When I worked on the Hubble Space Telescope in the late 90's I kept hearing the electrical guys talking about the new computer we were going to install. They called it simply, "the 486". I one day joked that I had had a 486 a few years ago in college and maybe they could use that and save some bucks. And they told me it was the same thing - literally an intel 486 processor. They said that the design happened to be very resistant to radiation, and met their purposes very well.
I haven't really kept up on it since, but I believe it's still running a humble 486. Granted, the rest of the machine isn't much like my old Dell, but I always found that interesting.
Yep, that radiation hardened 486 was used quite a bit. I believe it was used for the Space Shuttle as well.
The Mars rovers have moved to a more efficient PowerPC platform, but its still old tech. Radiation hardening CPUs isn't cheap and the processing requirements for what these projects do isn't high (its more sane to just send the data to NASA and have terrestrial computers do all the heavy lifting), not to mention you don't run a bloated desktop OS on these things. A minimalist VxWorks RTOS is typically used.
Price? The RAD750 on the Curiosity Rover starts at about a quarter million dollars.
The Space Shuttle should be so lucky - 8086 (though I read 80386 for a cockpit upgrade). First flight 1981, predating the 486 (1989).
It's interesting, NASA was raiding all sorts of old machinery to keep in stock of 8086s.
If memory serves, older process technologies are intrinsically a little more rad-hard, just because a 1um 6T cell is roughly a hundred times larger than a 14nm 6T cell... (giving it greater drive strength and larger "ballast parasitics" to absorb radiation events)
They probably built them with a lot more margin back then too.
I haven't really kept up on it since, but I believe it's still running a humble 486. Granted, the rest of the machine isn't much like my old Dell, but I always found that interesting.