Advertisement

Article

Does reliability count anymore?

One of my cordless phones died, so I replaced it and put it into the box where I keep the other dead cordless phones. In this box were seven other phones, replaced over the past 10 years. I decided to do my own informal “CSI” autopsy on these units, opened each one, and got down to the PC board.

Three had defective keypads, with some sort of sludge interfering with the keypad contacts (plastic migration and deterioration, perhaps) despite my vigorous cleaning with contact cleaner and alcohol; three had intermittent connections somewhere (it could have been the PC board traces or IC connections), and two were “cause unknown”.

Despite outside appearances, a cordless phone is a fairly complex product that sells for a modest price, so maybe I shouldn't complain. I also rationalized that these handheld products get a lot of user abuse. But even with this rationalization for these defective units, I know I am professionally irritated and worried.

Certainly, all these discarded products contribute to the global waste pile, but that's not what really gnaws at me. It's that we, as professionals and engineers, have admitted we design stuff intended to be soon discarded, after relatively little use or time. It's not something to be proud of, nor does it reflect well on the profession. Even though we know why these products don't last, the average user likely does not know or understand, and soon associates the message of these throwaway items with our presumed role of engineers.

Back in those ancient days before ICs ruled the industry, it was fairly difficult to design and build reliable electronic products. There were so many components and connections, and so many things that could go wrong, that doing a solid design was only a small part of the reliability challenge. Manufacturing was the real issue, because sooner or later the connections, electronic components, or a moving part would fail.

But the “sooner or later” aspect is key. Once the product got past its infant mortality or burn-in period, it would be fairly reliable for two reasons: it was usually treated with respect, and it also had enough size and mass for decent mechanical connections. When a component did wear out, it was generally replaceable, so all was not lost.

In contrast, today's products often are reasonably reliable in their initial phase, but have little stamina for the longer run, and are not repairable. Not all consumer products have this problem, but too many do, and we accept it as the cost of churning them out quick, cheap, and under pressure. Saying to ourselves that this limited lifetime is OK because the unit is not likely to be used for more than a few years is somewhat circular and disingenuous.

It really means we are too often setting a low bar for design integrity. That's not a good thing for engineers or our profession to admit. Think about this situation as you ride your car, loaded with electronics, over a bridge built a hundred years ago and still doing its job.

0 comments on “Does reliability count anymore?

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.