Systems Designers Must Know ‘All Aspects of Design’
One of the biggest challenges faced by systems designers is the need to understand all aspects of a design environment, according to Keith Ogboenyiya, vice president of Texas Instruments’ Systems Engineering and Marketing. This means understanding design issues such as thermal management, clocking considerations, noise immunity, housing and packaging, printed-circuit-board (PCB) integrity, and environmental conditions. All of these challenges impact semiconductor and electronic component selection and circuit layouts, as well as cost.
Aspencore editors’ recent discussion with Ogboenyiya covered topics such as the increasingly complex role of systems designers; old and new technologies that engineers need to add to their skill sets, particularly in automotive systems; and why on-the-job training and mentorship is important. His expertise has led to advances in several technology domains, including automotive instrument cluster, industrial motor drive products, digital power supplies, solar inverter systems, in-vehicle networking, and other technologies.
In addition, systems designers need to consider new technologies being implemented within their systems to achieve a desired functionality, but that implementation could have an impact on a circuit that has been in use for five or 10 years, he said.
“You may need to re-think or re-layout, or there may be an overall implication because of the integration of these new functionalities,” he said. “Designers have to think through the entirety of the system: not only the newer technologies that are being added but the impact to technology that has been around for a while, which may introduce new constraints and challenges.”
Systems designers also need to deal with older technologies that become “new” again in different applications.
Ogboenyiya explains why high-voltage isolation is changing automotive design across systems. He told us that while high-voltage isolation is nothing new, it is the first time it’s being deployed across automobile systems as carmakers release new electric vehicles (EVs) on higher-voltage platforms. High-voltage isolation will have a big impact on the automotive industry, and engineers should learn about it, he said.
Together with high-voltage isolation, there are fundamental changes occurring in the vehicle architecture and some of the critical components, including the CANbus and current sensing, which are impacted by the move to high-voltage systems. “There are quite a few different systems that now need to be isolated, and car manufacturers and Tier One developers are all in the process of developing this isolation technology and understanding what it means to the system,” he told us.
At the same time, systems designers also have to consider that electric motors will operate at voltages as high as 400 V and 800 V, depending on the size of the vehicle and the amount of load, which translates into different methodologies for driving power to the motor. He explained that designers are looking at more efficient technologies to drive the motor, such as silicon carbide (SiC)-based FETs and even gallium nitride (GaN)-based systems, as systems designers look to drive these motors efficiently with the best power density and control.
Functional safety and how to design for it is also a growing need. Functional safety is a “big care-about” across automotive and industrial systems, Ogboenyiya said.
Several years ago, functional safety was mainly a topic discussed in automotive and in only a handful of automotive systems — things such as airbags, braking, chassis control, and stability. Now, functional safety is a topic that is prevalent across almost every automotive system but also in industrial, where functional safety is becoming more prevalent across a host of different systems, he added.
Engineers also need to keep on learning, whether they are just out of college or decades on the job.
“You have to think about the engineering process as a continuum,” Ogboenyiya told us. “In many instances, new college graduates coming out of school will have some of the fundamentals from the coursework in school, but there’s no substitute for the things that they will learn on the job and in interacting with customers.”
He also noted said that mentorship is significantly important for new engineers, but it’s also a continuous learning process as new technologies are introduced. “While the mentorship piece is critically important, there are also newer technologies and things that are new in the field of electrical engineering and new in the field of our customers’ systems that both new and experienced engineers are learning together.”
The following is an excerpt of our conversation on a couple of key topics.
Richard Quinnell, editor-in-chief of EDN: What should up-and-coming engineers be learning? Or perhaps there are older technologies getting a new significance that they ought to know about?
Keith Ogboenyiya: There are a couple that come to mind. The first is related to automotive. Many of us understand the trend associated with vehicle electrification and the fact that the automotive manufacturers are releasing or planning to release new electric platforms or higher-voltage platforms. But what’s interesting is for automotive systems designers, [these are] some of the first generations that will require high-voltage isolation in the vehicle.
[High-voltage isolation] is a technology that’s been around … but it is the first time that we actually see it being deployed across the automobile. Some platforms are releasing now with a higher-voltage battery. We know that today, the majority of automobiles operate off of a 12-V battery, but there are — over the last couple of years and will continue to be released — new car platforms that will have a 48-V battery.
In addition, when you think of electric vehicle platforms, which go to even higher voltages of 400 V and 800 V and even higher, isolation becomes even more of a critical design care-about for car OEMs and Tier Ones.
It’s not just because of the fact that there are high-voltage systems that need to interact in a safe way, but also for safety reasons with respect to the vehicle operator and the electronics in the systems. Typically, high-voltage components, whether it’s for HEVs [hybrid electric vehicles] or EVs, isolate the chassis for functional safety reasons but also for occupant safety reasons. You need to prevent DC bus voltages and uncontrolled transients that can flow between two points, so in some systems, you need to protect components that are operating on the 12-V side from higher-voltage components that operate on the 48-V or even higher-voltage side.
Along with high-voltage isolation is how it fundamentally changes the vehicle architecture and some of the critical components that are required in order to drive the vehicle.
There are quite a few different systems that now need to be isolated, and car manufacturers and Tier One developers are all in the process of developing this isolation technology and understanding what it means to the system. It’s a new technology for automotive that will be pervasive over the next year to the next couple of years.
Quinnell: That’s one I hadn’t thought of, but it certainly makes perfect sense.
