Category: Interviews

Applying iterative learning control can improve the performance of motion systems tremendously. Often, by as much as ten-fold. However, the approach is fundamentally different to that of existing feedback and feedforward techniques. In order for it to be implemented correctly, a thorough understanding of the underlying learning mechanisms is required. The Advanced feedforward & learning control training provides the tools with which control technologists can understand and apply iterative learning control techniques.

.
Photo: Tom Oomen, with the steering mechanism of a desktop printer that, despite numerous limitations, has reached an astonishing level of performance due to an iterative learning controller.

Control experts in the industry are also looking at this, says Tom Oomen, from the Control Systems Technology group (TU Eindhoven). He shows us the steering mechanism of a desktop printer that has been converted by a lab assistant into an experimental setup. ‘This is a simple arrangement costing roughly one hundred euros, with a lot of friction and cheap mechanics,’ says Oomen. ‘But despite these limitations, we can still achieve perfect performance with iterative learning control. The measuring accuracy of the printhead is 42 micrometres and with our learning management we stay within that limit.’ He can hardly suppress his enthusiasm: ‘That’s just amazing. Motion control experts also wonder how this can be achieved. Participants in the Advanced feedforward control course will also work with the same printer.’

The Control Systems Technology section, headed by Maarten Steinbuch, at the Technical University of Eindhoven has a long tradition of working with high tech companies. Researchers are working with ASML, CCM, NXP, Océ, Philips Innovation Services and Philips Healthcare on control through iterative learning control. The High Tech Systems Center (HTSC) also plays a role in this. The high tech companies implement and validate the developed algorithms directly on their systems.

Oomen: ‘They all really want to apply iterative learning control, because that’s the way to significantly improve performance. One company wants to achieve nanometer accuracy, the other higher productivity or to employ cheap sensors, actuators or mechanics. Because qua implementation, learning control is very cheap. The solution shifts from expensive measuring instruments and drives to software and smart algorithms. Our starting point is that measurement data are very cheap and that we can full compensate for what is reproducible in the measurement data.’

Back to basics. Everyone in the control world of the PID-controllers knows about feedback. Feedback generates a correcting action based on a servo error. It is a correction made subsequent to the error. PID controllers are popular because the system is in a ‘closed loop’ and is thus insensitive to changes. Another plus is that developers can work very intuitively with the help of rules of thumb.

The performance of these traditional control systems can be improved considerably by adding feedforward technology. Feedback control is, after all, like shutting the stable door after the horse has bolted: the corrective action follows the error. With the feedforward method the system anticipates future disturbances. If the trajectory is known immediately beforehand, the controller can use this knowledge to significantly reduce the tracking errors. ‘That typically leads to a performance that is ten times better,’ says Tom Oomen. Therefore, it is not surprising that now almost everyone uses feedforward.

But we can top that. By using the repetitive character with which many mechatronic systems work, control systems are able to learn from previous tracking errors. The result is another step forward in performance. Oomen: ‘Compared to traditional combined feedback-feedforward designs, the tracking error can be improved by a factor of ten or more.’ This is the basic idea behind Iterative Learning Control (ILC), with repetitive movements.

Design techniques in the High Tech industry

Learning techniques are central to the three-day Advanced feedforward & learning control training, of which Tom Oomen is one of the trainers and course leaders. ‘On the first day we start with PID controllers and we teach participants why you can only achieve limited performance. Feedforward makes more of a difference, because you give the desired task to your controller and thus look into the future. The idea of ​​learning control is simple: every time you do the same task, you know what will happen in the future. With a good learning control implementation, you can achieve perfect performance. That is also what people in my field say: everything that is reproducible can be perfectly compensated for.’

Photo: Tom Oomen, trainer and course leader of the Advanced feedforward & learning control training course.

Iterative learning control ensures that performance improves step by step. With every experiment, every cycle, the system collects measurement data which is subsequently examined by the learning controller: have I done it better now? If it is perfect, the controller will retain the feedforward control signal. If there is still a fault, a small correction will follow, in order to achieve an even better feedforward control signal. With a good design you see an almost perfect performance after five iterations.’

