Jennie Hwang: Get Ready for Disruptive Technologies
At IPC APEX EXPO 2019, I met with Dr. Jennie S. Hwang, a columnist, author, and all-around an expert in PCB assembly. We discussed some of the changes she has seen since joining the industry, and disruptive technologies that technologists we are going to have to face in the near future. Are you ready for the future?
Nolan Johnson: You have been one of our columnists for a long time with SMT007 Magazine. You were saying before that you might be our most tenured columnist.
Dr. Jennie Hwang: Yes, I started at age six (laughs). I’ve been in this industry for nearly four decades. I have observed and participated in its most dynamic, challenging, and rewarding times, and I am continuing to do so. I would be glad to share some of my retrospective and prospective thoughts about our industry. Surface-mount technology (SMT) has served as the backbone of electronic products and enabled the printed circuit board (PCB) manufacturing industry to thrive since PCBs are the brains of electronic products. SMT was implemented in the early ‘80s, give or take a couple of years, which made personal computers possible, sitting on the desktop, instead of occupying a whole room.
For the last three or four decades, the industry has had periods of abundant breakthroughs or drastic changes, as well as the periods of incremental changes and gradual advancements throughout the industry hierarchy in materials, design, equipment, processes, and manufacturing infrastructure. The 1987 Montreal Protocol on eliminating the ozone-depleting CFCs led to the development of no-clean chemistry for soldering, solder paste, and overall CFC-free cleaning. Also, the 2001 lead-free legislation initiated by EU ROHS that implemented lead-free electronics is another example of an event that led to major industry changes.
Concerning SMT, we experienced continued changes in packaging and component design, from PLCC and SOIC components to QFPs, fine-pitch QFPs, BGAs, fine-pitch BGAs, small form factor bottom-termination components, PoP, and other 2.5D and 3D packages. Thus, the evolution of new package designs affects solder paste formulation, the printer and stencil design, the equipment, and all of the accessories and supporting services, which involved a large number of suppliers. Accordingly, I have seen countless refinements and improvements and worked hard on them to serve the industry.
One of my professional development courses or industry lectures is under the title, “Emerging Technologies of the Electronic Industry.” In this lecture, one of the viewgraphs that I have often used to illustrate the industry from the manufacturing perspective is the hierarchy of the electronic industry, which includes semiconductors, IC packages and passive components, the module level, the PCB level, and the system level. SMT essentially focuses on the packages and components, the module level, and the PCB level, and to a lesser degree, the system level.
Practically speaking, the SMT industry is driven by the semiconductor sector. However, those phenomenal semiconductor chips would have no use if they couldn’t be reliably connected to the real world to deliver the intended performance. That’s where the critical role of SMT comes into play—to connect semiconductor chips to the real world to deliver performance through a PCB.
Johnson: I come from the IC design tools side. I wrote code as an engineer for for a major EDA company for a number of years, specializing in IC design. I think you’re exactly right. You need to have that conversation. “What can you do on the wafer? What can you do on the silicon? What package does that require to get it out in the real world, and how does that interface with the board?” You were just touching a little while ago on new product development. Where do you see us right now?
Hwang: With the incessant introduction of new packages and components, surface mount manufacturing has been relentlessly improved and refined. Today, we are moving into advanced manufacturing using IoT and AI to truly move into what I call intelligence-teaming manufacturing.
Johnson: We’re starting to talk about the dimensions now on the PCB that were the dimensions we designed to with integrated circuits when I was working in that industry. We have dramatically reduced the dimensions, and that changes what you need to do as far as fabrication on the boards; assembly drastically changes.
Hwang: Yes, the miniaturized dimensions drive all sectors of the industry—materials suppliers, PCB fabricators, SMT production operations, equipment manufacturers, and others.
Johnson: There’s a lot of attention at IPC APEX EXPO this year on CFX and Hermes, being able to pass data, having machine-to-machine interoperability. Where does that fit in your perspective? Is that disruptive or a refinement?
Hwang: The future factory is driven by the nascent Industry 4.0 and its ultimate goal to achieve intelligence-teaming manufacturing or operations in an integrated manner. In manufacturing—such as surface mount—reliability, agility, flexibility, efficiency, and desirable costs are the name of the game. However, becoming truly automated is a daunting task.
Johnson: That’s a provocative comment. Let’s go into that.
Hwang: We always say, “Surface mount is an advanced manufacturing technology,” and that is true. In many ways, it enables us to produce the electronics we have today. But true automation in the realm of Industry 4.0 must deliver all of the features I previously mentioned: high reliability, agility, flexibility, efficiency, and low cost. Over the years, surface mount equipment and processes—such as the printing, inspection, reflow—have advanced tremendously, which have possessed some level of automation. But if you look at surface mount as a whole operating system, it’s modulized and not integrated. We need to have advanced manufacturing, which means intelligence-teaming automation.
One caveat in automation is that it can produce a substandard product, which is certainly not an outcome that we want to have. Reliability, particularly for electronics, is critical. Even for consumer products, such as smartphones that are not mission critical, we still don’t want them to malfunction or “die” on us (except the battery).
