Issue link: https://iconnect007.uberflip.com/i/1213413
44 PCB007 MAGAZINE I FEBRUARY 2020 SMT machines have interfaces with plenty of ways of getting data; we integrate with those. There are also machines on the mechanical and assembly sides that are older and more legacy; maybe they're just purely mechani- cal machines like presses. We also have some hardware components for those particular machines to get data out of them. We can in- terface with PLCs directly with industrial pro- tocols. Whatever is required to get data out of any machine, whether it's modern or legacy, we're never blocked. Johnson: Now, you can get all the data from all the machines, regardless. You now have this heterogeneous manufacturing floor all wired up and your data collection correct. You can do something with that. Burke: There are a couple of things we find our customers really want to do with it. One is they want to be able to provide it on-demand to their own internal applications in a way that doesn't require them to duplicate the work of the direct machine connection every time. We call this our data broker. No matter where the data came from—if you are a home-built MES, a local site application for dashboarding, or a traceability application that is home built for a customer—rather than each one of those individual products going back to the same machine, asking for the same data from the same vendor, and getting the same support, slightly different but basically the same, they get it from the broker. It's one connection with one system, and it feeds the data to the various people. That's one thing we help our custom- ers do. Johnson: That means that if you have a het- erogeneous software environment—using dif- ferent software tools in different sites—by working with the broker, you not only have a unified place to bring the data together on the hardware sensor side but also on the analysis software side. This starts to make a lot of sense. One major challenge that seems to show up for factory automation is having the data end up in silos. You have been talking about getting past that. Burke: Let me give you some very specific ex- amples of what we mean by that. Let's take, for example, MES data. MES data is very fo- cused on products, so that's very critical for a factory. What happened to my product? It went through machine A, then machine B, and then machine C. If that's the only way you store your data, it's very easy for you to track across a single product what conditions it saw, which is critical for your factory. What's hard in that case is seeing what hap- pened for a given machine. What are all the products it saw? And if you only look product by product, you don't see trends like the ma- chine is slowly getting out of alignment, slowly getting out of process control. By putting to- gether the data in a vertical rather than hori- zontal way or vice versa, suddenly, you can develop different insights from the same infor- mation. Imagine doing that with MES and specific data augmented from the machine as well as ERP data and other data from SPC and pro- cess control databases at the customer. It's im- pactful, putting all that together and starting to see what new insights I can draw that weren't present in any given system because it wasn't a job of any given system, but it was a new thing for Industry 4.0, looking across these big data sets with modern tools that excel in this large data environment. Johnson: Who are your customers at this time? We can interface with PLCs directly with industrial protocols. Whatever is required to get data out of any machine, whether it's modern or legacy, we're never blocked.