Issue link: https://iconnect007.uberflip.com/i/1267313
78 DESIGN007 MAGAZINE I JULY 2020 out, and if you put too little, it may not be enough. That had to be a trial and error that required some experiments to perfect the pro- cess. When you start dropping the components with the SMT machine, the third phase is we may need to fine-tune the robotic head to make the precision be that much better. There is an opportunity to fine-tune the robotic heads and hardware, as well as software, to make this process a lot more precise so that when it's dropped into these nests, once the IC goes in, it will perfectly sit inside. Those are a couple of areas that immediately come to my mind: knowing what the processes look like today, not altering today's processes, and then how do you fit into those processes? Those may be some of the areas that we have to experiment and work with, and this is where we need a partner who has the resources to make this thing a reality. Matties: Oftentimes, we see new technologies come along—and there are some great ideas—but without the proper funding or distribution, they fail. There's a big risk in the mind of the user to even invest in experimenting with this new tech- nology. Not yours, per se, but we've seen the his- tory time and time again. It repeats itself. Yaung: Every new technology or invention has its own risk. We have seen a lot of new tech- nologies go through the process, spend billions of dollars, and never even go anywhere. In our case, the first barrier that we already removed is not to alter today's manufacturing processes. That's already a plus. Now, how do you work with the processes to make this thing a reality? A lot of it is trial and error. And to do that, you need the time and resources. Matties: How many shops are currently using this? Yaung: We have had five shops, and we pur- posely experimented in different places: one in Korea, one in the United States—because we invented it in America—two in China, and one in Taiwan, but Taiwan was short-lived. We don't produce these things in Korea anymore because the cost is high. The two shops in China are making them. Now, with the pan- demic, we are thinking about bringing it back to the U.S. I'm having discussions with a cou- ple of shops. They're very intrigued. They look at this and say, "Wow, but let us think about it. How are we going to do this thing?" Johnson: To be clear, this is in production in those shops. Yaung: Yes. They are volume production shops. For one reason or another, these two shops in China are able to do this thing very well. They signed the NDA with us, we tell them exactly what our criteria are, and we are not altering their processes. They go out and do it. As long as they meet our specifications, we'll be fine, and they came back with it. Matties: You said the shop in Taiwan stopped using it. Why did they stop? Yaung: We stopped using them because they were having trouble making the height of the solder mask even close. We were very for- giving in terms of meeting exact spec, even though we have a specification call-out for it, our tolerances were quite generous, but they were still having a difficult time. They tried it for two or three lots, and there was a 30–40% yield loss. Greenberg: Right now, we're talking about our current product line, which is the prototyping product line. I want to make sure that's clear. We are not talking about the mass production product at this point. Yaung: If you look inside of our current prod- uct, you'll see canals. Where the IC leg sits, the trace is pretty long out there, and the purpose for that was for manual hand soldering with a soldering iron. You push the solder in the canal toward the legs to adhere to it, but the volume production one that we are talking about, or that we are taking it to, is not bad. It's just a pocket. The pocket is as long as the chip leg will sit.