Johnson: To pick up your point, I was amazed at the automotive executive forum. Someone commented that the field failure rate we have in phones is very good, but for automotive, they need to be something like two orders of magnitude lower in field failure rate. That’s the reliability that the industry needs to strive for. So, if someone is thinking that it’s challenging to get to cellphone requirements, that’s nothing compared to automotive.
Hwang: Right. In automotive, safety is the top priority, which is another level of reliability. We’re assessing from every functionality. For example, looking at the solder joint, I just taught a course yesterday titled, “Preventing Production Defects and Product Failure.” Interestingly, a lot of the attendees came from the automotive industry, military, and high-reliability electronics manufacturers. Obviously, they are concerned about potential product failure. So, we keenly care about the premise behind reliability. How you define reliability and look at it? Each industry has a different level of reliability they’re seeking or required to deliver. Not all products and industry sectors need the same level of reliability.
In my lecture, I always remind the audience that reliability is relative to your product service environment. Aerospace is different from the design of a dishwasher. Further, reliability is correlated to cost. How to optimize reliability to balance the performance features, manufacturability, cost, and time to market is a challenging question to address. After all, we produce electronics products where the business goal is often focused on introducing them to the market in a timely manner. Therefore, we must balance those factors. For a car, we want the electronics to perform whether it’s for engine control or the glove compartment. So, how do we produce that? It’s about making tradeoffs and reaching a balance with manufacturability, reliability and time to market. We can say, “We will design a most reliable product,” but if it will take another 15 years, that may make the product not viable anymore.
Johnson: What do you see as being some of the disruptive factors in the next five years?
Hwang: Our industry, like other industries, is and will be affected by the megatrends, namely, AI, IoT, and 5G. For example, IoT is going to drive product design and operation. How this industry is going to embrace these changes is critical to future success. In another one of my professional development classes, we had people coming from companies specifically in the area of IoT because they don’t want to see failure. IoT and other megatrends will greatly affect how the electronics are going to be built. What reliability will we have to deliver? Then, to deliver reliability, how does that affect material innovation? And how are they going to affect equipment changes?
It also depends on how you define AI. If we look at the ultimate AI, I think that will happen sometime down the road, meaning things will become AI integrated with human intelligence to operate everything in a harmonic, reliable manner. But if you isolate AI to certain functions, and I think it’s already here and some industries are doing that already. AI will link to certain electronic hardware as well. So, that type of AI will affect the electronic design, semiconductors, and package; it would not be standard for semiconductor chips. To that end, you want a semiconductor design to perform in a certain segment of AI. That is going to pass down to this industry, and the industry will have to consider how to deliver and embrace that to best serve the market.
Johnson: AI is important to pay attention to because the specific semiconductor design requirements to make chips that process an AI function at high speed are going to change the demands around packaging, which then will trickle down into changing the demands on how you place those components on circuit boards.
Hwang: AI is such a common term today. Once, when I was in a taxi, the driver started talking to me about AI; it varies with your perspective. In its full deployment, AI will bring huge changes. There are a lot of the individual levels of AI, and over the next five years, we’re going to see more AI and a higher level of AI.
Johnson: So, people in our industry should pay attention to IoT and AI regarding manufacturing demands.
Hwang: We really need to understand what Industry 4.0 is, what it requires, and what it takes to get to Industry 4.0. I think that is going to take the next several years to develop.
Johnson: A burning question around the CFX working groups was this: Now that we have a standard and a way of interchanging data, how do we make it useful?
Hwang: Absolutely. Data is data. Although data is valuable, I am not playing down its importance. Everything is about data in our data world, but it’s more about how to separate the noisy data from the clean data, and there’s a lot of messy data to clean up. In the next five years, we are going to learn a lot about that by creating something like CFX, which is just a starting point. We have to start somewhere to get to our destination. To achieve the desirable outcome, it’s not just about collecting a lot of data and exchanging it; there’s much more to it.
Johnson: If you can’t get something useful that helps you and doesn’t add value, then it’s just noise.
Hwang: Yes, usefulness is what counts. In terms of disruptive technologies, the real disruption that is going to come to the electronics industry is when we have an entirely new array of materials to make electronics. We all know that silicon is very hard to replace, as vividly demonstrated in last the three to four decades.
Like my question to JB Strabuel from Tesla’s keynote speaker this morning regarding the electric vehicle: “What new materials are needed to deliver high-energy and power density beyond what conventional lithium can provide to reach the speed, range, cost, and other performance features of a well-rounded electric vehicle?” There is still a distance to go. There are certain limitations on the range and storage we can achieve. So, the availability of higher power density, storage, and associated technologies is necessary.
For electronics, to get another level of performance, we will need new 2D materials. This area has been rigorously researched for the last few years across the globe. Like AI and 5G, this is a global race and competition. Surely, 2D materials are going to revolutionize electronic devices, but when that will happen remains to be determined. That would be a huge disruption for electronics—an entirely different type of materials above and beyond the typical semiconductors we’ve been making.
Johnson: This has been great, Jennie. Thank you.
Hwang: Thank you, Nolan.