How to Use AI Technology & Factory-Wide Traceability to Optimize the Production Process in Electronics Manufacturing

E. Miscoll: Hello and welcome. I’m Eric Miscoll with EMS Now and welcome to this edition of EMS Now Up Close. Today, I am visiting with the guys from Cogiscan. We have Mitch who is the Director of Business Development at Cogiscan and Greg who is the Director of Product Marketing. I wanted to do a catch-up as we spoke earlier this year with these gentlemen. Actually, it’s been just over a year since Cogiscan was acquired by iTAC, so I thought it would be good to kind of get an update about how that's going and how Cogiscan is doing in the market what, almost a year and a half on now. So, gentlemen, welcome. Mitch let me start with you. The blending of two companies and acquisition, there's always the culture question so you know: how has the cultures of the two companies mesh together?

M. DeCaire: Hi and thanks for having us. Yeah, it's a great question. The cultures of the two companies were familiar to one another beforehand because we've been working with iTAC as partners for over a decade & now serving some very large installations globally, our systems have been seamlessly integrated via those prior partnerships. So, one nice thing to start from was the fact that we knew each other very well already and had already collaborated & cooperated. And we already had an integration between our systems so that has been great. One other nice element is that both companies have their roots in the electronics manufacturing industry. In fact, we know people at Cogiscan and people at iTAC (management teams, etc.) have known each other for years even before I worked for Cogiscan for example. I already knew some of the people that are now key players at iTAC so it's kind of like an extended family that became closer through this merger.

E. Miscoll: Good! That's good to hear! Well, let me ask that: Beyond the cultures there's the blending of the two companies’ solutions right. Between iTAC and Cogiscan. How has that been received in the industry?

M. DeCaire: Well, very well so far! You know, it's exciting for me since I’m on the business development side of the business so I have the privilege of being able to introduce iTAC to a lot of companies, specifically in the Americas market. And of course, iTAC is a very well-known name so no one has to be told who they are, but when you get to meet with people to present the products, it opens the doors to new opportunities that we probably wouldn't have been able to pursue when Cogiscan was autonomous and by ourselves. So that's very exciting. This includes new opportunities with existing Cogiscan customers who can benefit from the expanded portfolio that we talked about during our last interview together. And that's great because now we can introduce new and exciting things, and given the track record we have of having our systems integrated it's really a baby step for existing Cogiscant customers to take advantage of what iTAC can offer well beyond what we could do before. And then, of course, for new customer projects, we can now go after enterprise level MES-type projects, which, of course, that's where ITAC is a world leader. And also advanced analytics, through either business intelligence or more advanced artificial intelligence offerings that we can combine with our machine connectivity which has always been Cogiscan core product to enable smarter factory projects. Something that's very exciting about the business intelligence and artificial intelligence side of the iTAC portfolio is that because of the modular and scalable nature of their architecture, you don't actually have to even have an iTAC MES to use these other new products. That's something that I even didn't really foresee at the start but as I started introducing the products to people and I get this interest, even from companies who have another MES or their own internal MES, they're interested in some of the other modules that we can provide through iTAC that can integrate with other systems.

E. Miscoll: That's been very exciting and that's an important point. I remember that from earlier in the year, the whole modularity of it. You don't have to buy this whole big solution you can kind of pick and choose where and how you want to engage and as you said, you might bring in the cognition solution with a different MES system in there, but that would still open up opportunities I would imagine.

M. DeCaire: Yeah absolutely! I mean a simple example would be that with the iTAC business intelligence and AI offerings you could take an existing Cogiscan installation where we're connected to the machines - we're already collecting the data - we can feed the data to those systems to enable what they can do. The business intelligence product for example can perform statistical process control. So by feeding test data or process data to this system, we can now automate the sending of messages and alerts to the factory personnel, not when something's going wrong, but even before something's going wrong. So you know we're always talking about more advanced technologies such as that that we’re going from a sort of reactive model, or illustrative model, to a predictive model in terms of “Hey Mr. customer you need to go to line three now because in another hour from now this could exceed some sort of threshold.” And then on top of that we have the artificial intelligence offerings, which take it to a whole other level where we're looking you know we're analyzing trends and looking for outliers and bringing alerts to people for that. And some of this stuff, especially when you start getting into the artificial intelligence you know it sounds sort of space age, but we actually have practical real SMT use cases that are already in use today. So it's kind of like we're taking what I think some regard is sort of a futuristic topic and putting it together in a very practical pragmatic and usable way.

