culture of a real machine builder.” Daniëlla van Laarhoven

he complexity of high-tech systems has increased enormously over the last few decades. OEMs – for example in the semiconductor, analytical and digital printing industries – create machines that not rarely operate on the edge of physical possibilities. Their performance is ever more determined by software. “Creating a machine with optimal behavior, also in our high-tech world, requires a sound mechanical and electronic design,” says Laurens van de Laar, a software architect at NTS. “If that’s not the case, you can’t usually fix it with software. The laws of physics always apply. However, with the use of software, you can push the boundaries. It’s a technology to maximize results, both from the perspective of machine and process control. You can add speed, accuracy, usability, maintainability, freedom of choice, functionality, uptime and so on. In other words: there’s a difference between complexity and intelligence, and the latter always runs on software. That makes our discipline so crucial for NTS and our customers.”

Building bridges

NTS is a first-tier partner for OEMs, has extensive knowledge of optomechatronics and a strong focus on high-complexity, low-volume, 42

high-mix markets. The company is globally active and operates as a one-stop-shop in development and engineering, component manufacturing and assembly. Van de Laar has been working at the company for 13 years now. He’s a member of the software team within the Development & Engineering division. Over

the years, in line with the evolution of high-tech machine construction, his work domain has changed substantially. In the past, software was usually applied as a final layer once the physical design was done. Now, it’s part of development processes from the very beginning of the creation process.

“Our customers continually push technological innovation within new business models. Short development times, cost efficiency and quality are crucial to their competitive strength. As software developers at NTS, we contribute to that in various ways. First of all, through our ability to rapidly set up a conceptual system design within an integrated multidisciplinary approach and then realize it in a short amount of time.” “One of the resources we employ to achieve this is our NTS Machine Development Studio, an advanced engineering tool we develop ourselves,” continues Van de Laar. “This enables us to combine all sorts of building blocks into virtual systems and run, test and analyze it. In this way, we can assess possible system configurations and identify and eliminate bottlenecks in the early phases of the development path together with our customers. In addition, it allows us to make all kinds of testing and analysis functionality available very quickly, which is also important from the perspective of design for excellence and optimizing manufacturing.” At the same time, Van de Laar points out, it should be said that software tools are only part of the solution in achieving the goals of customers. “The use of previously developed solutions – the re-use of proven technology – also helps greatly and in doing so, NTS can build on an extensive and broad portfolio of projects. Also, our strong vertical integration allows us to easily build the necessary bridges between mechatronics, electronics and software. For instance, we have data available about all sorts of components that are measured or built in our production facilities.”

Industry 4.0

Colleague software architect Jurgen Wilsch underlines the advantage that the culture of a high-tech manufacturer offers him in exercising his profession. In addition to the joy that comes with a variety of projects and technological challenges, this was a major reason for him to switch from a large international OEM to NTS three years ago.

“There are a lot of clever and skilled software engineers who can contribute to the development and construction of high-tech machines,” says Wilsch. “They’re skilled in creating solid architectures, applying all sorts of tools and methodologies, writing code, you name it. We have these people at NTS, but we also hire them through specialized partners when projects require this. In that way, we’re able to quickly ramp up and down to benefit our clients.” According to Wilsch, the NTS software team has a core quality that can’t be found anywhere else. “In our men and women beat the hearts of real machine builders. Their shared knowledge in the fields of optics, physics, materials, dynamic behavior, thermal effects and all sorts of related disciplines is enormous. And more than that, they can quickly translate their experience in these fields into an image of a machine: how it should look, what components are needed, what the critical

According to Laurens van de Laar (left) and Jurgen Wilsch (right), software architects at NTS, software is an enabler for pushing boundaries.

factors are and what it takes to get it to do what it needs to do. On top of that, they can easily determine how users – clients, service engineers, developers – ultimately want to deploy and use the machine.” All these competencies add up in meeting the needs of customers, finds Wilsch, even in those cases in which they don’t know exactly what they’re asking for. “In my view, this is precisely what’s needed to add maximum value as software developers in converting technical and functional requirements into an efficient development process and contemporary technology. I’m not just talking about topics such as development speed and the actual performance, functionality, flexibility and robustness of the module or machine. I’m also referring to – given the emergence of Industry 4.0 – matters such as connectivity and preventive maintenance.” Edited by Nieke Roos

4 43

THEME SOFTWARE ENGINEERING

HIGH-LEVEL FPGA PROGRAMMING FOR NANOSECOND TIMING IN TERABIT COMMUNICATION Next-generation data communication using laser signals between ground stations and satellites will be at the terabit per second level. Given the high demands on data quality and processing speed, wavefront sensors and FPGAs are essential ingredients of the required communication terminals. Demcon and Qbaylogic demonstrate the potential of high-level functional FPGA programming. Jan Kuper Joost Kauffman

ithin the Tomcat project (Terabit Optical Communication Adaptive Terminal), part of the European Space Agency’s Artes Strategic Program Line Scylight, TNO is in charge of developing an optical ground station, including an optical ground terminal (Figure 1). From a satellite, a terminal will receive laser signals that are affected by atmospheric conditions such as temperature variations and turbulence, which induce deformations of the beam’s wavefront. Adaptive optics can counteract these deformations using a segmented deformable mirror in which each segment is individually actuated based on the input provided by a wavefront sensor. TNO and Demcon jointly built a wavefront sensor upgrade to the high sample rate (5 kHz) required for this application. It’s one of the laser communication instruments developed and marketed by the Enschede-based company with the Dutch FSO instruments consortium, supported by the knowledge institute. This project involved dedicated optical hardware and data-processing software, implemented in FPGAs, as 5 kHz sampling couldn’t be achieved on a PC.

The wavefront sensor comprises an array of lenslets that each focus part of the incoming signal on a subregion of camera sensor pixels. To calibrate the mirror segments, 256 two-dimensional points of gravity of the corresponding subregions of each image have to be calculated on an FPGA. The camera sends the image to the FPGA line after line, in packages of 8 pixels of 12 bits each, together with some control bits for validity and end-of-line, at a rate of

Figure 1: Within the Tomcat project, an optical ground station, including an optical ground terminal, for laser satellite communication is being developed.

80 MHz – whereas the Tomcat specification requires a 200 MHz FPGA rate. Hence, on average, one package of pixels arrives every 2.5 cycles of the FPGA clock. The transfer of a full image takes 157 μs and the calculations have to be completed within 3 μs after the last package has arrived.

