Omron Discusses SPI, AI and More

Estimated reading time: 4 minutes

At the recent SMTA Oregon Tech Forum and Expo, Tim Anderson, a senior account manager at Omron, shares his insights on SPI, the latest trends in the industry, and some solutions to consider. As he points out, the SPI, AOI and AXI systems of today do far more than simply catch defects.

Barry Matties: Now, when we talk specifically about SPI, what are the advancements? Are we seeing machine learning and that sort of feedback?

Tim Anderson: The biggest thing that we're seeing right now is in the Industry 4.0 initiative and getting that information back to the other systems that can utilize it to improve the process. Essentially, it’s to get that feedback information to the printer so you can prevent further defects from coming down the line. 

Matties: It's making real-time changes. 

Anderson: Exactly. After every board, it sends that information back. The printer can then react and adjust accordingly.

Matties: People are confusing AI with machine learning or that sort of feedback. Where would you put that in the scope of AI?

Anderson: Well, AI is definitely an important piece of all the SPI/AOI/AXI app products. It has made its way into our programming interface—looking at trends and things of that nature and feeding that information back into the software. You’re not relying so much on the human to make those judgments and changes within the inspection programs. It's definitely a key aspect and it's something our development team is working on heavily.

Matties: When someone wants to improve their solder paste process, what should they be looking at?

Anderson: That's a great question. When you're looking at printing inspection, obviously, alignment is one of the key components, but it’s looking at the geometric shape of the solder paste deposits. Omron can not only give you volumetric information about that paste, to tell you you've got the right amount of paste, but also to make sure the solder paste is in that ideal shape. It’s making sure the paste isn't slumped, that you don't have stencil release issues, and you're capturing things of that nature.

Matties: When someone looks at doing the ROI equation on this, if you will, what sort of improvement yield or otherwise should they expect from this sort of system?

Anderson: The industry studies it, and these studies have shown that 50–70% of your SMT-related defects can be caught at the solder paste inspection process. When you focus on that area and get that corrected, essentially everything else on the line should fall into place.

Matties: In terms of soldering technologies, we're seeing some vacuum and vapor phase systems, and some laser systems.

Anderson: Those are key initiatives as well because we're seeing a lot in our X-ray inspection. Many customers are having issues with voiding, for example. Obviously, the vacuum ovens will try and address that situation, but many of the components are our bottom-side terminated. You've got a lot of these large voids, there are heat sink pads and LGA components that typically solder terrible. You get large voids and such, and Omron has the capability to inspect down to the individual void size, then down to the cumulative percentage of voids on a given pad. So, whatever the customer's requirements are, we can tailor the inspection to catch those types of subtle defects.

Matties: What you're really doing, aside from catching the defects, is building a data library for your customers to draw from.

Anderson: Exactly. SPI, AOI and AXI can catch defects, but the main initiative of these products is for process control. It's getting that process dialed in, so that you're not worrying, “Did I ship bad product to my customer?” You want to make sure that the process is set up for ideal conditions, and that you're constantly getting that feedback and knowing whether you're within the guidelines you've established for yourself.

Matties: Now, the problem that exists is there's so much data. What's the right data to look at?

Anderson: A key thing to differentiate between different suppliers is that it all comes down to software and how the data is presented to the operator. Omron has a newer package called Q-upNavi, where we take all this information from up and down the line, not only from within the Omron equipment but also from the pick-and-place and screen-print manufacturers, and we coordinate all that information to a database and give it to the users in a web format. They can pull it up whether they're at home or sitting at their desk, and they have access to all this data in a way that makes it very intuitive in a side-by-side layout for true defect root cause analysis. You can watch as a board progresses down the SMT line and see all inspection measurements and images whether each phase was judged good or bad at any one of the inspection machines. It's a lot of data to capture and we do it extremely well.

Matties: All right. I certainly appreciate your insights today.

Anderson: Thank you.