Ask HN: Could we build a long-life probe like Voyager today?
6 comments
If they were built by thoughtful engineers with a huge nerd-on for historical methods and not much care for the budget, then yes it's very likely.
If it were built by the same rote as every other probe with short term life and cost as a high priority, probably not.
If it were built by the same rote as every other probe with short term life and cost as a high priority, probably not.
Aside from trace size, a big problem here is that devices nowadays are just so complicated. There is a ton of cruft and bloat; thanks to computers used in the creation of things, every new car is its own Godel Escher Bach.
You probably want to use older tech. I don't know how electromigration will work on 7nm chips. You also want to make software simple so its easier to test.
i think the hardware will not be the problem, but the software. i do not believe the level of scruitiny that has been set in the past, where multiple people went over PRINTED code line after line, multiple times, is something that would be done these days.
Given the technology of the early 70s (very large feature size, very low level of integration, wide traces, probably not CMOS, etc) it's been possible to understand them well enough to keep them running and repurpose them. The more modern mars rovers also exceeded their mission lifespans by several times, but had ultimate lifetimes of about 15 years. And I note that one of the probes had an early failure of flash memory, something that didn't even exist in 1975. So while they exceeded their mission by more than 30X, in absolute time it was much shorter than Voyager.
But outside that kind of almost unlimited budget environment are we still building things with unanticipated lifespans? Everything from electronics to housing is optimized for cost of manufacture, not long term operating value. And that extends far enough through the stack (from semiconductor to UI) that I wonder if design for the long term is even possible.
Sure, a few hackers can reflash an old phone to be a TV remote, but the bulk of the population can't or can't be bothered. More dangerously, in a few years people will be scattering IoT devices willy-nilly in fields, over provisioning because it's OK if a few fail if you have enough running. When they stop working they'll just drop more, newer ones in; it won't be worth picking up the old ones which will just become e-waste in situ.
Is it even reasonable to want active devices to be designed with long lifetimes? Outside space programs do we even have the skills and technologies?