Haskell

Demcon engaged University of Twente (UT) spinoff Qbaylogic, because of its design methodology for

high-level FPGA programming. The advantages are that dependencies in the various processes, for example regarding the order in which the data are received or transmitted, can be dealt with adequately and exact timing, down to the nanosecond level, can be achieved. This high-level functional methodology offers a fast design process with full control over FPGA code efficiency. A central element in the design methodology, based on the functional programming language Haskell, is the open-source compiler Clash, the result of over 10 years of (ongoing) UT research. It translates a functional specification written in Haskell into any of the standardized hardware description languages (HDLs, such as VHDL, Verilog and System Verilog). Haskell isn’t commonly used in industry, one of the reasons being that it offers less control over CPU performance than, for example, C. However, on an FPGA, the situation is the other way around: with Haskell, a designer has more control over performance than with C/ C++. The reason is that a functional Haskell specification is structural, while a C/C++ program describes behavior rather than structure. Hence, Clash generates HDL code in a structure-preserving way, whereas a high-level synthesis tool starting from C/C++ needs transformations that are hard to grasp and may generate unintended hardware with unpredictable performance. Worthwhile aspects of the methodology include its fundamental-

ly model-based character and the availability of adequate abstraction mechanisms, among which are higher-order functions, embedded languages and typing mechanisms. Another advantage is that cycleaccurate simulation on a functional level is possible at all design stages.

Correctness

Haskell, close to mathematics, is suitable for expressing the model of an application. In the Tomcat project, this was the basis for fast and effective communication about the precise functionality definition. In addition, such a model is executable, so that its correctness can be checked. The model then is the starting point for the design of an FPGA architecture that has the same functionality and meets the performance requirements. Note that all design steps are performed in the same language and hence are executable and testable. This increases the productivity (and the satisfaction) of the designer and greatly contributes to the correctness of the design. A first abstraction mechanism is higher-order functions (HOFs) – “higher-order” because they take another function as an argument. HOFs express architectural patterns in which a given function is used repetitively. For example, in Tomcat, the parallel pairwise multiplication of a vector ps of pixels with a vector is of indices can be formulated using the HOF zipWith as zipWith (*) ps is (see also Figure 2). The first argument of zipWith can be any bi-

Figure 2: The architectural pattern of zipWith, together with another higherorder function, fold, which expresses an accumulative application of a binary operation on a sequence of values. The dot product of two vectors is a combination of zipWith and fold.

nary function, in this case (*), for multiplication. In practice, vectors and matrices tend to be huge and the straightforward modeling with HOFs may lead to architectures that either don’t fit on the FPGA at hand or may give rise to a slow clock. In such cases, modifications of HOFs are available with proven correctness by which a design can be pipelined or otherwise executed over time. Embedded languages, the second abstraction mechanism, offer the possibility of hiding the underlying bit representation of certain constructs. They’re very practical for an instruction set of a processor or for the states of a state machine and are very helpful in avoiding errors. In Tomcat, a small embedded language is defined for packages of pixels arriving from the camera (Figure 3). Note that an embedded language is defined as a data type so that a function can be directly defined on constructs of such a language by using a technique called pattern matching (Figure 4). The importance of Clash’s typing mechanism, the third abstraction mechanism, can hardly be overstressed, because many, if not most, errors made in practice are typing errors. A strong type-checking mechanism catches errors in an early design stage. As part of the typing mechanism, Clash can derive the type of some component, even if the designer doesn’t indicate the type explicitly. In Tomcat, this feature was often used: when the spec4 45

THEME SOFTWARE ENGINEERING ification wasn’t accepted by the type system and its error message was somewhat cryptic, it would help to isolate a function and ask Clash what type it ‘thinks’ the function has.

Right the first time

The methodology yielded a fully pipelined architecture, by which incoming packages of pixels are first regrouped into vectors of the size needed for the application, then multiplied in parallel by index values and finally, the results are added in a tree-shaped adding mechanism. Then follows a step of accumulating results per relevant region, after which a pipelined division operation is applied for the actual point-of-gravity computation. For each subregion, this computation is completed 70 clock cycles after its last pixel has arrived on the FPGA. At 200 MHz, this corresponds to 0.35 μs, whereas the requirement was 3 μs. The

VHDL generated by Clash was adapted to the required interface and straightforwardly integrated with the VHDL for the surrounding architecture as developed in other project parts. The functional level and abstraction mechanisms of the Clash methodology made the communication between Demcon and Qbaylogic fluent and effective since the language used for the design process is close to the language in which the application was defined originally. The resulting architecture, which satisfied the requirements, was created within the required development time; in fact, it was right the first time. Jan Kuper is co-founder of Qbaylogic and Joost Kauffman is a senior system engineer at Demcon Focal, both in Enschede. Edited by Nieke Roos

Figure 3: The small embedded language defined in Tomcat for packages of pixels arriving from the camera. The first clause is for packages containing a Boolean for an endof-line marker and a vector of 8 pixels (defined as 12-bit words). The second clause is needed at those clock cycles when no new package of pixels arrives. Clash has a default translation of values of such types into bit patterns, but the designer can also define a bit representation.

Figure 4: Using pattern matching, a function f can be directly defined on constructs of an embedded language. Parts of packages may be automatically extracted and given names (eol, pxls), which may then be used in the body of the function definition.

FORTUNES OF HIGH TECH

A HISTORY OF INNOVATION AT ASM INTERNATIONAL 1958-2008 Jorijn van Duijn explores the dynamics behind the greatest high-tech innovation: the computer chip. This brand new book describes the history of ASM International between 1958 and 2008. techwatchbooks.nl/fortunes

Order now

pinion

SYSTEMS ENGINEERING Marcel Pelgrom consults on analog IC design.

