Tounge-in-cheek noted, there's really not a lot of margin in space parts. Generally they're so expensive because the purchaser wants to be hyper confident that piece is going to do what it needs to do. The cost goes astronomical (bah-dum) because of things like:
- Significant nondestructive and destructive testing (sometimes 20% of a lot is set aside for lot testing...or even more)
- Really low lot volumes (hundreds or thousands of parts instead of tens of thousands or more), so much higher per-part non-recurring costs (setup, etc.)
- Full traceability from material orgin through final installation location and everything in-between
While some space parts are indeed "better" than equivalent lower grade parts (automotive, etc.) due to things like tighter tolerances or uprated materials, the irony of many space parts (typically less important ones) is that they are literally the same thing as lower grade parts made on the same machines as the lower grade parts...just in way smaller runs. Part of the general COTS philosophy is that whatever you're losing by not doing significant lot testing or traceability exercises, you gain in sheer statistics--the variation in a COTS lot that's orders of magnitude bigger than a space lot is often way more favorable for the end user.
Maybe a little adjacent, and I've probably posted this before, but on the subject of parts selection Doug (basically one of the pioneers in 'volume' smallsat components) and Jonny (formerly the CTO of Skybox and one of the founders of Muon)
wrote a paper a while back that a lot of folks still use as guidelines.
Yeah, especially in low LEO (and a very short life), you really don't need a ton of tolerance for even the really important parts. A combination of shielding and redundancy is really all that's required. Shielding also doesn't need to be dedicated or fully enclosing either--it is more or less additive. Any bits and pieces that might 'block' your important components from those pesky particles count as shielding. One of the things you do toward the end of a satellite design is ray tracing from the important components to understand what physical directions of dosage are actually untolerable. Then you might only need to put a block of shielding on one side of that processor, or maybe to slide that reaction wheel (that is mostly a big block of aluminum anyway) a bit this way or that to provide "free" shielding.
In context, the Snapdragon on Ingenuity actually checks out. Mind, I'm sure it made the old guard go head-explody for a while in Pasadena, but if one actually evaluates the use case it's not some giant leap:
- It had a pretty short life in a radiation environment. Even though space and Mars is, of course, much harsher than LEO from a radiation perspective, its fair to assume Ingenuity was pretty well shielded during the transfer to mars and then also reasonably well shielded sitting on the surface of Mars (half the time not facing the sun). The major upside here is that permanent damage due to total dose would have been minimal (and likely a non-issue).
- In addition to what was likely some component redundancy to mitigate the effects of upsets, the inherently short mission duration of each flight would statistically minimize the possibility of some kind of single event actually happening during flight. No doubt the startup sequence scrubbed/reset the important chips prior to liftoff, so basically all the the processor (and other important/susceptible components) had to do was stay clean for a 1-2 minute flight.