The methods used in the Advanced feedforward & learning control training course fit in very well with the design techniques already known by the Dutch high tech industry. They make it possible to converge motion systems with learning very quickly. Oomen: ‘That is really different to the rest of the world. You see a lot of academic techniques that need hundreds or thousands of iterations in order to converge.’

The unique Eindhoven approach is based on very accurate models for mechatronic systems. Oomen: ‘The basis for this was laid out in the seventies, eighties and nineties at the Philips Natlab, for example in the development of the compact disc players.’ Things that result from this, such as frequency-response-function identification, loop-shaping of PID controllers, and notch-filters, are now to be found in the basic course Motion Control Tuning. Oomen: ‘In the follow-up training, Advanced feedforward control, we construct the learning control technique from the same philosophy from the very beginning, so that from day one, the participants themselves are able to design and implement a learning controller which gives an almost perfect performance after a few iterations.’

Theory
The second day of training contains a lot of theory. The aim is to give participants a complete picture of what is available in the world in the field of learning control. Oomen: ‘Surf the internet for learning control and you will find mountains of information. Many different types of mathematics, usually from a strong academic perspective. Curious technologists automatically ask themselves: Why don’t we apply this?’

There is a world of difference between the alternative mathematical descriptions and the techniques presented to students on the first day. Why then all that hard work? ‘We expressly want students to experience how alternative methods work together in a mathematical way,’ says Oomen. ‘That is indeed quite difficult for most students, because they often have to refresh their underlying mathematical knowledge. Nevertheless, we confront them with it and drag them through the methods consciously, so that they can understand these other approaches and are able to put them into practice.’

Worldwide publications about control techniques always speak of optimal design algorithms. ‘Almost everyone in our profession is working on this,’ says Oomen. ‘That is certainly in line with the criteria they set. We are also going to work on it. Participants in our course experience a lot of parameters that are relevant. Our experience is that understanding all these parameters and how they influence performance is complicated. By letting these participants experience it for themselves, they gain all the knowledge needed to assess what it is that specific algorithms can or cannot do together with their respective advantages and disadvantages. This gives participants the feeling that they can oversee the entire field of iterative learning control. Especially if they want to delve into it more deeply.’

You say that you have to drag participants through the theory and mathematics. Does that always work?
That always works. And once you’ve seen that, you can pretty well see the range of techniques that are available. It is not about reproducing formulas, participants need to know what is behind them, what the basic ideas are and how they can use them. Once they have done this part of the course, the rest is really easy. It’s quite something that they are able to implement Matlab code in two lines. But the most important thing is that participants can substantiate the advantages and disadvantages of specific techniques. It is also nice to have useful knowledge to bring to day three, where we examine recent developments and use automated feedforward tuning.’

How much experience in control technology do participants need to be able to do the Advanced feedforward control training?
‘People with experience in designing controllers and motion systems automatically qualify for this training. This applies to most people who design feedback controllers in the Eindhoven region. They know how to design PID controllers and also what state-space, loop-shaping and filtering techniques are and how to think in the frequency domain. A little Matlab knowledge is also very useful.’

‘The basic knowledge needed is in fact the basic training to be found in the Motion Control Tuning course. The description gives a clear picture of the expected prior knowledge. People may therefore draw the conclusion that they should first follow the Motion Control Tuning training course.’

Can you give some examples of the type of participants in the Advanced feedforward control training sessions?
‘Participants vary from young people who have just left the lecture rooms, to motion control tuning experts who have been working in the industry for twenty years and who design controllers every day. For example, motion control experts from ASML, K&S, Nexperia Itec or Océ who are not yet familiar with iterative learning control. And also, technologists who have already experimented with this new technique in their work and who are interested in it. Among them are many small companies that want to apply iterative learning control. After three days they get a sense of what it can offer for their machine and also, right away, a basic implementation so that they can directly get started on their own machines.’

Before, with regard to the number of iterations, you mentioned five cycles. Does it matter if there are a few more or a few less?
‘The great thing about learning algorithms is that they adapt themselves when the situation changes. If temperature plays a role, for example because of a day-night rhythm, it is nice if the system adjusts within a few iterations. If, for example, 100 iterations, of one hour each, are required, this can lead to production failure.’

