Nothing lasts forever. I don’t intend that as a philosophical statement, but as a physical one: entropy always wins, in the end. Maintenance, in a sense, is a long, losing battle to keep entropy at bay. Parts wear out, components fail, systems degrade: to keep a machine operating indefinitely will eventually require replacing every part of it, in the fashion of the philosophically debated ship of Theseus. For most systems, this means there comes a point of diminishing returns, when it is more practical from physical, financial, or logistical reasons to replace rather than to repair. It’s a problem engineers must consider in every design choice, and it’s at the heart of what is derisively referred to as planned obsolescence.
That phrase has acquired a distinctly negative connotation, mostly associated with smartphones and other emblems of modern consumer technology, and it is used to refer to the perceived practice of setting up some component in a device to fail after a period of time, thus prompting the user to spend money on a newer model. Deployed as part lament, part accusation, it’s come to be considered a rather underhanded business practice, but I can never quite join in the griping. There are plenty of things about modern consumer technology and its associated business practices which I find displeasing, but planned obsolescence, while it can certainly be deployed inappropriately or cavalierly, is a necessary part of any engineered product. In fact, planned obsolescence is arguably preferrable to the alternative, which would be random and potentially catastrophic failure.
Consider a standard satellite, with the usual subsystems: structure, thermal control system, reaction control system, electrical power system, command and data handling system, telemetry tracking and control system, and guidance, navigation, and control system. Recent advances may lead to changes, but spacecraft design typically requires spacecraft to operate in one of the harshest environments for long durations, with no physical maintenance for the entire operational life. There’s not even an option to refuel (that might change soon too, but then, I’ve been reading about satellite refueling demonstrations since I first grew interested in space). When we design a spacecraft, we design it with its intended mission life in mind, choosing components and making design choices with the goal of keeping it operational for that amount of time explicitly in mind. Propellant capacity, reaction wheel durability, battery capacity, even solar array capability are all designed based on anticipated degradation over the mission life to provide a minimum capability at projected end of life. Inevitably, this means one or two components will tend to be the limiting factors, failing before everything else. Running out of propellant is a common case, but so are reaction wheel failure, battery capacity running too low (which will sound familiar to most smartphone owners), and failed sensors. If we design it right, it won’t suffer those problems until after it’s completed its mission, but there’s another way in which spacecraft are similar to smartphones. They’re exquisite, expensive pieces of technology that are usually still capable after their design life ends, which means their owners want to keep operating them.
I think the oldest satellite I worked on was almost forty years old. That’s forty years of continuous operation in space, with no maintenance, no refueling, no going up there and doing a hard reset. It had a design life of seven years, so it kept going way past what its engineers anticipated. Of course, its companions, based on the same design, weren’t still around, so it was a bit of an exception. It had almost no propellant remaining, we had to teach ourselves how it worked all over again any time there was a new problem because everyone who designed it was long retired and its technology was markedly different from modern spacecraft, and it often felt like we were just limping it along because it refused to give us an anomaly so bad we couldn’t fix it. A remarkable vehicle, and arguably, remarkably over-engineered. I worked on another satellite program which had a design life of one year, and those started to experience pretty significant anomalies after about eighteen months. Their design often felt a bit slapdash compared to other spacecraft because, well, they only needed to last for a year. That never stopped the customers from trying to operate them for longer, though, or from being upset when they started failing after three years.
No matter how well engineered your device is, eventually it will fail. Even the Voyager probes, faithfully exploring the interstellar boundary space while limping along on the dregs of their nuclear power supplies, far longer and farther away than the designers expected to operate them, will eventually reach a point when there is no longer enough power to keep themselves at operating temperature, to run instruments, to send transmissions and receive signals from Earth. That point is not far off, no matter how many instruments and subsystems the engineers on that program manage to turn off in a bid to eke out just a few more months. Smartphones are exquisite pieces of technology, but they are consumer products, mass produced, not unique, arguably overengineered spacecraft (nor should we want them to be like spacecraft, for they would have price tags to match). Just like spacecraft engineers, the engineers at Apple, Samsung, and the other smartphone makers must make deliberate decisions about the expected lifespan of the device. Will it be the battery capacity that goes first? The antenna? Perhaps the charging port will wear out first, or the contacts behind the screen will degrade, or the adhesives holding the case together will degrade. They face another tradeoff, too: how much are their customers willing to pay for a bit longer life? Doubtless they’ve done analysis to determine that most people will themselves damage the phone to the point of requiring replacement within a given statistical period of time, and they make their design decisions with whatever that intended average lifespan is in mind.
Now, it’s true that phones could be made more repairable and maintainable, but that’s also a tradeoff, as we discussed in our review for Maintenance: Of Everything. Improvements have been made in that regard. Even Apple, notorious for making it difficult for customers to access the hardware they ostensibly own, has improved maintainability of their smartphones. Replacing the battery, for instance, can be done by removing a few screws, opening the case, applying a small voltage to the battery, disconnecting it, and putting in a new one. Sure, it’s more involved than replacing the batteries in your TV remote, but for what the smartphone is, for the complexity of the device and the standards of performance it is expected to meet on a daily basis, it’s not terribly complicated. How many people do you know who have replaced their phone battery rather than upgrading to a new phone? Unless you’re in a community inclined towards that sort of thing, it’s probably not many. Designing for end of life means the primary mode of failure can be a graceful dropoff in battery life, giving the user time to notice the changes and affect a replacement, rather than a sudden, unplanned failure of a major component.
There are plenty of problems with the consumer device industry these days, from the waste associated with devices which reach their end of life or are simply replaced with a newer version, to the massive invasions of privacy over which we have only the illusion of choice, to the frustrating debates about the definition of “ownership” and the endless subscriptions and license agreements without which the hardware can be rendered useless and content you thought you owned rendered unavailable. Yes, maintainability and repairability could be better, and yes, planned obsolescence can contribute to these ills. At the core, though, planned obsolescence is an engineering necessity. Everything fails eventually. It’s only a matter of when.