Mastering complexity

hen chess world champion Magnus Carlsen was asked what he had learned during his training sessions with former world champion Kasparov, the answer was simple: rapidly analyzing a complex chess position. And indeed, Carlsen is one of the fastest to find the crucial move in an unknown position. An important aspect of chess analysis is to recognize structures: specific arrangements of pawns and pieces that create a solid defense or allow specific attacking tactics. Experienced players cherish such structures, as they’re integral to their game. Similar mechanisms can be recognized in designing and analyzing complex technical systems, both in hardware and software domains. In the era where a telephone exchange required a big hall, the only way to keep control over the design was to strictly separate functions and structures. The same is valid for a ‘simple’ system-on-chip IC, with various I/O channels, a few processing units and a diverse collection of memories. Clear function definitions and interfaces allow off-loading irrelevant sidelines when overviewing the system. Managing complexity starts with selecting people. Brilliant engineers design simple systems. Electronics and software mavericks, on the other hand, are a real danger, as their smart and elegant ideas mess up the simplicity of structures and turn mastering complexity into a nightmare. Equally dangerous are people with hammers that treat every issue as a nail. These engineers, once successful with a trick, apply the same solution to every problem. Engineers must be able to adapt their way of thinking to the system

they design. That starts with acquiring a variety of tools and skills, which allows you to easily switch from analyzing a control loop in the time domain to its frequency domain or root-locus view. Real masters of complexity can simplify the system to its essence without getting distracted.

Organizing the daily workload is paramount They can change their perspective and analyze a problem from various sides. A broad technical experience allows them to use ideas and methods from other disciplines. There they outperform their run-of-themill colleagues. They see the red line running through the project and by experience discriminate essential operations from nice-to-haves. Implementing a system and choosing its technical structures is a consequence of a rigorous analysis of the required objective for the system. Forty years in IC design have taught me that there’s always only one objective that really matters – the rest is negotiable. A system built for serving multiple objectives is by definition too complex to master. Extensions, upgrades, modifications or even repairs drown in an uncontrollable workload. The objective comes sometimes with the specifications, but mostly it’s more than that. Finding the crucial objective can be trivial. For Miele household appliances, it’s reliability. For Ferrari cars, it’s ap-

pearance. For Apple electronics, It’s ease of use. When Apple wanted to introduce the smartphone, Steve Jobs had every proposal brought into his office. He knew what his company objective was and how it should translate into a phone. Every proposed device not satisfying that was smashed into the wall. During the implementation phase, the objective is leading and recognizable in the backbone of the system. Engineers have an unhealthy tendency to economizing their designs. As if the Excel generals in the company would be able to spot the waste in hardware or code. Every structure has its own set of supporting functions; optimization isn’t advisable in an early design phase. Complexity requires an appropriate level of administration. In the days that a single engineer could build a chip or software application, some scribbling might suffice. Today, progress in a project is virtually impossible without adequate documentation and communication. Every engineer can easily identify a dozen more attractive tasks, yet a clear way of working is a must to master complexity. Organizing the daily workload is paramount: keeping track of experiments, sorting, labeling and storing test and simulation results, and keeping modifications separate from the golden code are elements necessary for avoiding communication pitfalls in modern system design. Designing complex systems requires a comprehensive approach, from personnel selection to documenting the daily progress. Mastering complexity is foremost an exercise in discipline, a virtue desperately needed in today’s world. 4 PB 47

THEME SYSTEM ARCHITECTING

“HIGH UP IN AN ORGANIZATION, YOU’RE BUSY WITH KEEPING MANAGEMENT AT EASE” Ben Pronk gained fame as a system architect at several Philips divisions. A year ago, he decided to round off his career by joining a startup in robotic surgery. In the run-up to his keynote at the Bits&Chips System Architecting Conference, we ask him about his 30 years of experience as a system architect. René Raaijmakers

n the autumn of 2019, Anupam Nayak and Maarten Steinbuch contacted Ben Pronk. With their startup Eindhoven Medical Robotics (EMR), they were working on devices that can drill, saw and mill bone independently during medical interventions. With this technology, they promise to change the surgery game and have the ambition to be a market leader in robots for the operating theatre in 2028. Pronk, a system architect who made a name for himself in countless Philips divisions and spinoffs, saw an opportunity and decided to prove himself once more. A few weeks later, Pronk selected a chair and a desk between a team of youngsters and a few experienced guys. “I’ve passed the age of sixty and wanted to do something fun again. The medical world and the application are very interesting. What mainly appealed to me was the small scale. The organizations in which I worked consisted of thousands of people. At Philips divisions and later Signify, you have a specialist on hand for each area, or at least someone who can arrange it. Little or nothing has been arranged here. I have to go out for a thermal simulation and order my own PC. I also program it 48

myself,” he says, adding laughingly: “If I have to.”

System thinking

In the past 30+ years, Pronk saw the system architecting profession evolve. In the 1980s, a system architect was often a domain specialist. Someone with expertise in the dominant technology, such as X-ray or MRI in medical diagnostics. Somebody like that knew everything there was to know about X-rays. But with digitalization, everything became computer controlled and more intertwined. The boundaries of disciplines blurred. Mechanics, electronics, mechatronics, optics – it all became interwoven with software. Wherever possible, electronics shifted from analog to digital and in the last decade, products also became networked and connected to the cloud. As a result, the system architect became less and less a specialist in the dominant discipline. Pronk: “Very few products are really monodisciplinary mechanical. Nearly every product these days has an app to go with it. In any case, system thinking is almost a necessity. Of course, there are still architects who work purely on basic techniques such as optics for lenses,

but they’re niche specialists. Even small companies can’t develop anything without networking and Wi-Fi.” How do you see the role and tasks of the system architect? “For me, there are two sides: the multidisciplinary aspects in R&D and the broader view. The system architect must keep an overview, connect all disciplines and distribute them across the system. The challenge is to organize it optimally. What do you do in software? What in mechatronics? You shouldn’t design everything in mechanics if you can do it in electronics. You shouldn’t do it in electronics if you can do it in software. How do you ensure coherence?” “Second, the system architect needs to look wider, not just at R&D. He needs to coordinate with marketeers and the people who deal with logistics and manufacturing. You can make a beautiful design, but it also has to fit in with your logistics flow and manufacturing.” With the growing palette of options and technologies, system architects have to oversee more and more. At the same time, R&D organizations are asking their employees to start

thinking at a system level earlier in their careers. This is a problem because experience is essential to be able to do that. Pronk observes that organizations are paying more attention to this in their training and course offerings. “But more effort is needed there.” The nice thing, he says, is that it’s easy to spot system thinking talent. “They’re curious by nature. You recognize them pretty quickly.” What was the big challenge for system architects in medical systems? “Optimization. To achieve maximum results with minimal development. Diagnostic devices are large machines and have an almost unlimited demand for new functionality and expansions. You always have to keep up with technological developments: new generations of processors and new programming languages.” With the digitalization of medical devices, functionality grew overwhelmingly in the nineties. “Every-

