Issue link: https://iconnect007.uberflip.com/i/1284035
SEPTEMBER 2020 I SMT007 MAGAZINE 57 when you're trying to build a very highly reli- able product? Where do the failures mostly happen? I'm not saying there are a lot of fail- ures; it might be 1%. Malmrose: One thing that has been an issue, as well as the other places where I have worked, is the calibration of the torque wrenches. When you're putting bolts on, or putting in screws, if the drivers or torque wrenches are not in cal- ibration, the biggest concern is overdriving or underdriving. It must meet the specification of the build on the mechanical. If you look at reli- ability, something's going to have vibrations, and that's going to be an issue. Another one I saw at a different company was a latent reliability issue from hydrogen embrittlement of bolts. We found that the bolts in the actual final industrial product were being subjected and exposed to water, and we had bolts breaking in the field. When the bolts were being manufactured by the manufacturer, they were not being processed correctly, so we were having field failures of the mechanicals there. The quality and calibration of the equip- ment are also concerns, particularly if you're a small company. If you're doing riveting, it's important to make sure you're not overdriv- ing or underdriving what's going on. We run into this all the time, where a company says, "Here's my drawing. I'm going to launch it over to you so that you can build me a hun- dred boxes," and they were supposed to be prototypes, but they end up not being proto- types. Many small companies may end up sell- ing them for production. But it's very com- mon to not have assembly instructions with requirements for how tight the bolt should be or how hard something should be driven. If you're driving a screw into a piece of plastic, you overdrive it, and it's going to fall out. It's a lot of looking at the product usage and making sure that the specifications are being followed. It's not right for any customer just to launch something, throw it over, and say, "We just expect you to figure out what the specifica- tion should be for driving this screw into a piece of plastic." There is mechanical engineer- ing needed to create the specifications from the customer. We can always do it, but it takes money. And that's typical with small companies. If they're worried about product reliability on a build on the mechanical side, make sure that it's properly specified in terms of what the manufacturing specs and performance need to be. Johnson: As we move toward harsher environ- ments and more mobile, automotive, medical, and life-critical applications, the reliability of the enclosures becomes just as important as electronics in many cases. Malmrose: Absolutely. If it's going into a harsh environment, you have temperature and humidity. Recently, in one of our box build situations, we did different kinds of coating, and one of our coating processes was crack- ing. If that got out to the field, you could have a latent humidity problem and corrosion in the cabling if cracking occurred. Johnson: With that in mind, what do the test- ing qualification procedures for box build look like in your factory, and are those changing? Malmrose: For some, they'll do drop test. It's like when we used to drop things from three meters and ask, "Does it still work?" But we'll do that, and then we do full burn-in, so we'll do accelerated life testing. We can do humid- ity testing as well as burn-in on the mechanical side. We have one very large burn-in chamber. If you don't have sufficient thermal conduc- tivity, then you're going to have a board-level failure. Also, if you have adhesive that's part of the box build and is not rated for a certain temperature, the adhesive could fail. Some boxes or enclosures are adhesively attached, not mechanically attached. But we can acceler- ate life failures for those, both hot and cold, as well as doing basic drop testing and shaking. Johnson: Are the requirements you receive from OEMs, or the ones who are specifying Roger Malmrose