Estimated reading time: 4 minutes

Learning With Leo: Where Reliability Actually Hides

In my many years of training, it’s funny to see what still surprises people. For example, they’re surprised that by the time the solder melts, most of what decides whether that joint survives has already happened before the iron and even before inspection. In fact, it’s even earlier than that. In my column last month, I argued that most soldering problems are really problems of interpretation, which leads to the materials themselves, and where reliability is won or lost. So, let’s talk about what we’re actually soldering today.

We’re not working with simple copper leads and forgiving materials anymore. On any given week, what comes through our classes includes:

• Hardened alloys like Kovar and Alloy 42 that don’t tolerate forming stress well
• Gold-plated surfaces that can introduce brittle intermetallics
• Multilayer ceramic capacitors that can crack from something as simple as board flex
• Flux chemistries that are expected to perform across wildly different material combinations

When we expect one soldering process, whether wave, selective, or reflow, to handle it all, that’s a tall order. One of the biggest surprises for customers is how much variability is built into their process before the solder ever melts. Lead materials, laminate properties, component mass, and thermal profiles are the process, not just small details.

For example, I still run into manufacturers treating “gold is gold” as a working assumption when it isn’t. There’s a world of difference between a thin gold flash and heavy gold plating, which shows up right at the solder joint. If you don’t know the thickness, and many teams don’t, you’re taking a risk because excess gold forms brittle gold-tin intermetallics. Those joints can fail in ways that don’t show up until the product is already in the field. This is frustrating because it’s completely preventable. All it takes is verifying the plating thickness with your supplier rather than relying on a label. I find that step gets skipped far more often than it should.

Rework raises the same theme. Toward the end of almost every class, someone asks, “Can we rework a Class 3 assembly?” Yes, the standards allow it, and IPC-7721 exists precisely to govern how. But rework is not a free pass. It’s a controlled process that has to restore the product to full compliance with the documentation and approval the job requires. Companies find success when they treat rework as a formal engineering process. The opposite is true: You’ll struggle if you treat it as a last-minute fix. That difference shows up later when the board’s reliability is put to the test in the field.

One tool I wish more organizations used regularly is microsection analysis. When you look at a solder joint under 200x or 300x magnification, you see wetting, intermetallic formation, voids, and cracks—things that aren’t revealed in external inspections. I’ve seen companies debate inspection criteria for weeks, only to resolve it in a single afternoon once a proper microsection was on the table. The proper data changes and accelerates these conversations.

Few failure mechanisms keep reliability engineers awake at night more than MLCC cracking. These failures are often invisible at first:

• A board flexes slightly during handling
• A placement force runs a little too high
• A thermal profile ramps too quickly
• A microcrack forms

Everything passes inspection, but months later, it fails in the field. It’s one of the clearest examples of understanding the “why” behind the process matters. You can meet every visual requirement in IPC-A-610 and still ship a latent failure built right into the product.

I’ve spent years as a trainer, consultant, and VP/technical director at EPTAC. I’ve watched these patterns unfold. I believe we have an interpretation problem where the organizations that consistently build reliable products do three things differently:

• They connect acceptance criteria to real failure mechanisms
• They validate their processes with data, not assumptions
• They train their people to think beyond the picture in the book

I come back to that last point most often because soldering is about managing a complex interaction of materials, heat, and mechanical stress, and doing it consistently, assembly after assembly.

Nearly every question I get, whether it’s about toe fillets, gold removal, or flux selection, is really a variation of the same question: Will this product survive in the real world? The standards can guide you, but they won’t answer that for you. That requires understanding, discipline, and a willingness to look below the surface. It’s also the work we spend most of our time on at EPTAC, in the classroom, and out on the floor. From where I sit, that’s where the real work begins.

Resources

• Modern Solder Technology for Competitive Electronics Manufacturing, by Dr. Jennie S. Hwang, McGraw-Hill, 2002.
• IPC-A-610, Acceptability of Electronic Assemblies.
• IPC-J-STD-001, Requirements for Soldered Electrical and Electronic Assemblies.
• IPC-7721, Rework, Modification and Repair of Electronic Assemblies.
• Solders and Soldering: Materials, Design and Theory, by H.H. Manko, McGraw-Hill, 2001.
• Surface Mount Soldering Techniques and Thermal Shock in Multilayer Ceramic Capacitors, by J. Maxwell, AVX Technical Publications, 2001.
• Product Reliability, Maintainability and Supportability Handbook, by M. Pecht, CRC Press, 2009.
• “Reliability of Lead-Free Solder Interconnections,” by P.T. Vlanco, Journal of Materials Engineering and Performance, 2004.

Leo Lambert is the technical director at EPTAC Corporation.