Jeff Duntemann's Contrapositive Diary Rotating Header Image

Falling Back to Dieselpunk

My writing time has taken some hits in the last few weeks, but the weather has hugely improved. It got up to 72 here today, so with joyous enthusiasm I took a long walk. As usual, something occurred to me, this time when a badly adjusted dump truck went past and bathed me in fumes. Ahh! Dieselpunk!

The insight followed soon after: If the world went to hell for some reason and I had to build a vehicle, it wouldn’t be steam. It would be Diesel. With some study and care, you could render farm animals or even roadkill and make Diesel fuel. Diesel engine technology requires machining and some skill, but not exotic materials nor computer models. (The same is true of gasoline engines, but gasoline is harder to make than Diesel.) You can do it in a garage. Clean rooms not required.

I had some experience in thinking about recovering technology after a societal crash while creating the Drumlins world. The glitch there is that all the inadvertent colonists’ knowledge was in computers, but they didn’t have the critical mass of technology to make more computers, nor even fix the ones that broke. (Quick! We need ten pounds of indium! Jimmy, Sam! Go dig around and see what you can find!) So when the computers died, their technology died too, and they were back to a medieval style of life that might have stayed medieval except for the Thingmakers that shared the planet with them. Advanced technologies build on simpler technologies, which in turn depend on simpler technologies still. It made me wonder if there were a sort of minimum technology level, one that, with common sense, an oral tradition, and few old books, might be constructed more or less from scratch.

Speculation: Steampunk might be a consequence of ignorance (i.e., we don’t know enough yet) whereas Dieselpunk might be a consequence of a sort of poverty of connectedness (i.e., our societal matrix is neither large enough nor rich enough to build what we might find in old books or otherwise imagine, even if we knew how.)

It occurred to me that there was an interesting plateau of sorts between about 1920 and 1940. Most of the stuff that existed in 1940 existed in a slightly cruder form in 1920. During those two decades, we got better at doing the stuff we did before, but we didn’t invent a great deal of truly new stuff. WWII changed everything, of course, and nuclear energy and transistors and many exotic materials showed up by 1950.

The era 1920-1950 was the Golden Age of Back Room Science and Technology. You could do lots of interesting things with an engine lathe, a microscope, a slide rule, a gas stove, a source of electricity, and raw materials you could buy at the local drugstore, hardware store, and feed store. The science was straightforward, the technology simple. Most important of all, it was still possible to be a generalist. A hundred books (Ok, maybe two hundred) could teach you most of what we knew in the hard sciences. You could usefully master physics, chemistry, and math in less than a lifetime. Specialization has always existed, of course, but I think it became mandatory after 1950. After that, you could no longer hear street traffic near a university for Asimov’s Sound of Panting.

If the population of Earth were reduced by three quarters (especially by something limited to human beings, like a very nasty flu virus) high-tech civilization might no longer have the critical mass of human skill it would take to maintain itself. The computers would work for awhile, but after they died, all the support infrastructure (chip foundries etc.) would die with them, and what would be left after a few decades would be less Mad Max than Dieselpunk.

All that’s debatable, of course, and I could be completely wrong. I bring it up only as an insight obtained by getting out in the sunlight for the first time this cold season and making the blood pump a little. I’m taking notes on a fictional setting involving a new Ice Age, and now I’m sure it’ll be a Dieselpunk culture, with no computers but a great deal of steel, vacuum tubes, carbon black soot, and internal combustion. Neanderthals, too: Brute muscle mattered before WWII in a way it may never matter again. And airships–hey, they were huge in the ’30s! (Why let the steampunkers have all the fun?)

Now for the time and energy to finish what I’m working on now so I can get on to The Gathering Ice.


  1. Andy Kowalczyk says:

    I guess it depends on the scenario. If it was more of a Connecticutt Yankee in King Arthur’s Court sort of thing (where you have the knowledge but no infrastructure) then Diesel makes sense – you only need steel and not copper wire (Otto engines need generators and distributors and spark plugs and coils and points). But if it is a post-apocalyptic world, where scavenging is the main source of technology – then I think what the Germans did with wood gas in WWII is more reasonable.

    1. Actually, Diesel engines are tougher to build than I had thought, and one of my friends is an expert in the technology. (I’ll explain in an upcoming entry.) The scavenging point is interesting, and leads to a question I’ve always had of the Mad Max-type scenarios: How long does it take for all the useful pickin’s to be picked, and what happens after that? We haven’t had VLSI long enough to know clearly, but from what I understand, after fifty or so years any wafer-based item will fail because of devices “fuzzing out” due to metallic atoms wandering into junctions, and so on. Discrete transistors would last longer, but not forever. Furthermore, there are not a lot of discrete transistors in electronics anymore.

      1. Andy Kowalczyk says:

        Once you know how, is it easier for the basement engineer to build a vacuum tube or a bulky transistor? Geranium diodes where around long before transistors – so what chemical engineering is involved in making semi-conductor materials?

