PCB007 Magazine

PCB007-Sept2020

Issue link: https://iconnect007.uberflip.com/i/1288481

Contents of this Issue

Navigation

Page 105 of 137

106 PCB007 MAGAZINE I SEPTEMBER 2020 of sorting. Currently, we encourage people to avoid higher stacks, so you will see a Motor- ola presentation on that company's require- ments for its own product. The IMEC testing on the test methods will be referred to as TM- 650 2.6.26 and 2.6.27A; they involve multiple reflow cycles simulating assembly with contin- uous resistance monitoring. Microvia Failures When I first heard of this problem, I thought, "This is just a common problem out there of mi- crovia failures." Figure 1 is a roughly 10-year- old slide, showing a one-stack microvia that obviously did not include a well-cleaned land- ing pad in the separation between the microvia fill; the internal layer is obvious on this cross- section. This is not the simple problem that we encounter today. In Figure 2, a cross-section from Motorola Solutions shows an unpolished cross-section of a three-stack microvia on the left. On the right, with reflow and careful micro-edge, it is possible. You can see a line there, but this is a functional resistance that shows up in the test- ing. And you might pass this just with optical cross-section. It is an almost invisible problem because it's subtle—and that really bugs us. Figure 3 shows the methodology of using in- strumentation to find these problems. Figure 3 shows the instrumentation involved in finding these weak microvias. The two graph- ics on the left show thermal cycles. Against that, the resistance of the microvia sections is includ- ed in the daisy chain. On the third thermal cy- cle, the resistance went up infinitely, and the microvia opened. However, you can also see in all the cases the resistance comes back down after the end of the thermal cycle. In Motoro- la's case, with this data, when they stagger their microvias on the right, the thermal cycles did not change the resistance of the microvia daisy chain, and those are shown as good microvias. With information from Raytheon (Figure 4), further analysis shows that there was no single Figure 2: Two microvia stacked constructions with the typical microsection on the left, where the problem is invisible, compared with after reflowing, where it is almost invisible. (Source: Motorola Solutions) Figure 1: Older microvia failure due to contamination on the landing pad.

Articles in this issue

Archives of this issue

view archives of PCB007 Magazine - PCB007-Sept2020