Issue link: https://iconnect007.uberflip.com/i/1424540
NOVEMBER 2021 I SMT007 MAGAZINE 79 computational power of a smart thermostat— that would have been a nice desktop computer 10 or 15 years ago. It's fully connected to the in- ternet; it's the transformation of interconnect- ed communication, especially wireless devices, into a much larger infrastructure typically con- trolled by soware, or it is used downstream for analytics. e electronics that go into a ve- hicle are anything but simple, especially for au- tonomous driving. Some of the most serious- ly complex computational horsepower avail- able today goes into an autonomous vehicle. I'm not saying those are super high volume yet, but the amount of computational power elec- tronics, the number of sensors, the complexity of sensors, the complexity of assemblies, all go into what we take for granted. Johnson: In developing this product, overcom- ing those challenges, what came out of the R&D? What became the primary features or attributes of this equipment? Cain: One of the things decided early on was automation, which is two-fold and plays into the question that you asked me earlier about platforms, factory automation, and data inter- play. ere is physical automation; and data automation, which leads to data analytics in- sights, especially if you get to predictive and prescriptive insights, because those feed back into a method to make sure that what you mea- sure is helping you improve. One of my favorite quotes is from Lord Kel- vin, who said, in essence, "If you cannot mea- sure it, you cannot improve it." ere is a lot of work that goes into the scientific process, analytics, and knowing root cause; then you act and continuously improve. We established a platform that could enable customers to do that, starting with physical automation, while thoroughly putting the hooks in for data auto- mation to make sure that customers could use the information that is gathered, to gain in- sights and act quickly to improve yields, im- prove their time to market, and to be first in market. ose are things these high-volume electronics customers hold very dear. Johnson: It seems to be at a critical juncture. We have Industry 4.0 infrastructure and data pro- tocols that capture data, add a process, and can communicate that from machine to machine up and down your line or elsewhere in your fa- cility. at information is available, but only as raw data. Now we need to process that data to turn it into actionable, meaningful control and guidance, whether that means fine-tuning process windows automatically between ma- chines, or providing management-usable in- formation back to the operators and supervi- sors. I'm sure there is some interesting R&D going on there. Cain: Yes. ere are two primary forms to physically automate. One is to do a convey- or system. at's well-known and used in the industry. e other is to use islands of robot- ics to load and unload printed circuit assem- blies as they get manufactured. We do both. is particular system uses a guided rail for au- tomation and there are things like IPC CFX, Hermes, and other standards we fully comply with. ose standards then make these systems plug and play into the larger context. ese are not islands unto themselves; they're part of a whole chain of equipment that other vendors create, and at the end, that customers create themselves to essentially manufacture and ver- ify that what they're manufacturing is indeed ready to send to customers. We've also been investing in an analytics plat- form for manufacturing called PathWave Man- ufacturing Analytics, which has been deeply integrated into this massively parallel test plat- form and other manufacturing test platforms that we make. It is also built to be integrated with any test platform—anything that produc- es data. Industry standards continue to evolve, so we aren't as concerned about the format. We will take it all. If we don't have a way to convert