Issue link: https://iconnect007.uberflip.com/i/1372612
MAY 2021 I PCB007 MAGAZINE 29 Stepinski: An engineer should have this as a subset of their skills already. Matties: ey just need to make that part of their daily function. What are you doing with this data and why? Zajac: is is a soware platform into which we can enroll all our equipment and other equip- ment as well; anything throughout the facto- ry can be connected to this as a single pool of information, a "unified namespace." Consider, your ERP has some data, your MES has some data, your WMS has some data, and you need all this available for you. Part of the system we are building is obviously the MES that will gov- ern the operation, recipe management, and all these activities, but to have this working in tan- dem with other systems to make the decisions automatically, which would have to be available. First, we must build the ecosystem for it, so we are currently defining a system that allows for that. Among the features we want to provide are some very advanced control schemes that would otherwise be impossible to do on stand- alone equipment pieces. Some of those would be physiologic, statistical process control, AA recipe management, predictive maintenance, and dynamic recipe corrections. ese would be aided by machine learning, and in some cases, could affect some cloud-based data col- lections for remote troubleshooting in case of maintenance issues. We are trying to put a sin- gle blanket over all our product lines with this soware and make this a mesh that ties all the pieces together. Matties: I know this is much broader than North America but if we look at the North American market, a lot of these shops are us- ing equipment that is still 10- or 20-plus years old. Will your system retrofit, or will they need to invest in new lines? Zajac: ere was no necessity for direct re- placements. We consider that some of the equipment might be dated. However, there is still straightforward potential. We provide ad- ditional edge devices that allow for intercon- nection of legacy equipment. Matties: at makes it a lot easier for them to move in that direction. And in some cases, bringing in new equipment may make more sense. It's just going to depend on their circum- stances. Zajac: We know that buying equipment direct- ly off the shelf, out of the box, ready to connect will be the most favorable option. But, if you weigh in some of the equipment prices, obvi- ously there will be some necessity for retrofit- ting. Matties: When you start talking about dynam- ic recipe correction, that's where you're likely going to see the newer equipment. Zajac: Right. at would rely on collection of specific recipe parameters. is is hinted at in the wet process in vertical lineup, so you're go- ing to be able to collect all your parameters that you set up in your recipes and see the out- come. Based on this information stacked his- torically, the AI or machine-learning algorithm will be able to decipher the impact of all these parameters through your end result and cor- rect accordingly. Matties: As we've talked about with Green- Source, this puts a heavy demand on your front end. You must set up the files to be recipe driv- en, because you're not going to be making manual corrections during the process; it's go- ing to be planned out upfront. How challenging is that? We hear that information from design- ers never comes in complete, that there is a lot of front-end work, a lot of back-and-forth. How do you help a fabricator get over that hurdle? Zajac: ere is still a necessity for your process engineers to have first say in your recipes. For