E. Miscoll: Yeah! And Greg let me pull you in on that, on the AI side because that's an area I know that you're very involved with. So, talk to some of the AI technology that's embedded and how customers are utilizing it.

G. Benoit: Yeah! So one of the products in the package that Mitch has been talking about we call the IIoT.Edge. So, this is basically an AI platform that you can use for predictive analytics. It's able to take data from a bunch of different sources as inputs like MES data, data from any sensors the customer might want to use for things like temperature vibration, and then as Mitch said we've integrated our Co-NECT platform in it to deliver all of the machine data. So, it can take all these inputs and run ai algorithms live on this data stream and look for patterns and predict certain key events. To give some of the practical examples that Mitch was talking about, the use cases that we've actually uh deployed for this kind of thing. One of them is for predictive maintenance for machines. Monitoring the machine data coming in from Co-NECT, the machine statuses, alarms and warnings and also taking the data from any of the sensors that the customer might want to use for things like temperature vibration. It then analyzes this kind of complex data set using it to predict when the next machine breakdown would occur if nothing were to be done. And then telling the maintenance staff when they need to perform maintenance in advance in order to avoid that downtime. So this, of course, improves utilization OEE and things like that and it's not just kind of waiting for something to break and then fix it or using some kind of like preventive maintenance schedule. Another exciting application we have is to deal with the problem of false calls on AOI machines and so. As we know, this is always a major problem with inspection of the AOI reports some defects that are not actually defects. And the result is you have to have a verification station after the inspection machine with an operator who looks over the defects reported by the AOI to verify whether they're actually true defects or false calls. So what the algorithm does here is it can use the AI to pre-screen those defects and identify which of them are actually false calls and then filter those out so they're not even reported. So this results in a lot fewer defects that have to be manually verified by an operator which can really reduce your labour costs. This is just a couple of the examples of what we've been doing with the platform. We have some other use cases and applications related to things like optimizing cycle time, analyzing quality data to help identify problem areas and speed up root cause I won't go into all that right now but you can expect to hear a lot more about this from us coming soon.

M. DeCaire: Yeah! One other kind of example that we're working on currently with a large defence contractor in the USA. Just a simple example but taking that what Greg just talked about. Connecting it to sensors inside the screen printers, for example. So we're already tracking the soldered paste in the screen printer, we're already tracking what allegedly the floor life should be of the solder paste in the screen printer. And if you think about what the floor life of a solder paste is, I mean what it’s most dependent on? It's the humidity. So by just simply connecting a sensor inside the machine and feeding the humidity to this platform, now we can make adaptations to floor life based on real-life conditions, not just what's on the jar of paste in terms of the floor life. That’s just a great little simple example where we can put this on the factory floor now.

E. Miscoll: Now hey Greg, the new inspection product and make sure I’m not confused, is that what you refer to as the Focus AI?

G. Benoit: It is that yeah. So the one I was talking about is a separate AI platform that can be used in general, but then we do have another AI-based product, also related to inspection. In this case that's what we call the Focus AI. It is a hardware-based station with a 2D camera that we've developed with our partner Fitech in Poland. What this is really ideal for is for kind of putting optical inspection in places where you normally wouldn't have it. So, to catch things like basic defects: if a part is missing, or if it's been placed with the wrong polarity, or if there's something wrong in terms of the colour of a LED (those kinds of sort of simple defects). We found a particularly useful for use after through-hole placement and before solder (so before the parts are actually soldered to the board) to do a kind of a quick optical inspection. It automatically catches some of these basic defects before the board is soldered and you have to do a repair. It could also be used after final assembly, for inspecting like surface features of the product (so making sure that screws are all in place or there's no scratches on the product). The really nice thing about this product is that since it uses AI there's very little programming involved to set up a new product. You can basically just feed the system a few examples of kind of a golden product (basically feeded examples of how the product should look) and then very quickly you can start running inspections on that product to catch these defects. So this is another AI-based product that we're really excited about too.

