Issue link: https://iconnect007.uberflip.com/i/1509873
12 PCB007 MAGAZINE I OCTOBER 2023 the same problem that I saw presented around 2004 in the U.S. It's now 2023, and they're pre- senting the same problem, with the same test methods, and almost the identical test vehicle design. e U.S. establishment went through this 20 years ago as well with almost identi- cal stackups and structural configurations, and have actually not advanced that far past this point, but it was far enough that there was still value in a conversation. I offered some input on the topic, but in sit- ting on these reliability review meetings, you can see why nothing gets done. It's all "coopeti- tion"—everybody is some kind of competitor. I don't understand how we can get them to do anything at all. Everybody's silent when some- one asks a question. I think that's why there's no progress on the topic, and that's why people still talk about the same problem 20 years later. ere needs to be well thought out incentives to manage the coopetition and improve the quality and capability of the ecosystem for tech advancement. Matties: Are outside organizations more apt to solve problems like this? Honestly, anybody who's in that business and running HDI fabrication facilities should have potential product value/willingness to pay, so fabs must make sure that they correlate their testing results to their product design attri- butes. But that precise regression is missing in a lot of cases, and there's a lot of diversity of test/qual methods. e big question is: What are the defects and/or parametric conditions that lead to reliability failure and can they be normalized? Today, many people are talking about HDI failures. Stacking microvias has been a common topic lately. ere are really three failure modes that occur, and everybody has a different way to deal with them. A large, influential, unnamed semiconductor company will deal with them in a completely different way than a typical fab shop. e semicon actually biases toward one failure mode by design, because they bought equipment that is designed to deal specifically with that condition, and they take advantage of it. But most regular fab shops got started on plated through-hole (PTH) technology, and the equipment is not set up for HDI. ey try to do both processes with the same equipment and the trade-space is not favorable very oen when it comes to HDI reliability. On top of that, the whole process comes down to this: What chemicals and equipment are you using, and what are the steps? What's the whole process. What is the copper sup- posed to look like? What is the metallization that you're supposed to be using? It may not be a magazine-selling idea (laughs), because many of the challenges are oen control prob- lems and embarrassing, but that's really it. e process methods that are proven to drastically improve stacked microvia reliability are almost never discussed. Nolan Johnson: Fabricators still struggle with these challenges, so is there a lack of progress here? Yes. I recently got involved in reliability again for a conference in Munich. A major European aerospace organization had a problem, and I was solicited to talk about it because it was Alex Stepinski