Ogboenyiya: Also related to high-voltage isolation is how the vehicle is propelled. We think about combustion engines moving to electric systems, but that electric system is now driven by an electric motor, and that electric motor will operate also at high voltages of 400 V and 800 V, depending on the size of the vehicle and the amount of load. That also means different methodologies for driving power to that motor. So typically, when we think of motors — and motors are pervasive, whether it be automotive, industrial, or even personal electronics — a motor is driven based off of field-effect transistors — FET-based systems with control loops to support it, so a microcontroller, analog, etc.
When you move into this high-voltage domain, now customers are looking at implementing new technologies — higher-voltage, more efficient technologies — to drive that motor. You’re seeing some customers move to, instead of silicon-based FETS, SiC-based FETs and there’s even discussion about moving to GaN-based systems to drive these motors, so that is a fundamental new technology for automotive that will be, I believe, pervasive across the systems as we look to drive these motors efficiently with the best power density and the best control.
Gina Roos, editor-in-chief of Electronic Products: When you were talking about systems moving from silicon to SiC and even GaN, when do they make that kind of move, and does that require additional attention to other things like thermal management, EMI, and layout?
Ogboenyiya: Most definitely. One simple thing that happens when you move to some of these more advanced power FET technologies is that switching frequencies can go higher, and as switching frequencies go higher, there are some benefits. For instance, passive components may become smaller, which can have a positive effect on power density, so layout may change. But as layout changes, there are also impacts related to high currents close together, which can impact EMI and noise, so noise immunity design has to be thought of. Similarly, as you move to higher voltages and higher currents, thermal dissipation definitely becomes important. So I think that relates to some problems that engineers have seen for many, many years that will still be relevant.
EMI and noise immunity will consistently be an issue for some of the reasons I just highlighted, but in addition, there’s just more electrical content being added to systems, whether it be automotive or industrial. As you add more electronics, you add systems that also can create static or active magnetic fields, so, for instance, electric motors, noise immunity, and how you design for that becomes a definite challenge for systems designers.
Roos: As you integrate more components into a small space, products are shrinking, and there’s a need for higher levels of integration, are there challenges there? What are the biggest challenges for the engineers?
Ogboenyiya: There are definite challenges. Some of the ways semiconductor suppliers have been looking to address that has been through innovative system-on-chip (SoC) integration techniques, but there still are system-level challenges for customers. A good example is radar sensing. I would describe radar sensing as a very innovative technique to address a wide range of applications.
We (TI) have our millimeter-wave (mmWave) radar sensor technology that we use for automotive but also in factory systems such as factory automation, and typically, that had been a very difficult system integration challenge that was done discretely — analog front end, data converters, antenna, processing, clock and timing — all within a tight space and all having to interact with each other, which made for a difficult design challenge. TI’s mmWave technology integrates that technology and makes it easier for customers to develop these types of systems.
But the design challenge starts to change where now customers have to think through, for instance, positioning of their sensor in order to get the best angle measurement, and to get the precise timing in their system, they may have multiple different sensors together in concert with clocking considerations, noise immunity, and how to manage thermal and power dissipation, so those types of challenges still exist in the system.
The other thing that will be a challenge is that, in many instances, these PCBs that customers are developing can be in harsh environments. So, for instance, we talk a lot about smart factories and the additional technology that’s been driven into factory systems. Those can be factories that can be very dusty or smoke-contaminated environments; ambient light can be low. So in these systems, they have to think not only about just the electrical interaction but also how to manage the housing and packaging of the systems in order to ensure robustness over the lifespan of their system.
Quinnell: It sounds like what’s happening is the engineers need to know something about all of the different aspects of design because they interact so much. You mentioned the housing and packaging and the design environment. How does that affect the electronics designers’ actual design decisions? Do they need to be making packaging-aware decisions as they select components and lay out the circuit?
Ogboenyiya: I’ll give an example on the packaging front. Let’s go back to automotive as an example. In the vehicle electrification push moving forward, it’s not just about the traction inverter and high-voltage motor control; you can also see systems that were mechanical, using a belt-based system, moving to electric motors. And those systems can be in locations in the vehicle that can be difficult to access, also with high amounts of vibration and severe temperature constraints.
So the design engineers need to think about not only how to ensure the correct functionality for the system but also about, from a packaging perspective, if it can withstand a certain amount of vibration, which can sometimes cause shearing and package stress, and if it can cause warping within the PCB. There are several different considerations that need to go into the system not only from an electronic functionality but from a design, thermal management, and PCB integrity perspective.
Roos: Are there courses or in-house training that you have in place to help these engineers understand some of these issues like thermal management, EMI, safety functionality, risk assessment, things like that?
Ogboenyiya: Good question. You have to think about the engineering process as a continuum. In many instances, new college graduates coming out of school will have some of the fundamentals from the coursework in school, but there’s no substitute for the things that they will learn on the job and in interacting with customers.
Second, we do have industry experts. For instance, we have experts at TI that have been doing power management design for decades. So you can consider it almost a mentorship process to ensure that new engineers that come in can learn from some of the senior experts and also through structured learning as well. So I wouldn’t consider it training, but it is a structured process by which we ramp our engineers into those specific discipline and focus areas. I just firmly believe there’s no substitute for on-the-job learning. As you work with customer systems and start to develop and release products to market, that learning expertise that you gain from working with real customer problems will ramp engineers appropriately.