E. Miscoll: Well let me ask you: I like that you've used some case studies which I appreciate, but how are customers utilizing the combination of the two solutions and what are the features that you see them kind of leveraging more? What are they really deriving from that? What are you seeing?

M. DeCaire: I mean the first thing that comes to mind is factory-wide traceability. You know we've been doing traceability since our inception as has iTAC, but if you look at the typical electronic manufacturer, a lot of them do a really good job on the conveyorized portions of their operation (the SMT line whether you have a Cogiscan system or maybe you bought a traceability solution from your pick and place vendor). You know it's relatively straightforward and the traceability solutions available through us or through other companies are very robust for the SMT line. But, once you go past SMT and your different batches or work orders of partially built boards go off to the back-end as we call it, it's a lot harder to track and control interlock operations, assure quality and ensure traceability data integrity. This is where iTAC is really a leader in the industry. So that’s something that comes out quite often as the number one driver of discussions with new customers. So, by being able to take that traceability to more of an enterprise level, included with a deeper integration with the customer's ERP, and that kind of lead to another very current topic which is materials management. Again, managing materials on an SMT line is relatively straightforward because all the machines are connected and there's only one place for the materials to go. Once you're in the back-end operation, you have different workstations doing different things and materials being taken manually from location to location. It's a lot easier to lose track of things and this is again where iTAC's deep presence in back-end and box-build has really helped us and is leading us to new opportunities. Combine that with electronic kanbans for making sure you have the right material at the right place at the right time. Combine that with controlling robotic delivery systems like AGVs and we can take it to a whole new level of automation using the tools we have available now.

E. Miscoll: Greg, did you want to add to that?

G. Benoit: No. I mean other than maybe in full disclosure I used to work for iTAC actually for a number of years. I joke with the management that if they really wanted me back this badly there were cheaper ways to do it than buying Cogiscan. You know I was “super stoked” when I found out that we were being acquired because, to Mitch's point about the culture, these are people who, I mean, I knew a lot of them from working there but even before that. We have kind of a shared history. We both come from electronics so it's a real sort of natural fit. But then also from the product standpoint, they have a broader offering in one direction, but we go kind of deeper in another level in terms of really integrating directly into the shop floor and the equipment. So I really saw this as really a great natural fit for the solutions to come together. I’m really happy with it.

E. Miscoll: Good! Well, as long as you're happy Greg. One of the things I did want to follow up when I was looking at this, because we're talking about the lines and everything, but doesn’t Cogiscan also has kind of the post reflow kind of piece too? Don't you have a scanning opportunity beyond that as well?

G. Benoit: That kind of what Mitch was referring to. The back-end part right. So Cogiscan is traditionally focused a lot on the SMT part and we have a lot of great coverage for that, whereas iTAC has expanded a bit more broadly towards the full process. So now when you combine the two we have a great solution for covering all the things you need for SMT and iTAC comes with a lot of the things for the back-end assembly, the manual assembly, PLC connections, shipping, those kinds of things. So we really have a full solution for electronics manufacturers now.

E. Miscoll: Yeah and there's a big movement in the industry to try to automate more of that anyway. So it seems like you're kind of positioned well then for as that ultimately happens right. You know chasing this lights out kind of factory type thing right.

G. Benoit: All with one platform essentially, you know, so it's really exciting.

E. Miscoll: Good, good. Listen, last thing I wanted to ask before we go: What's coming in the next few months? What's on the docket? Is there anything you guys can talk about? Is there anything coming up?

M. DeCaire: Well one thing that comes to mind that is that July 13th 2022 we're having a webinar. This webinar will be educational, practical, it will provide real-life examples of how our customers can improve their operations with this expanded portfolio. There will be quite a bit of emphasis on the new products that we have available through the iTAC.MOM.Suite. So that's happening on July 13th and that's something we hope many of our customers and prospective customers will want to join.

G. Benoit: We'll be sending out some promotions to our list, but we'll also be promoting it on social media and so forth. It'll be visible. People will be able to find the invitation in one of those places.