Oomen gives another extreme example. In collaboration with the researchers from the Eindhoven research institute Differ, the motion control group applied learning techniques to nuclear fusion experiments for the Tokamak reactor (TCV) in Lausanne. ‘Physicists have traditionally relied on complex physical models to simulate these fusion processes. There is a big gap between the use of data and control technology. My colleague Federico Felici has expertise in nuclear fusion, in addition to a background in iterative learning control technology. He is now dabbling in that world from his technical background.’

In Tokamak reactor experiments it comes down to plasma formation by means of the correct actuator signals. Such a shot takes a few seconds and is very expensive. ‘Because a complex computer simulator had already been developed, we were able to calculate how to adjust the signal to make it better. To do this, we linked the simulator to the measurement data from the experiments. It turned out that with our iterative learning control technology we had an almost perfect control signal within a number of iterations. That has had a lot of impact in the nuclear fusion world.’

 This article is written by René Raaijmakers, tech editor of Bits&Chips.

High temperatures lower not only the lifetime but also the performance of many electronics. A clever, cost-effective thermal design translates into a commercial advantage. ‘Your competitive edge increases  if you can deliver performance at lower costs or offer more performance at the same price through better cooling’, says Wendy Luiten, who delivers the Electronics cooling thermal design training.

High Temperature impacts the performance of everything that depends on computing power or memory. It is also an important factor for the performance of image sensors and energy converters, lamps and power supplies.
One of the difficulties in thermal design is that key decisions often go unnoticed. Says Luiten: ‘If you want to do it low cost, many critical cooling parts are not directly visible. They do not show up on your Bill of Materials. That makes it tricky. Passive fan-less air cooling is a preferred low-cost cooling solution, but air is not on your BOM. Neither are the open spaces that accommodate air flow. So, if a mechanical designer incorporates an open space as an air path in the design, it is not registered anywhere by default as a cooling component. That means that at a later time in the design process changes can be made that disable the original cooling concept. It is just not in the documentation, if you don’t put it in specifically.’

Don’t design tools to take this into account?
‘No, not really. Thermal behaviour is the combined result of electronical and mechanical design and the effect of a change on cooling performance will show up in CFD (Computational Fluid Dynamics) Thermal simulations that combine the inputs of both. But it will not show up in stand-alone Electrical and Mechanical design tools. You often see in product development that a thermal simulation is ran based on the finalized mechanical CAD and electronical layout EDA files, but essentially then what you have done is replacing the hardware prototype with a CFD simulation, after the detailed mechanical and electronical implementation has been finished. If the thermal simulation shows a problem with temperature, part of the mechanical or electronical design process has to be re-done. Obviously, that is a waste that you want to avoid.’

How does your training offer a handle on this?
‘We teach a way of working as well as the physics. Thermal design needs to be taken into account starting in the architectural phase and thermal risks addressed pro-actively. If you think the IC might need a heatsink, address the issue and ask the layout to put in heatsink mounting holes.  Do not wait until the complete lay-out is finished and you can do a test.  If layout has just completed a complicated multi- layer board, and part of it has to be re-done in order to accommodate the mounting holes for the heatsink, people are not going to be happy’. ‘So, if you think there is a risk, ask for the holes.  Maybe they are not needed after all, but unused holes in the PCB are no big deal. The other way around carries a much larger project risk: if you do need the heatsink but the holes are not included, the layout will need to be party re-done, and that can cause a development time delay’. ‘With a good thermal design, a lot is possible, but you have to make sure that the thermal concept is sound from the beginning of development to avoid re-designs.’

High temperatures have a negative influence on the life expectancy and reliability of components. Therefore, some chip manufacturers integrate temperature control software in their IC’s. Intel started by incorporating a thermal sensor in its P6-processors that simply shut them down if they heated up too much. Intel’s Pentium 4, Xeon and Pentium M processors had an additional over-temperature protection that slowed down the IC’s clock speed if it got too hot.

