Designing For The Long Term

I was at the gym the other day and glanced over at a fellow gym-goer on their cellphone. I did a triple take as the phone was a flip phone that was maybe 4 inches wide and 5 inches high on each flap of the flip (making a 10 inch phone when completely extended).  On my third glance at this monstrosity of a phone I realized it was in fact a Blackberry that he had pulled out of it’s case/holder but the case looked like the bottom half of a flip phone. It got me thinking about design longevity.

I think back on the cell phones of the past and recent past and remember how clunky and awkward they were. That was maybe 5 years ago and those phones have long been sitting at the back of peoples’ desk drawers or hopefully donated to causes that recycle phones. I am amazed that these phone manufacturers continually get away with phones that will be obsolete in 5 years maximum. Why don’t we expect more from our mobile devices (in terms of longevity)? Do we really think your phone will last more than 3 years?

My most recent phone just passed away after 2 years. In my case I MAYBE dropped my phone in a bowl of soup, but I think it just got one of the external speakers; really I think the kiss of death was something a bad battery (which was not contaminated with soup). But even if it lasted another year and THEN died, would I have been upset? I don’t think I or most people would be because we have come to expect consumer products to have a shorter life span.

How do we design electronics for the long term? There are a bunch of great examples of electronics that have been built to last:

1. Military designs –Aside from the humongous budgets that most contractors have for their military products, the specs on military designs can be equally large in scope. Translation: The military gets high quality products that were expensive but are built to last. These products are often ahead of the technology curve (thanks to the money available), so the technology often goes obsolete later too. The final piece is that the harsh environments encountered by military personnel requires gadgets that are sturdy enough to last for a long time; the ones that function in the field can continue to do so for a long time. A good example would be this emergency radio which was recently torn down by EETimes after an eBay purchase. The 1950s internals reveal high quality workmanship with components that match.
2. Space designs — Although NASA’s budget has been cut back since Bush has taken office, this research intensive organization has produced some of the finest inventions for human kind. Not only that, they have a mandate to create equipment that can last for long periods of time. My favorite example is the Voyager 1 satellite, currently exiting our solar system and headed further than any other human instrument has ever been. Not only that, but this advanced spacecraft started taking up close pictures of Jupiter and Saturn before I was even born (first passed Jupiter in 1979). The fact that this machine is still functional, still running tests and still capable of sending back data until 2025 (est.) is mind boggling. Not only that, but the spacecraft has not had the advantage and protection of the earth’s ionosphere, so it has been taking much more direct cosmic radiation than normal electronics.
3. Power companies — These terrestrial behemoths don’t have to worry about cosmic radiation quite as much as the NASA folks, but they often have materials carrying hundreds of amperes of current over long distances. Unfortunately, these systems are in need of some updating (especially to accommodate new renewable energy resources onto the grid), but once they are built, I’m sure they will hold up. Usually power companies achieve longevity in their equipment by using high quality, high strength materials that are designed with enough overhead to manage higher loads that they expect to see (i.e. A copper wire that is designed to carry 1000A of current, but only carries 600A on a regular basis).
4. Nuclear Facilities — Some of the remnants of the Cold War include the control systems that decided whether missiles would fire or not. These are still some computers operational today in Russia that (we hope) are still making logical decisions. While I don’t agree with these computers in the first place, I sure hope they continue to hold up, otherwise it will prove to be a doomsday device. Proper shielding from radiation and free radicals help to prevent aging damage to electronics from fissile material, in addition to starting with high quality, military-grade products.
5. Autos — While the auto industry might be falling on its face currently, the designers in Detroit used to help drive new technologies in many other walks of life. Looking at cars that have lasted since the 50s and beyond, we see examples of simple yet elegant electrical designs that were meant to last. Cars have not always had the GPS systems of today (which I’m guessing will have a much shorter lifespan), but have had electronics powering the wiper blades and the spark plugs for a long time. These systems in vintage cars require some maintenance and the occasional fuse replacement, but on the whole are sturdy enough to continue powering well-cared for vintage vehicles.

So these industrial/military and some commercial applications obviously present the need for longevity in finished products. However, designers need to consider many different parameters of a system in order to produce the best product for the long term.

1. Communication protocol — This item applies most directly to cell phone makers and is a decent excuse for their short life products (but does not excuse everything about them). Unfortunately for phone users (and fortunately for phone makers), wireless protocols are always changing in order to try and achieve the highest bandwidth, usually through higher frequencies or different transmission methods. So once a technology changes for good, older phones become obsolete (and the phone makers happily sell you a new shiny one). This problem also exists when looking to the internals of products; to prevent obsolescence due to outdated protocols, they should be standard to the industry in which the product will be used, simple enough to incorporate into a new standard (and included legacy support) and well documented. Nothing is worse than having a 20 year old device that works fine but can no longer transmit information. An example might be an industrial test fixture on an old computer that only has a 5.25 inch floppy drive. The test fixture might work great, but getting data off that computer is no longer viable so the entire setup is obsolete. A tried and true method for machines to communicate has always been serially and with good reason. While a newer communication protocol might require myriad signals that are not available on an older product, most improvements to a serial signal are often speed (increasing the frequency of the oscillator driving the serial line) or encoding. Since devices can be re-programmed to send a new encoding or you could slow down the device on the receiving end, serial communications seem to be a viable solution for lots of applications.
2. Long term drift of components — Designing for 10’s of years often requires attention to detail and deep pockets. The most important first step is to watch for this parameter on a data sheet for any critical component (marked as “long term drift”, often given as a percentage change over a specified period). But beware, many vendors simply leave this data off of their spec because they either do not think it is relevant, do not want to display poor data or because they don’t know what it means. In any of these situations it is critical to demand this data or to perform testing yourself in order to create lasting products.
3. Susceptibility to thermal stress — Size matters when it comes to handling thermal stress; this is partially why older electronics hold up so well. The smaller components on a device get, the less heat they can dissipate (assuming similar materials in a larger package). A good example would be resistors. A 0603 resistor (.6mm x .3mm) can only dissipate 1/10 of a watt while a standard through-hole component can dissipate 1/4 watt on average. This is a trade-off that must be made in any system designed for portability, but could result in lower product lifetime (especially in high heat or high current situations).
4. Standard packaging — The chip industry is a highly competitive environment where silicon designs are always being touted as the next best thing. Unfortunately for older products, this can often mean that components such as op-amps or a buck converter will no longer be produced. It’s a symptom of being in a dynamic industry and has to be dealt with. The best way to combat obsolescence is to create projects that have standards designed in to them. Thinking about creating a great new analog circuit with a non-standard pin-out in a device package that is so obscure that you have trouble finding it in catalogs?  Why not try making some other compromises on your circuit board and squeezing in a proven SOIC-8 with a pin-out similar to 4 other op-amps. You’ll be happy you do in about 4 years when that op-amp you’re using goes obsolete.

There are probably other ways to help design a product with a long life span, but these are a good start. A common theme is to pay more for higher quality components, which might not be preferred in certain situations. However, designing products for the long term can help save money year after year by not having to replace products or maintain sold products so spending a little more up front could pay off in the end. Some newer consumer electronics industries create new products each year either to drive demand or to fulfill needs after older devices break (which they may have produced).  In the process, they try to drive cost down by using the cheapest parts available; this can cause failures and unhappy customers. To design a long term product, costs and long term design considerations must be balanced.

What’s the longest period you’ve ever had a piece of functioning electronics? What kinds of changes did you see over the years? Have you ever created a low cost design that lasted more than 5 years? Let me know in the comments.