Bill Cardoso Discusses Creative Electron’s Inspection Strategy
At PCB West in Santa Clara, California, Dr. Bill Cardoso of Creative Electron held a class on advanced packaging and X-ray inspection strategy. Guest Editor Tim Haag and Publisher Barry Matties met with Bill to further discuss his class and the importance of turning inspection data into information.
Tim Haag: Bill, can you give us just a little bit of background information about your company?
Bill Cardoso: Yes, I'm happy to. Creative Electron is now the largest designer and manufacturer of X-ray machines in the U.S. We design and manufacture everything in our facility in San Marcos, California. And we have been growing very fast for the past few years mostly due to our focus on solving customers’ problems. I know that’s easy to say and everyone says that but we have Ph.D.s on our staff who really take the time to understand our customers’ applications, so we can both recommend off-the-shelf systems and our custom systems if necessary.
Haag: So, what are the primary uses of these systems?
Cardoso: Most of our sales are in the SMT world to inspect for solder joint quality inspection, BGA voids, QFNs, and other types of defects that occur in the SMT process. We also have a wide range of other quality assessment applications from batteries, bulletproof vests, to refurbished cellphones.
Haag: That’s a wide range.
Cardoso: It is. The point of X-ray is to give our customers the ability to see inside their product. From that perspective, it is easy to understand why we have such a wide range of applications. Some of our most successful products today are in material management. Our X-ray parts counter allows our customers to count how many components are on the reel without having to open the package. It allows you to see through the package, and we image the whole reel of components from 7” to 16” reels. In a matter of seconds, it identifies how many components they have without the manual process of reeling and unreeling.
Haag: When you mention X-ray, the first thing that pops into my head is medical. Do you have anything to do with the medical area, or is it all just dealing with components and bulletproof vests, etc.?
Cardoso: Correct. So, if you look at the X-ray market as a whole, there are three major areas. You have medical diagnostics X-ray machines (e.g., hospitals, clinics), safety and security (e.g., airports, border patrols), and industrial. We only operate in industrial. Within industrial, we have a sub-segment, which is the electronics industry. That’s why we’re here at PCB West with companies putting products together and needing X-ray technology—especially for bottom-terminated components like BGAs and QFNs—to check on the quality of the assemblies they make.
Haag: I understand you just taught a class here.
Cardoso: We just finished this morning. That's why my voice is going to be done in about 10 minutes.
Haag: How did the class go, and what kind of students did you have?
Cardoso: The class was on the iPhone X. We like to use high-end devices like the iPhone X as a vehicle to talk about X-ray inspection and advanced packaging, which is one of the main themes for the class. It was well attended and well received. We conveyed to the students that as things get smaller and we keep pushing miniaturization of our devices, the need for inspection is growing as well as the need for high magnification and computed tomography.
Haag: With where you're at now, what's going to be the next big step for your company?
Cardoso: We have an exciting product we released—the TruView Fusion AXI. As an inline, fully automated X-ray inspection system, it's going to revolutionize our industry, and it is AI-based. Our software team has been focusing on AI for the past four to five years, and we've seen AI solve problems that we've never been able to solve before.
We have one example of a customer that refurbishes 100,000 iPhones or smartphones a month, and in the process, you have to open them, replace a battery or component, and close it back before you ship it. Often, a screw is left loose inside the case, the operator forgets to put in a screw, or one of the components is not placed correctly, and that's a problem that standard image processing software can't solve. But you can use X-ray, and if you know exactly where the problem might be, you can design an algorithm to find it. In cases like this where you don't know where the problem might be, that's where AI has made a huge impact in solving very complex problems for customers.
Haag: When many of us have worked in high tech, we’ve been able to bring some of our work home to enjoy. What do you get to bring home? A big X-ray machine?
Cardoso: Well, I did the opposite! I started Creative Electron 10 years ago in my garage. So, my company is just a bigger garage now (laughs). I went from 500 square feet to a 30,000-square-foot manufacturing facility. We were able to design the first machines, sell them, and eventually move to this new facility last year, having outgrown our old facility. And I think there are several important messages in that.
One message is that it doesn't take much to start. I'm a big believer that you should start with what you have. If you really have a good idea, your garage is more than enough to get those first prototypes put together, and then you go from there. Just don't be afraid to get started!
The other important message is that last year, we were in a 10,000-square-foot facility in San Marcos, California. We quickly realized we were going to outgrow that facility by the end of 2017. We had a decision to make. We could have outsourced manufacturing to Asia or somewhere else and stayed in the same facility because the manufacturing would not require as much footprint. Or we had to move to a bigger facility and keep everything in-house. We opted to keep all of the manufacturing in San Marcos, go to a bigger facility, and hire more people. And as recognition, the state of California gave us a $500,000 tax incentive for staying.