thing was possible and we knew that competitors were also chasing after all these new possibilities. In the meantime, we had limited resources, people and knowledge. To fulfill all these wishes, you have to keep innovating your platform. First, you make a coffee grinder with some mechanics, then, suddenly, chips and software have to be embedded. Soon, the 286 microprocessor and your real-time operating system are no longer up to date. That also means you have to work on architectures for your platforms.” In the nineties, the Fusion project was launched at Philips Medical, with the ambitious goal of covering the entire X-ray portfolio with a single platform. The sharing of operating systems and the reuse of software components had to justify this extensive operation, but at the end of the decade, the effort collapsed. “One of the solutions to reduce the need for development capacity is to

work with platforms and share the software that runs on it. But that’s complex. That was the problem with the Fusion project: we were trying to cover the entire X-ray portfolio. In the end, it was reduced to cardiovascular intervention, to the Allura systems (which support interventions like implanting stents, RR).” Pronk sketches the R&D rat race for billion-dollar markets where technological opportunities are constantly increasing. “Because of the digital explosion and the necessary software, the development demand is endless. If we at Philips Medical had had ten times as many people or had been able to develop ten times faster, we could have used it all.” “In software, there have been waves and trends to deal with this growth. Platforms and re-use are examples of this. It’s a treadmill. You have to keep running, deliver sufficient functionality and at the same time keep your system technologically up to date. If you don’t do that, 4 49

THEME SYSTEM ARCHITECTING you keep leaning on your old system and at some point, there’s only one solution: starting all over again. That, too, is often an almost deadly action. The Fusion project was such a moment. We had to take a big technological step and nearly suffocated.” How do system architects ensure that they keep an open view of their world? “To be perfectly honest, at some point, you have to leave your organization. System architects run the risk of becoming demigods after twenty years. The real risk of such a guru status is that people will no longer question you, even if you say that the earth is flat. Moreover, they hinder the continued growth of others.” Pronk himself moved from the medical world to the semiconductor division of Philips and later to Signify, the spunoff lighting business. These kinds of moves provide new experiences for system architects but are also a major hurdle. “You may know the tricks of the trade, but you have to build up the domain knowledge for a totally different market from the ground up.” For the new environment, such a switch also means a substantial investment. “They’re attracting a very expensive colleague who won’t be productive for a while.” At multinationals such as Philips, this job switching is policy. Pronk acknowledges having had exploratory talks with other employers during his time there. During these discussions, potential new employers always asked how long it would take before he would be worth his salary again. Pronk: “For very complex developments in a specialized branch, it can take one or two years before you’re really back on track.” This isn’t only depressing for the company, but also for the person involved. The new environment sets high expectations. On a new site, system architects have to win the confidence of a team all over again. “I once spoke to a colleague who had just come from another location. He hadn’t proven himself yet and 50

thought it was terrible not to receive the respect he was accustomed to. That’s what’s stopping some people from taking this step.” How about your switch to semiconductors? “Chips for the consumer market was a different world compared to medical devices. At Philips Medical, we sold three hundred expensive systems a year. At Semiconductors, we shipped 10 million one-chip TVs in just a few months. You can’t enter a hospital with bad stuff. Diagnostic instruments have to have the right functionality and reliability. Delivery time is important, but if you end up six months late, you won’t go bankrupt. Only when you sell rubbish, you have a real problem. In medical markets, the order of priority was: quality, functionality, costs, delivery time.” Pronk points out that reality is completely different in the semiconductor market. “You simply have to be on time to enable customers to deliver at Christmas. Otherwise, your chip won’t make it into their TV design and you miss a year’s revenue. Chip factories keep running, costs are huge. That means that time to market comes first, before functionality and quality. Really a totally different mentality.” This also results in other development considerations. “For three hundred machines a year, the total size of the code isn’t that important. For one hundred million pieces per year, that’s completely different. In that case, it’s worth using a bunch of extra software developers of ten thousand euros a month each to squeeze the software in the cheapest memory possible.” “Yes, of course, I overlooked a lot,” Pronk responds to the question to reflect on his biggest mistakes. Over the years, he says he has become more interested in non-technical aspects. “By that, I mean that your organization has to be fit for the task. You can make an architecture, but you also have to have the people to carry out the development and production.”

You also cause problems if your architecture is at odds with the organization. “For example, if you merge or replace functionality that was built by two departments before, you can be sure that there will be conflicts and struggles. The introduction of an architecture is therefore not possible without adjustments in the organization. If you have to change the existing organizational structure, or worse: if departments become superfluous, then there’s work to be done.” Experience taught Pronk to make more down-to-earth decisions. For example, in the past, he had a strong inclination towards making platforms very future proof by taking into account a lot of reuse. “That meant many layers of abstraction. Over the years, I’ve come to realize that you tend to build in too much extendability too quickly. You’ll never use 90 percent of those extra features again. But what’s much more annoying is that I always found out that the extensions I needed later weren’t there. Customer or product managers always come up with questions you didn’t foresee. My point is: the system architect doesn’t have a crystal ball either. You often put a lot of money and effort into a future that doesn’t come.” The 10 percent extensions that do turn out to be useful don’t make up for the 90 percent unnecessary costs? “That’s the question, of course. I don’t dare pass judgment on that.” Pronk doesn’t have to think long about examples of decisions that have turned out surprisingly well. “I’ve never regretted switching to standard components. At the time of the Fusion project at Philips Medical, for example, we replaced dedicated operating systems with Windows. At Philips Semiconductors, we switched to Linux for TV.” Those choices weren’t obvious at the time. “Both at Medical and Semiconductors, we had strong internal headwinds. There’s always hesitation. Are Windows and Linux

As technicians often experience: more nagging on your mind.”

suitable? The general view was that they would be too big, too heavy or not stable enough. But it paid off. It almost always does,” he says. With some self-mockery: “Once you get your own stuff stable.” So it’s important to use parts of the shelf wherever possible. “Even if standard components don’t seem quite fit for purpose at first glance. It often delivers a lot of benefits. I’ve never regretted that. By the way, the worst thing you can do is rebuild standard components. Then you have the worst of all evils.” As a system architect, you fight against the urge of your colleagues to add their own tastes and fun ideas? “There’s often a very big tendency to add bells and whistles. You should limit that as much as possible because particularities make it expensive. But I’m a technician myself, and I admit that I, too, sometimes have a blind spot for that sort of thing.” Once a colleague said to him, “Ben, I’m not worried until you start to worry.” Without saying it in so many words, Pronk clearly sees the remark as a compliment. It says a lot about the position and responsibility of the people within the organization who keep the technical overview at the highest level.

