[perlgeek]

I don't know what they did in this particular instance, but basically there are two things you can do: shield, or in the case of electronics, use larger parts.

The smaller some piece of electronics is, the smaller the charges in there are, and the easier it is to introduce errors with ionizing radiation.

[_ph_]

And the design of the electronics might help with radiation tolerance. Computation paths could be redundant to detect/eliminate bit flips due to radiaton.

(Some server Power cpus seem to run parallel in pairs, comparing the output at critical parts to detect cpu errors)

[rurban]

Basically the same as in space exploration. Also running at very low MHz.

[FiatLuxDave]

There is a third thing, which is to use a dopant which reduces minority carrier lifetime and provides radiation resistance when manufacturing the semi-conductor. Platinum works really well for this. Unfortunately, this can be a poor engineering tradeoff for some kinds of electronics (power transistors for example) and usually increases electrical resistance. AFAIK nobody manufactures platinum-doped microprocessors.

I used to be a product manager for radiation measurement electronics, and we used platinum-doped diodes for the part intended to go in the beam.

See for example: https://academic.oup.com/rpd/article-abstract/17/1-4/527/311...

http://ieeexplore.ieee.org/document/5436014/