Haag: That's excellent.
Cardoso: First, we stayed in California, and then we received the grant. It wasn't the other way around.
Haag: I've worked for a couple of companies that started in garages, and I love to see that kind of innovation and “make it happen” attitude.
Cardoso: That's exactly what we have, and the big challenge as we grow is to keep that startup mentality. I think that the day a company stops thinking of themselves as a startup is the first day they start dying. Because that's when you stop thinking of making as much as you can with as little resources as you can—being scrappy, innovative, and fast to market. You can start to overthink, confuse, and lose.
Haag: What does the remainder of the show hold for you?
Cardoso: We have a couple of other meetings, and I am looking forward to catching up with a lot of the vendors and customers that we have here. From what I can tell so far, it's a very successful show.
Barry Matties: You said you're really inspecting a lot of solder joints. What have you learned from all the inspection? Is there a common problem that keeps reoccurring?
Cardoso: That’s a good question. BGA voids are always there; it's a problem that doesn't go away. Voids in QFNs are also increasingly prevalent because of the way the QFN package has been designed, and the uneven thermal dissipation or distribution outside the package. But we've seen quite a resurgence on through-hole via fill inspection. That's where some of the automotive companies are trying to get away from cleaning that requires less and less flux, making wetting of the solder harder into the barrel fill. So, we've seen a resurgence of customers asking for automated barrel fill inspection.
That's another area where AI is critical for giving those customers pass/fail decisions instead of just giving them an image they have to figure out what to do with. We like to go one step further, and instead of giving the customers data, we like to give them information they can act on.
Matties: So, go backward in that process and improve. When in the selling process, if someone asks you what sort of advice you would give them on their inspection strategies, what would you say?
Cardoso: Inspection is graduating. Inspection for 20–30 years has been a cost center. In this new generation of connected equipment—thanks to Industry 4.0, CFX, and other initiatives—it is graduating into a data center.
Matties: A value add.
Cardoso: Yes. If you look at an SMT line or a process, the only machines giving information or data are the inspection machines. Now, the challenge when designing an inspection strategy is to understand how to turn the data into information. Data you don't use is just noise, and noise only creates a problem in your process. So, think hard about how to design a strategy that can maximize the data you're collecting from inspection instruments, and turn that information into action so you can close the process loop for continuous improvement, efficiency, etc.
Matties: In terms of Industry 4.0 and CFX, how is your equipment integrating into that?
Cardoso: We are in the committee for CFX. At IPC APEX EXPO, we were a part of the demonstration for the CFX both with our parts counter and our inline X-ray inspection machine—the TruView Fusion AXI. But let's be clear; CFX is just one small step towards what we envision as a fully connected factory. CFX gives us that ability, and it's a very important step, but it's that very first step to having machines share messages. The next evolutionary step in this process is to create intelligence that can take that data and connect a void on the X-ray image to a broken stencil on the printer.
Matties: Who's driving that conversation right now?
Cardoso: Right now, that thought is spread out. One of the features of our industry is the fragmentation of ideas. We are putting a lot of the intelligence now in our machine related to X-ray inspection. The future of this company, and for companies like ours, to be successful is to have the ability to deal with the data we generate—to have the data science on top of domain expertise. If you have only one of those two, you don't have anything. If you're a domain expert but don't know what to do with the data, you don't have anything. If you only deal with data but don't have any domain expertise, you're also not very valuable.
So, what we have in our company is data science—AI and machine learning—sitting on top of a lot of domain expertise. We have people with Ph.D.s and scientists who know everything about X-ray inspection. That combination is very powerful. We are driving that from an inspection perspective, and we hope that other companies—the big players in the market from pick and place to printing—will join us and converge in what we think is going to be the future of the factory.
Matties: Is there anything we haven't talked about that you feel we should share with the industry?
Cardoso: It's an exciting time. We've been talking about the lights-out factory for 30 years now, and I think we're getting closer. There are a lot of very interesting conversations related to what jobs will be available in the future, and not only what jobs will be available, but where they will be available. Again, we just the class about the iPhone X, how it was fabricated, and what level of infrastructure was required to make a device like that happen.
Matties: But the history of manufacturing going back to Henry Ford. Where did all of those jobs go when he created the assembly line? The new economy will have to create new opportunities for employment. It doesn't mean we shouldn't have manufacturing that's automated because we're afraid.
Cardoso: Absolutely. I think that's a healthy way to approach the topic. If you approach it from a protectionist perspective, you think, "We need to protect jobs and delay automation because we want to make sure that these classes are protected." It's not as productive because it's going to happen anyway, and in this case, it's better to be a leader than a follower. Followers tend to get a much smaller share of the market.
Matties: Thanks very much for that, Bill.
Haag: Thank you.
Cardoso: Thank you so much for your time.