The disadvantage is that the top architects spend a large part of their time managing managers. “The higher up in an organization you are, the more you’re busy keeping the management calm. You’re the first point of contact for all technical problems. If a difficult project makes little progress, you have to sit down with the CEO every day. Fusion influenced the crown jewels of Philips Medical and had the attention at the highest management levels. When I came to work in the morning, I’d meet the business unit manager at my office, so to speak. At some point, you’re mainly trying to reassure those people.” “Managers obviously have a frustrating profession. They have very little grip themselves and have to motivate others to execute. If things don’t go well, they’re going to be upset, asking for three reports a day. The senior technicians and the project leaders are the ones who suffer the most from that behavior. They have to keep their bosses well informed and ask for their backing.” “As a technician, you really need to want this and be able to deal with it. Gaining confidence and speaking a language that non-technicians understand is a skill you can develop, Pronk says. “Not everyone likes that. It’s not just technical in the long run. It’s also about strategy and budgets.

Where can things go wrong in contact with managers? “Transparency is essential. You have to provide inspiration, but at the same time, you have to sketch a reliable picture of product development. That means balancing. No political correctness, but also no doom stories. You always have to paint a realistic picture: the biggest risks, the delays. You have to include managers in scenarios and point out to them that things get technically deadlocked if they don’t make specific interventions or investments.” “It’s very important to adjust your tone. With a CEO without a technical background, you don’t have to go into the details. Don’t shout at young technicians that we’re going to go bankrupt next month if things continue like this. You have to convey a certain degree of urgency, but not sow panic.” Pronk underlines that it matters what kind of manager is sitting across the table. “There are big differences.” He experienced seasoned marketing managers who understood how the R&D of a large organization worked, but also the ‘boys in shorts with walnut spectacles’ who mainly went for a fast career. Customers didn’t respect the latter group. Clients often ask the system architect to be present in discussions to avoid elaborate marketing stories. “Because they know system architects will tell them what’s really going on,” says Pronk. At the Bits&Chips System Architecting Conference, 24 September 2020 at Igluu in Eindhoven, Ben Pronk will deliver the keynote speech. He’ll talk about what has actually changed for system architects and what the constant factors are. Pronk will also extensively discuss whether there’s still room for system engineering with the speed of product development and digitalization. The complete interview is available on bits-chips.nl. 4 51

THEME SYSTEM ARCHITECTING

Back to basics with Gerrit Muller

TRUE ARCHITECTS UNDERSTAND THE ART OF OMISSION AND KNOW WHERE TO DIG DEEPER System architects are in increasing demand in the high-tech industry. They provide focus, overview and results in complex development projects. This means value for customers and euros for their own business. We ask Gerrit Muller, founder of the Sysarch training courses at High Tech Institute, about the secrets of good system architects. René Raaijmakers

he image that most people have of system architects resembles that of architects of buildings and constructions. They expect these professionals to divide complex machines or products into parts, give them properties and define the interfaces between them. It all boils down to sketching and drawing. In practice, these tasks are also the most visible. In buildings, but also in the technical industry, where sketching is expressed in block diagrams, CAD drawings or piping and instrumentation diagrams. All the parts are made visible and you can see how things connect. An architect indeed has to make a system or product transparent. But that’s only the basis and not what the job is really about. “If you cut it up into pieces, look at those pieces and at the connections between them, all you have is a static image,” says Gerrit Muller, professor at the University of Southeast Norway in Kongsberg and founder of the Sysarch training courses at High Tech Institute. Of course, drawings are useful. “The interfaces allow us to disconnect the components. They’re important and interfaces need to be well de52

fined, but with them, you still have a collection of parts, a box of parts.” The problems, Muller explains, arise when those parts start interacting with each other. “That’s where the value of the system is. Because together, they take care of the intended function and together, they do it well enough, accurately enough, fast enough, reliably enough, safe enough – a lot of that kind of namable qualities.” Behavior and qualities, therefore, stem from the parts that are interacting with each other. “As a system architect or systems engineer, you design to get the desired behavior and the desired properties and you prevent undesired behavior and annoying properties.”

Far from trivial

But in practice, this interaction is so complex that we can’t foresee and understand everything. “Getting the behavior we want is far from trivial. Designing a system without undesirable properties is also far from easy. In the integration phase, when parts are made, usually unforeseen things pop up – you don’t get the desired performance. Usually, things don’t turn out the way you intended.”

The better the system architect, the better he or she can assess whether the design will work? “Yes, but I want to take it one step further. They mustn’t only make estimations but also be able to visualize and communicate. That can be done with sketches and models. The goal is to communicate with many stakeholders such as designers, product managers, customers, the boss and other architects. Good architects make the system explicit and thus discussable and reasonable. In this way, they ensure that everyone can think about it and contribute their ideas. For example, by asking questions such as: suppose we do this or that, what will happen? This leads to better decisions in the design or specification. Making this communication optimal in teams and companies is the core function of the architect.” As an example, Muller remembers a description Guido de Boer made when still working at ASML. De Boer wrote down on paper the path a silicon wafer travels through a lithographic stepper: via the wafer handler and wafer stage, including all operations such as moving, measur-

“The challenge is to make the events visible that have the greatest impact”

ing and exposure. He called the story “Life of the wafer.” “‘Life of the wafer’ was a set of drawings that showed what was happening. It helped to understand what happens to a wafer, during alignment, measuring the profile and all that sort of things. The downside was that everyone used it to discuss their problems, precisely because it was such a handy tool.” This proved to be ineffective. “To discuss a so-called aerial image, for example, it’s useful to know what happens in the light path of a stepper: from the light source via the illuminator, mask and lens to the photoresist. Such descriptions of dynamic paths next to each other provide a great deal of insight into how a system works. They provide understanding and the opportunity to discuss and reflect on the whole. To grasp the dynamic behavior, you often need a whole lot of complimentary drawings or models. This way, you make the whole system discussible.”

ronment. In that infinite mountain of interactions, you want to visualize the context. This means making the events visible that have the greatest impact. It means the system architect has to know what he can delegate to others and what he can ignore because it will have too little impact. That’s where true architects come in, the professionals who know where to dig deeper and who understand the art of omission.”

