Issue link: https://iconnect007.uberflip.com/i/1372612
34 PCB007 MAGAZINE I MAY 2021 driving this with us because it was a brain- storm I had with them when we first started the project: "Why don't you send us some of your board business? We will have your peo- ple come here. We'll all build it together, and if you have any concerns with how we're build- ing it, as we're designing your factory, we can make changes. We can modify things; we can do it the way you like." It's really a big coop- erative effort between ourselves, SEL, and Vi- cor to go ahead and put this training and pro- gram in place for all three companies. en, as we add more factories to our little cooperative, this is going to get easier to do. We will have a program in place and will to continue to im- prove it by adding levels and training. Johnson: It makes sense to be doing that sort of training for the human operators. Absolutely. At the same time, how much of this gap at your client, the OEM, do you anticipate filling with the AI technology? Stepinski: We have some initiatives relative to AI, specifically addressing indirect labor in PCB shops where we have skill gaps. One of the biggest skill gaps is in the plating depart- ment, copper plating. ere is a lot of lost art there, that's for sure. ere are very few peo- ple anymore who really have a clue as to how it works. First, the equipment set is quite old on average, so you're dealing with archaic equip- ment, like the "Millennium Falcon of the facto- ry"; without Han Solo it doesn't run that well. ere are not too many Han Solos out there for this area. As part of these factory projects, for instance, we're making a whole system that mounts on the front of a plating line and we are building plating lines now as well. e system on the front of the plating line has metrology built in. So, you do partial plating, measure before and aer, and weigh and image things; we have a whole sequence of metrology piec- es that are in this. en you can say, "Instead of going 100% of the way in one shot and having a good probability of screwing it up, you do a partial plate, analyze the results, extrapolate it out, make a correction, and it's a simple algo- rithm." en this is automated. at's what people do now: ey put something in, pull it out, cross-section it, and they make adjustments— or at least that's what the sophisticated folks do. e less sophisticated folks just pin gauge it or use a CMI probe, but then you're leaving a lot of things up to chance. We are really focused on developing metrology that can be used; I think Rob- ert mentioned gravimet- ric. is is a big initiative from us using gravimet- ric techniques for indirect measurement. e weight change should always be the same. Nobody is us- ing weight to monitor the process. Why would you use all these sophisticated tools when you can just do a weight loss? It is very sim- ple, and we are trying to keep things simple. at's a tool that you can use across all processes.