But temperature safeguards don’t automatically have a positive influence on the thermal behavior, indicates Luiten. ‘Many people think that using an IC with temperature protection solves the thermal problem, but this is not the case. The safeguard does not cool the IC but typically lowers its energy consumption by lowering the performance.  If the products thermal design is weak at system level the component will get hot sooner and more often and the end performance at system level will be jeopardised.’

Luiten found this out first-hand in a recent consultancy job on an image processing device. As soon as the video processor became too hot, the display went black. ‘It turned out to be an intended feature, not a bug. The video processor had embedded memory that was susceptible to high temperatures so the component supplier had added a temperature protection. As soon as the internal temperature sensor experienced over-temperature, the thermal protection kicked in. Normally this would not have been a problem, but in this case the product was developed for use at higher temperatures so the black screen was an unpleasant surprise.’

Because of this components temperature protection, the thermal design of the product became more critical. ‘If the IC’s temperature went up too high, it switched off. So, the protection seemed like a clever thing to do at component level, but at the same time it made thermal design more critical at the system level. In the end a partial re-design was needed to make sure the product worked way as intended.’

What is the most important topic in the training?
‘Thermal problem solving and better cooling design. Heat is a major performance-limiter in many electronic systems, from computers to lighting. The moment you can cool your product better at the same price point, this translates to better performance at the same price and this is a commercial advantage.’

Where do attendees work?
‘We have people attending from all levels in the system design, from component to module to complete system. Many former attendees worked on components, small electronic products and LED applications, but we have also had people working on large systems, like radar systems or heat sinks in large power supplies and we have had people from cooling component suppliers. With the increase in automotive electronics we also see more people coming in from that field.’

What makes people sign up?
‘Part of them come by word of mouth. Clemens Lasance and myself are both known in the international electronics cooling community and so is this thermal training.  The training is in English and draws an international public. We have had people coming in from as far as the United States that had heard they had to go to Eindhoven in order to get a really good thermal training.’

What makes the training special?
‘Clemens and I really teach electronics cooling in an application-oriented way. The training covers all aspects and is very hands on: we go from high level system architecture to implementation level details such as layout and locations and dimensions of air vents. And this is backed up with physics and best practices.

You don’t learn how to swim by just looking at other swimmers. Our goal is to send participants home with applicable skills, and that includes the hands-on ability to do basic calculations.’

‘I want people to get a sense of sizing, to get a gut feeling for thermal estimations.  If you have this ability, you will be able to take much better design decisions and you will also be much more confident in doing thermal simulations because you better know what is happening.’

Can you bring your own case study?
‘We always ask participants beforehand if they have a case of their own to share. If applicable, we will discuss it during the training. This is fun, because it leads to lively discussions. In addition, we have some standard cases.’

During the training: learn to thermally interpret specifications

Lasance and Luiten discuss the physics of electronics cooling, how to benefit from best practice thermal ways of working and how to implement them during the product development phases. This also explains why time management and project management are part of the course. ‘We discuss specifications and the way to interpret them thermally as well. We have people from system, sub-system and component level – this leads to interesting discussions on the interpretation of specifications.’ ‘The last learning activity of the course is a case study.  We split the group into two teams and spend two to three hours to crack a case. This is an eye opener to many people because it is the first time that they take the factors and specifications at all levels into account and see for the first time how mechanical and electronic considerations interact into the thermal behavior from component to system.’

When you have finished the training…
‘You have applicable skills on estimating thermal effects that you can use to estimate how to cool a product properly. Many participants indicate that this is the unique selling point of our training. Both people that are new to the field and experienced thermal architects and designers comment on how much they have learned during our course and recommend it to their colleagues. In addition, the know-how increases confidence in your results if you happen to do thermal simulations because you better know what you are doing.’

What people attend?
‘Attendees typically have higher education in a technical field such as electronic engineering, mechanical engineering, physics, optics, or industrial design. We often see product developers and architects with a couple of years development experience get into the thermal discipline. In Europe we have no higher professional education or academic education in this field. The United States do have universities that offer cooling of electronics courses, but they have a more academic approach.