Everything to get a better grip on the dynamic behavior? “Yes, because there’s infinite dynamic behavior of a system and its envi-

Hiding the details is part of getting to grips with the complexity. System architects are aware of the software stacks, circuit boards and chosen al-

How does that process of omitting work in practice? Muller explains that this is a real art because system architects operate in an environment with a lot of noise. “There will always be a team member asking for more detail, while someone else is shouting that his part isn’t visible. But as soon as you see too much, details start to dominate and the function and application disappear to the background. You no longer see how it works and what the effect is.”

loys, they can also discuss them with their software engineers, electricians and mechanics, but they shouldn’t exaggerate. They’re forced to focus on the more abstract levels. The first level is quite recognizable for everyone, that of the modules, units or subsystems. “Whatever you want to call it,” says Muller. “It’s the things that are produced, that you can touch. These fit well into the mindset of technicians. In lithography, for example, these are units like a stage, a wafer handler or a lens.” On top of that comes an abstraction layer, which usually is about functionality. “Placing the wafer or moving a wafer plane.” On the level above that, the qualities are discussed. “Good overlay, good depth of focus, speed – that sort of thing.” Then comes the layer where the qualities come together in properties of the application. “Those are the things your customers are waiting for, such as yield,” Muller points out. “So you need to understand what role that depth of focus has and what depth of focus is exactly essential and what deviations the patterns on a processed wafer may have. At that level, you place everything more in context.” According to Muller, system architects should be able to switch between multiple viewpoints at all those levels. “Is your product about speed or accuracy? If it has to be accurate and fast, exactly how accurate and fast? I can make something fast or super accurate, but most of the time, you want both speed and accuracy. Then you have to find the sweet spot – that’s what it’s all about.” How do you recognize the potential system architect? “I’m sorry to say, but I don’t have a recipe for that. I know good system architects. They’re often peculiar figures, each with their own qualities. They often entered into the profession from different angles. In the first place, they’re generalists by nature. They shouldn’t shy away from the broad picture, never be afraid of things they don’t know or that are 4 53

Object-oriented analysis and design – fast track

ELECTRONICS

12 – 15 October 2020 (4 consecutive days)

Signal integrity of a PCB

Software engineering for non-software engineers

21 – 23 September 2020 (2,5 consecutive days)

Online

Starts 29 October 2020 (2 evenings sessions)

Electronics cooling thermal design – Online

Multicore programming in C++

2 – 6 November 2020 (5 consecutive afternoons)

2 – 4 November 2020 (3 consecutive days)

Ultra low power for Internet of Things

Design patterns and emergent architecture

5 & 6 November 2020 (2 consecutive days)

9 – 12 November 2020 (4 consecutive days)

Switch-mode power supplies

Speed, Data and Ecosystems

Starts 11 November 2020 (2 modules of 3 days)

18 & 19 November 2020 (2 consecutive days)

Power integrity for product designers

Good software architecture

17 & 18 November 2020 (2 consecutive days)

Starts 30 November 2020 (4 consecutive days)

Thermal design and cooling of electronics workshop

Modern C++

18 – 20 November 2020 (3 consecutive days)

Online

Starts 7 April 2021 (4 days in 2 weeks)

Advanced thermal management of electronics – Online

Secure coding in C and C++

8 – 11 December 2020 (4 morning sessions)

12 – 14 April 2021 (3 consecutive days)

Design of analog electronics – analog IC design Starts 1 February 2021 (11 days in 18 weeks)

New

Solid State generated RF & applications 3 – 5 March 2021 (3 consecutive days)

SOFT SKILLS & LEADERSHIP Online

28 September 2020 (1 day)

Advanced feedforward & learning control

Time management in innovation

30 September – 2 October 2020 (3 consecutive days)

Starts 8 October 2020 (1,5 day)

Mechatronics system design – part 2

Effective communication skills for technology professionals – part 2

5 – 9 October 2020 (5 consecutive days)

Advanced motion control

9 – 11 November 2020 (3 days + 1 evening)

26 – 30 October 2020 (5 consecutive days)

Presentation skills for powerful public speaking

Metrology & calibration of mechatronic systems

11 November 2020 (1 day)

27 – 29 October 2020 (3 consecutive days)

Creative thinking – short course

Basics & design principles for ultra-clean vacuum

12 November 2020 (1 day)

2 – 5 November 2020 (4 consecutive days)

Consultative selling for technology professionals

Actuation and power electronics

23 & 24 November 2020 (2 consecutive days + 1 evening)

16 – 18 November 2020 (3 consecutive days)

Creative thinking – full course

Passive damping for high tech systems

24 & 25 November 2020 (2 consecutive days)

17 – 19 November 2020 (3 consecutive days)

Beneﬁt from autism in your R&D team

Dynamics and modelling

3 December 2020 (1 day)

23 – 25 November 2020 (3 consecutive days)

Effective communication skills for technology professionals – part 1

Motion control tuning

23 – 27 November 2020 (5 consecutive days)

14 – 16 December 2020 (3 days + 1 evening)

Experimental techniques in mechatronics

SYSTEM

30 November – 2 December 2020 (3 consecutive days)

Thermal effects in mechatronic systems

System architect(ing) in Eindhoven

1 – 3 December 2020 (3 consecutive days)

28 September – 2 October 2020 (5 consecutive days)

System requirements engineering improvement

OPTICS

1 & 2 October 2020 (2 consecutive days)

Modern optics for optical designers – Part 1

Introduction to deep learning

Expected January 2021 (15 weekly morning sessions)

8 October 2020 (1 day)

Modern optics for optical designers – Part 2

Design for manufacturing

Expected January 2021 (15 weekly morning sessions) Starts 26 October 2020 (15 weekly afternoons)

SOFTWARE

21 September 2020 (3 morning sessions + e-learning)

How to be successful in the Dutch high tech work culture

MECHATRONICS

Applied optics

Presentation & body language skills – Online

Starts 15 October 2020 (3 days in 3 weeks + assurance session) New Location

System architect(ing) in Leuven (Belgium) 16 – 20 November 2020 (5 consecutive days)

Introduction to SysML 4 March 2020 (1 day)

Object-oriented system control automation

Systems modelling with SysML

Introduction to deep learning

Value-cost ratio improvement by value engineering

Starts 17 September 2020 (2+3 consecutive days) 8 October 2020 (1 day)

