Issue link: https://iconnect007.uberflip.com/i/1284035
18 SMT007 MAGAZINE I SEPTEMBER 2020 on performance out of the box, and making sure that when the customer opens the box, it turns on. The cellphone companies are great for this. My lab sees nothing from the cellphone compa- nies because they're not concerned with prod- uct life five years from now. They don't neces- sarily want their product to last too long. They want you to buy next year's phone. They really don't need to do reliability testing because their expectation for their product in the field is such that they have a very short life, and they only have to worry about the short warranty period. They don't spend a lot of money or a lot of time doing reliability testing. They're focused solely on cost savings in manufacturing. Johnson: Any parting thoughts on reliability? Neves: I spoke in general terms about reliabil- ity in my discussion here, but many people talk about reliability and what it takes to be reli- able. It's a hot topic. My goal here centers on motivating people to look closely at reliabil- ity versus robustness and getting it into peo- ple's heads that they're not the same and can't be used interchangeably. Companies need to understand the testing they are using and why they are using it. It's really about understand- ing what you truly get out of the reliability or robustness testing you are doing. If you're doing a robustness test, understand that you may be paying a little bit less now, but it may be cost- ing you more because you're setting a bar to a higher level than you really need for your prod- uct to be reliable in your environment. Short- cuts can cost money, or it may be the other way around. You may set a bar that's too low. If you use someone else's experience and data and set a bar that's too low, all of this money and time you spent on testing doesn't give you what you really want to guarantee the acceptable life of your product to your customer. Right now, companies seem to be focus- ing more on legal liability—doing something that looks tough—rather than understanding what it takes to ensure true product reliabil- ity. They look at these tests and say that going more cycles in a high-stress environment gives you more reliability. It's just not necessarily so. That confusion has been carried on in speci- fications, standards, discussions, and papers. People refer to these torture tests as reliabil- ity testing, when—in fact—they're robustness tests. Matties: Bob, thank you so much for all your time. Neves: Thank you. SMT007