Do you update the course regularly?
‘Absolutely. Recently we decided to split the training and you can choose either to go more in-depth with Clemens Lasance or do more hands-on exercises with me. That change was well received. At the moment, I am working on Design for 6-sigma and thermal design. In the thermal architectural phase, you already have to check how mechanics and electronics work together, identify the demands on the thermal behavior and design, optimize and verify how to make it work out. This combines well because of the system level approach in DfSS. And you can get great results in design and optimize phase with the combination of computer CFD simulations and Design of Experiments.’

Photo by: Bart van Overbeeke

Wendy Luiten started her career halfway the 80s as a thermal specialist at the Philips Centre for Fabrication techniques, at that time just split from the well-known Philips Research Laboratory (Natlab). In the late nineties, she contributed as a thermal architect in the development team to the first flat-screen televisions, made by Philips Consumer Electronics. ‘These early plasma screens had fans on board, and the fan noise was not liked so the challenge was to take them out.’ Luiten wrote a paper on the thermal design of the first plasma TV in the world without fans. This won her the best paper award on the Semi-Therm conference.

Cooling of electronics in consumer TV at first was seen as a temporary phenomenon. Managers assumed that existing heat problems would not be there forever. In reality, whenever a generation of televisions had outgrown its heat problems, product managers would pile on new demands on the development teams, going from plasma to LCD screens, the rise of HD TV, LED TV, 3D TV and smart-TV.

‘Every new generation would have its heat problems’, Luiten says. She ended up working sixteen years on a temporary problem. She is laughing about it. ‘I have done a temporary, non-structural activity for sixteen years straight.’

Since 2000, Luiten teaches courses Cooling of electronics. She has been all over the world for this, travelling to China, Singapore, Taiwan, Korea, the United States and several countries in Europe. Luiten also has been teaching in Saudi Arabia, at a summer school in Turkey and last year she presented the pre-conference short course Fundamentals of thermal system design at the European Therminic conference.

Together with Clemens Lasance, Luiten has been teaching the workshop Thermal design and cooling of electronics for fifteen years. This particular training has been split up into two modules and is being held online and Eindhoven, but the two of them were also presenting a pre-conference short course at the Semi-Therm conference in the United States. The combination of Luiten’s years of electronic product development experience and Lasance’s broad and deep knowledge of their discipline works particularly well. This marks their training as distinctly different from other available courses around the world.

Luiten is passionate about her subject, because she loves solving puzzles. ‘Designing the right cooling architecture for a range of TV’s forming a certain generation is similar to high-level puzzling. Together with the electrical and mechanical architect, you have to cover as many models as you can with the smallest set of different components. You need a flexible and scalable cooling strategy to cover the product diversity at acceptable cost.’

Nowadays Luiten is principal of her own firm: Wendy Luiten consultancy. ‘Not quite an original name, but highly practical.’

Getting the thermal design right for the Internet of Things

Thermal aspects are only part of the total problem, but Luiten expects the significance to continue. ‘Electronics cooling is important in the energy transition. Converting solar and wind energy to the grid requires power electronics, and this has thermal limitations. In addition, the materials that convert light to electricity and vice versa also are known to be temperature sensitive.’ ‘The transition to electrical and self-driving cars is also a hot item. Currently electronics make up 30 percent of the total costs for a car, and that is rising.’

Says Luiten: ‘Automotive electronics can be safety critical, failure for thermal or other reasons is not acceptable.’

Meanwhile in data communication, cooling is a well know cost issue ‘Data center cooling can cost as much as processing data, and in telecom 5G is expected to be a thermal challenge as well. For the Internet of Things also, getting the thermal design right is important. In future, there will be sensors everywhere.’

Partnership High Tech Institute

The training Cooling of Electronics is part of the T2Prof portfolio. T2Prof has continued the technical trainings on electronics and optics originally developed at the Philips Centre for Technical Training. T2Prof brings its courses on the market in exclusive cooperation with its partner High Tech Institute. High Tech Institute focuses on the marketing, sales and organization of these courses.

This article is written by René Raaijmakers, tech editor of Bits&Chips.