12 – 15 April 2021 (4 consecutive days) 20 & 21 May 2020 (2 consecutive days)

hightechinstitute.nl

THEME SYSTEM ARCHITECTING

out of their sight. They shouldn’t shy away from new things. In fact, they should be energized by them.” “An architect is somebody that everyone can talk to. Imagine a large building with a room where colleagues always drop by. That’s probably where the system architect has his desk, even though he may not officially have that job title. The interaction with him is a naturally occurring phenomenon in the team because others experience that this person helps them.” If a company doesn’t have a system architect yet, this is the person to look for? “Exactly. If you put the system architect’s profile next to that, as we define it in the Sysarch course, it usually matches nicely. One more thing: system architects are always multitasking.” What exactly do you mean by that? “Being able to continuously change viewpoints, as we call them. Looking at a problem from different angles. You can learn that or might be forced to learn it. This multitasking is essential but can be very tiring. Some people are very good at systems thinking but are completely lost when they need to multitask.” What are the biggest challenges for people who are new to the role? “People often concentrate too much on the system and the technology. You have to help them get out of the system and into the world of customers, the product lifecycle and the business. They need to get out more and they need a push to do so. Communication or soft skills are also useful.” The complexity of systems is increasing. Does this increase the need for system architects?

“You’d hope the problems of twenty years ago are so well known that we can now solve them in a more structured way. That would enable today’s architects to focus on the more complex problems. There’s almost no system that’s not connected to other systems anymore. There are almost no functions and features that don’t depend on multiple systems anymore. I need to understand the system I’m working on but also other systems, including the interaction and the people around it. That complexity, that growth, that’s a fact of life.” You’re a professor in Norway and are working one day a week at ESI in Eindhoven. What’s the nature of the problems that companies ask you about? “All questions that are also in the Sysarch training. What’s the role of the architect in my organization? How do I take long-term strategy into account? How should I help architects do their job in the best way?” “Some companies say right away: I want to do model-based system engineering, MBSE. Then I’m always curious about their real question. Do they have an administrative need? Do they have to comply with the rules imposed by the American FDA? Or do they need to investigate or communicate better? You can model for many different reasons.” “A lot of companies are struggling with the same question: they want to create a platform because they have products 1, 2 and 3 with a lot of synergy between them, but all different. Or they constantly have projects to make different product variants. Platforms, standardization – I often get questions about that. For an architect, this is a balancing act because standardization can make things rigid, thereby reducing the value for customers.”

“Imagine a big building with a room where colleagues always walk in. That’s probably the office of the system architect, even though he may not officially carry that job title”

Can the knowledge in the field of system architecture be packaged in manageable chunks? “This begs the question: what’s the ability and what’s the art? What can we offer people in terms of methods and means, and what can’t you transfer as a teacher? Competent system architects have gone through quite a development. That’s an accumulation of time and experience. But if you’ve done something for a long time, it doesn’t mean you’ve developed the skill. Seasoned experience is needed. It’s about recognizing situations and thinking about them. Knowing why some things don’t work because you’ve experienced it and the next time, you know how to do it right the first time. Such a cycle of reflection is actually essential for a system architect to learn and reach a useful level of experience.” Want to know more about system architecture and systems engineering? Visit the corona-proof Bits&Chips System Architecting Conference on 24 September 2020 at Igluu in Eindhoven or check out the possibilities for system architecture trainings at High Tech Institute. 4 55

INTERVIEW MARTIN LANGKAMP & MARTIJN BOUWHUIS (IMS)

INNOVATION AND CHARACTER LIGHTS THE PATH TO IMS SUCCESS In today’s high-tech environment, companies of all sizes are looking to stay at the cutting edge of innovation. According to team leaders Martin Langkamp and Martijn Bouwhuis of Almelo-based IMS, the equation is easy. It comes down to a few key factors: keeping the employees interested, keeping the workplace light and focusing on personal development through training. Collin Arocho

utch innovation in the hightech sector comes from businesses of all sizes. While big names like ASML and Philips are recognized around the globe, there are also several small and medium enterprises (SMEs) in the Netherlands playing a big role in global high tech. Take, for example, Almelo’s IMS. IMS has been around for just over 20 years, opening its doors in 1999 after it was spun out of Texas Instruments through a management buy-out. Now, in 2020, the automation and technology expert has delivered more than 750 production lines with an emphasis on the medical device, smart device and automotive domains. “We’ve grown a lot since the early days. Now, we see our role as helping our global customers realize their production goals,” explains Martin Langkamp, technical sales coordinator at IMS. “We do that by delivering our innovative machines all over the world that excel in the high-volume production of small, precise and sometimes extremely complex products.”

Character

While IMS’s global customer base is certainly large, the company itself has a relatively small footprint – employing more than 120 people at its 56

Almelo and Groningen locations in the Netherlands. Despite its small stature, it’s having a big impact on consumer electronics. Currently, the high-tech machine maker is active in delivering machines used in the assembly process for the smart device and automotive sectors, in addition to next-generation headlights and sensors for cars. “The character of IMS is that we’re always focused on innovation, not just locally, but globally,” highlights Langkamp. “That means we do a lot of international projects, which offers our engineers exciting opportunities to travel, learn and share knowledge. That’s part of our DNA.” Another focal point in the character of IMS is the focus on the personal development of its employees. “One of our main focuses is on continuing education for our workers. We find that trainings, workshops and conferences are a great way for our engineers to develop both personally and professionally,” comments Langkamp. “In fact, as we look to the future as we continue to innovate, the necessary competencies of a position can expand and the engineers may be guided to specific courses to bolster their skills. We actually use education-based developmental plans in our evaluation pro-

cess, to help people and the company meet our goals.”

