Issue link: https://iconnect007.uberflip.com/i/1275152
28 SMT007 MAGAZINE I AUGUST 2020 at the same time, but there were micro-vari- ances in processes and key materials. And if there was a big variance, it would have been caught as a defect in the factory. But a combi- nation of several influences put together can create a critical weakness, such as a joint fail- ing after a certain time. With traceability, you can pinpoint exactly the unique circumstances that contributed to the failure, and then with software, monitor and ensure that the same condition is avoided, including poka-yoke machine control where we can command a machine to stop because what it is doing is likely to create a defect. And that's the way to use traceability as a closed- loop system, in a positive way such that you can learn from even the most complex and rare defects. Johnson: I'm reminded of the nursery rhyme, "For want of a nail, a horseshoe was lost," which is essentially a supply-chain story. Ford: We all depend on our suppliers, and we are all links in the chain. We need to trust that what we're getting is what we think it is. We see the loop closing now, where the product goes out into the market, where main- tenance and repair work is done in association with the original manufacturing data, such that we have feedback about product reliability and performance back to design such that compari- son can be made against expectation. Johnson: In a previous job, I wrote software for integrated circuit design. Back annotation information, even 20 years ago, was the weak- est link. How can you determine how rev B performed in the field so that you know what to change on rev C? Ford: As a designer, you think, "I did every- thing possible. I did all my design rule checks, and everything was perfect." And then you have the product made most likely in a distant place where they use an alternative material that you didn't take into account. You should run all the DFM again, but it's too late from the design perspective since the board is already produced, you have to live with it. Until now, there hasn't been a mechanism to say, "For rev C, let's learn from what we did for revs A and B and look at the performance, the production compromises, the material variances, and the equipment capabilities that we know is going to be used to make this product, and let's use the feedback to influence design improve- ment." This logic can boil down to a design feature, as opposed to a product, such that you can have the same benefit when making new designs. Johnson: Is there a responsibility on the part of an individual company to do their part within their small loop in the cycle? Can we talk a lit- tle bit about what the benefits are, and why those individual players should get involved in providing more information for traceability? Ford: From the traceability aspect, it is impor- tant that everyone works together. Internal traceability benefits manufacturing in terms of quality. Externally, each entity in the chain will benefit from having incorrect responsibilities perceived, and also increased business oppor- tunity in that the gray market becomes secure, and DFAR processes opened up. The way of working together, however, has to be defined in that data from all of the different parties must be made somehow compatible. In this way, a greater benefit from the whole can be enjoyed. Within IPC, another standard that has taken shape recently is the IPC-2551 digital twin. The approach that has been taken is quite unlike solutions for digital twin that have been appearing recently in the market. We have created a framework—an architecture—that defines the interoperability of how information from all aspects of the overall manufactured product lifecycle should flow. As with trace- ability, there needs to be a benefit for each part involved for it to be sustainable. Nobody wants to have to do something for somebody else's benefit without getting a benefit themselves. You may be a contract manufacturer that needs to get a quote done very quickly. You may be designing a manufacturing line, and you need to understand the capability, throughput, risk,