Semi-feasible alternates to silicon electronics for a Von Neumann Probe?

Question

Disclaimer: this is a modified cross-post of my post on the Worldbuilding SE site, adapted to better fit the Space Exploration community.

Some background

A Von Neumann Probe (VNP) is a sci-fi probe which explores the universe in a self-replicating fashion: it finds a location to recreate itself, does the necessary mining and processing to duplicate every part of itself, and then sends off the duplicate to start the cycle anew. VNPs have been associated with the Fermi Paradox.

Investigations into assembly I know of

There have been various explorations into the idea of a VNP, which varying degrees of scope. Importantly, all of them seem to have the fundamental barrier of silicon chip production.

1. NASA's Advanced Automation for Space Missions project - a deep-dive into the feasibility of creating a self-replicating factory on the moon. Most components were theoretically capable of being manufactured in situ, but in terms of electronics, Wikipedia describes that "a more speculative, more complex microchip fabricator was specified to produce the computer and electronic systems, but the designers also said that it might prove practical to ship the chips from Earth as if they were "vitamins.""
2. RepRap and Snappy: 3D printed 3D printers - These projects are quite cool, but they are limited in that they're unable to print any of the electronics. Snappy, for example, has peaked at around 70% 3D printed parts. I assume this problem would exist for CNC'd CNCs and metal-sintered metal sintering machines too.

Potential solutions

The way I see it, producing silicon chips is far from being achievable for a VNP at this point in technology, since producing a full-scale clean room and electronics factory would take a VNP the size of cities or larger. I will assume graphene electronics technology is impractical as well, since that's still in its infancy and would probably need clean room conditions as well.

With this in mind, are there any somewhat practical alternates to silicon electronics that could be used for a VNP? Vacuum tube technology, pneumatic instead of electric transistors, mechanical computers, something else? Emphasis on the practicality of the designs.

Answer 1

As I said in a comment, I think this question might belong in worldbuilding. Here however is a proposal for how you might solve the problem.

I think valves (vacuum tubes) would be very manufacturable using relatively simple technology, certainly compared to semiconductor manufacture. For a probe which spends its time in hard vacuum or which is willing to do the self-replication in fairly hard vacuum, then they are also delightfully easy to evacuate – indeed for any parts of the system which live in hard vacuum all the time you might not need envelopes at all, certainly not for components which only need to switch, such as the parts of a digital computer.

Another advantage would be that the system would be very resistant to radiation and other insults which hurt electronics in space.

Significant disadvantages would be power requirements. Long-term reliability might be a problem as well, although this might be soluble with enough research.

This looks like a mad idea, but it is not perhaps quite so mad as it seems. One reason it looks mad is that modern processors have billions of transistors, and a machine made of billions of valves would be huge. But, if we accept the Church-Turing thesis (which is certainly true for the capabilities of any modern computer), and we are willing for our computation to be slow (which might be perfectly acceptable for a probe which (or the offspring of which) needs to last for millions of years) then we can live with computers with relatively tiny numbers of active components. What limits computation is really memory, not how big the active part of the machine is. It would be possible to build a machine with only a few thousand active components, but an enormous memory, which would be perfectly able to compute anything you like, just rather slowly.

Of course this might be seen as simply pushing the problem elsewhere, to the construction of a very large, low-power (ideally zero power except when being read or written), extremely reliable memory. However I suspect that this may still be much more achievable than semiconductor manufacture. For instance tape meets some of these requirements (as large as you like, zero power when not being read, probably not anywhere near reliable enough, probably still fairly fiddly to make).