Modularity

Recently, IMS found a golden opportunity to utilize training. Looking to continue to grow and push the cutting edge of complex part manufacturing, the company took on a new role for its customers, helping lead them in the design of production machines by offering series-based machines, rather than one-offs. “For many years, R&D operated more reactively for development, finding solutions for the customers as they arose,” recalls IMS R&D team leader Martijn Bouwhuis. “More recently, however, we’ve started to adopt new methods to become more proactive in the process and we’ve focused our efforts into making standardized products that can be tailored to fit our individual customers.” To get these standardized products, IMS decided that modular thinking was the best way to achieve the new goals and it started laying the foundational work to get its workforce aligned on the idea. However, it was during the Bits&Chips System Architecting Conference, the team found that their modular approach fits perfectly with the principles of system architecting. Langkamp: “For a few

Credit: Fotowerkt.nl

Martijn Bouwhuis (left) and Martin Langkamp (right) sometimes refer to IMS as a high-tech playground for engineers.

years, we’d already been adjusting our processes, but we were looking for a better structure with more continuity within the whole of the company.” Bouwhuis: “While we were assessing the best way to progress, we found that often in the design process we would focus on subsystems because that’s where the value was added. Somehow, we forgot to look at things from a system level. But as the complexity of the parts our machines are making continues to explode, it’s clear that software engineering has become more important than ever and it was time to update and professionalize our working methods.” Rather than sending a few team members to a relevant training, IMS reached out to High Tech Institute to develop its customized in-company edition of the System Architecting training, allowing the Almelo-based company to bring in a broad and diverse group of its team. “It’s important in our transition to establish cohesion among all the different disciplines and departments,” says Langkamp. “From mechanical to electrical and software engineers to

the sales team, the goal was to get everyone on the same page, thinking at a system level.”

Added value

“The reason we selected High Tech institute was because of the strength of its instructors. Their knowledge and expertise matched our needs precisely,” emphasizes Bouwhuis. “What we appreciated the most was that the trainers found ways to trigger discussion, which got our group of about 12 trainees really participating. This interaction between the team and the instructors, all with different perspectives, really enhances the training with a lot of added value.” Does IMS use training to attract or keep its skilled engineers? Is it difficult to compete with larger companies in the high-tech domain? “Yes and no. Yes, training and education opportunities are a great tool to attract and retain our engineers. But, as far as competing or losing our skilled workers to the bigger companies, no, that’s not the case. In fact, I think the size of IMS, the scope of our

work and our approach is something that draws people to us and makes them want to stay,” illustrates Langkamp. “In the Brabant region, it’s pretty common for engineers to bounce around from place to place, but here at IMS and in the Twente region in general, it’s just not as common.” “Because we’re small, we’re able to keep things light and fun in the workplace. Of course, we’re extremely professional in working with our customers. But the people here are more than just a number and embracing that mentality means we can operate as a family and have fun,” adds Bouwhuis, joking: “Sometimes we refer to IMS as a high-tech playground for engineers.” “Yes exactly. Because of our roots from Texas Instruments, we sometimes joke about having people working here for 40 years, but the company is only 20 years old,” laughs Langkamp. “By keeping our people interested with exciting projects, a light-hearted informal workplace and a focus on our workers and their development, IMS is in a strong position to continue innovating.” 4 57

EVENTS AS A SERVICE Not having real-life events doesn’t mean you can’t stay in touch with your (potential) clients. Hire Bits&Chips to organize your (online) event or webinar. We’ll arrange the registration and hosting and provide online promotion to make sure your event becomes a success. Contact us via events@techwatch.nl to receive more information about the possibilities and rates.

bits-chips.nl/events

UPCOMING ISSUES

Technologies for the IoT

Bits&Chips 5 | 23 October 2020

Bits&Chips 6 | 4 December 2020

Connections between devices are largely wireless these days. Aside from data exchange, RF technology is also increasingly used to transfer energy, to charge batteries and even to heat food. This issue focuses on recent developments and applications.

From consumer devices to vehicles and machines: these days everything’s plugged into the Internet. The IoT is playing an increasingly central role in our daily lives. This issue dives into the current state of the technologies that make it possible to connect everything to everything else.

Interested in contributing? nieke@techwatch.nl

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 Jessica Vermeer – editor tel +31 24 3503534 – jessica@techwatch.nl Alexander Pil – editor tel +31 24 3504580 – alexander@techwatch.nl

Administrative Mathilde van Hulzen – finance tel +31 24 3503532 – invoices@techwatch.nl

Design Justin López – graphic designer and illustrator tel +31 24 3503532 – justin@techwatch.nl

Publisher René Raaijmakers tel +31 24 3503065 – rene@techwatch.nl ISSN 1879-6443

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

Advisor Maarten Verboom External staff Ilse de Boer, Femke Veldhuis Contributing writers Wim Hendriksen, Joost Kauffman, Jan Kuper, Daniëlla van Laarhoven, Cees Links, Marcel Pelgrom, Anton van Rossum, Han Schaminée

Publisher in Belgium René Raaijmakers Biesheuvelstraat 1 2370 Arendonk, Belgium Printer Vellendrukkerij BDU Barneveld

Bits&Chips membership Bits&Chips is free of charge for delivery to addresses in the Netherlands. You can also join as a premium member for 59 euros per year. That gets you (in addition to the magazine) substantial discounts on admission to Bits&Chips events held by Techwatch BV. For companies we have a business membership costing 159 euros per year. That includes two copies of Bits&Chips sent to the company address, plus two discount codes for every Bits&Chips event. To request a membership, please visit bits-chips.nl/subscribe or write to info@techwatch.nl. Back issues: write to info@techwatch.nl.

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.

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.

Cover Main photo: Philips/Thales

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 4 September, 23 October, 4 December Copyright All rights reserved. (c) 2020 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.

bits-chips.nl

24 SEPTEMBER 2020 • IGLUU • EINDHOVEN PROGRAM

Silver sponsor

Ke y n o t e

Bits&Chips 4 | 4 September 2020 | Trends in software development

Articles inside

Innovation and character light the path to IMS success

keeping management at ease” True architects understand the art of omission and know where to dig deeper

High up in an organization, you’re busy with

High-level FPGA programming for nanosecond 44 High-level FPGA programming for nanosecond timing in terabit communicationtiming in terabit communication

A pressure cooker for software talent40 A pressure cooker for software talent

No systems engineering without digital engineering32 No systems engineering without digital engineering

Optimizing your high-tech development and machine performance42 Optimizing your high-tech development and machine performance

Cracking the code to craftsmanship36 Cracking the code to craftsmanship

Performing an ultrasound on a wafer to align it

Pandemic accelerates ASML’s adoption of AR

Surveilling aging infrastructure using fi ber-optics

Has Intel lost its mojo?

US-Chinese trade war leaves semiconductor industry in limbo

Pink RF and America’s Odyssey link to bring RF repair to Nijmegen

Noise

Holst betting big on the foil battery

Dutch silicon anodes inch closer to the gigafactory