        I don’t think we need VLSI to improve life – but discrete transistors would certainly be useful.

        1. Andy Kowalczyk says:

          Sorry – I have been fuzzy headed lately: Connecticut ends with a single T and the semi-conductor mineral is Germanium. (if only we could harness flowers to do our calculations 🙂

        2. As I understand the physics, it might be easier to make a transistor than a vacuum tube–but it’s harder to make multiple transistors that perform identically than vacuum tubes that perform identically. Doping semiconductors is black-art chemistry. Forming grids and plates is a difficult machine-tool challenge, but once solved, operating the machines to create identical devices is mostly trivial.

          Without VLSI, many things become impossible, and many other things become epic projects. (Vacuum-tube computing comes to mind.)

          All that said, I need to do more research on kitchen-table transistor fabrication.

        3. Bruce C. Baker says:

          hand making vacuum tubes Part 1:

          hand making vacuum tubes Part 2:

  2. Tom R. says:

    Jeff, Have you seen this video on making a point contact transistor at home? Of course she has to start with a block of doped germanium, but it does show that it is possible.

    Technology builds on prior technology and the steam engine’s necessary precursor was the technology used to bore cannon barrels. I still think an external combustion engine would be easier to build with limited technology and in those you really can use anything that burns as fuel. Still diesel punk sounds like an interesting concept! Good luck with it.

    1. Alex Dillard says:

      I think to make the above link go to the correct video you may have to rewrite it in the following way:

  3. Lee Hart says:

    I agree that Diesels are *harder* to make than spark-ignited engines.

    Great video links to making vacuum tubes!

    Point contact diodes (and even some point contact transistors) were made in the 1930’s, but they were too temperamental to be practical. A master craftsman might make one that worked for a while; but they couldn’t be duplicated.

    We made silicon diodes in college, but started out with a silicon wafer already made and sliced and polished for us. The amount of technology needed to make diodes and transistors that way is formidable. I think that traditional silicon or germanium transistors are just to complex for “steampunk” level technology.

    However, there may be alternate approaches that *could* have been done. There are simpler ways to make inferior transistors that we ignore because we have better tech. For instance, thin film FETs are fairly easy to fabricate. They have really poor characteristics, but are good enough for low speed digital switching. They are in fact used in some LCD displays.

    And, I wouldn’t underestimate vacuum tube technology. We gave up on it because silicon became “fashionable”. If the transistor hadn’t been invented, we might well have seen the same micromachining technologies used to produce microminiature tubes, and even vacuum tube “ICs”.

    Even relay technology could have been used. Reed relays can be as small as ICs, and can implement all the basic building blocks (AND and OR gates, flip flops, counters, etc.) They are even lower tech than vacuum tubes, and can be made arbitrarily small.

    We often start with multiple technologies being used to solve a particular problem. But as soon as one of them gains a slight lead, we tend to put all our R&D into that one, and abandon all the others. The other technologies stagnate and die, even if they held promise for further advances. We end up with a technological monoculture.

    1. GE proposed “fat” Compactron-like tubes circa 1960 that would have all five active devices required by an AA5 superhet AM receiver inside a wide, low glass envelope shaped something like an 832A. (And we all know what an 832A looks like, no?) That would include certain standard passive devices as well, though which ones were not made clear in the writeups I’ve seen.

      They never actually did it, and I’m not sure that would be a clear win. If the filament burned out you lose all five active devices at once. This kind of “vacuum IC” would actually make a lot more sense for computers. Something smaller than a Compactron could contain an entire flip-flop or logic gate–or, with some extra care, several.

      There actually was a sort of gas-tube decimal counter that “remembered” a count applied by pulses to an element. They were called “Dekatrons,” and were wicked cool.

      Making transistors at home sounds easy (at least with some methods like diffused alloy or possibly meltback) until you consider the difficulty of getting suitably doped Germanium or tiny dots of pure Indium. Tube fabrication, by contrast, is more of a mechanical challenge and uses less in the line of exotic materials.

      1. Lee Hart says:

        Compactrons were a pretty crude attempt at integration. You might say they were just “more of the same”.

        Nuvistors are a better example. Instead of laboriously forming each little piece of metal and spot welding them together, nuvistors etched the structure out of a few larger pieces. First, it allowed really tiny tubes. Second, it allowed mass production similar to what we do with ICs.

        The best modern example of where this leads are vacuum fluorescent displays. They are flat glass panels, with planar microetched plates and grids, and a single filament for the entire package. Each segment or pixel is actually a tiny little triode!

        With this technology, you *could* build vacuum tube computers of reasonable size and performance.

  4. […] was wrong about Diesel engines being easy to make, as I suggested in my entry for March 5, 2012. Fuel injection, as it turns out, is a bitch. You’re trying to divide oil into a multitude of […]

Leave a Reply

Your email address will not be published. Required fields are marked *