Skip the Tubes: On to the Circuits!

So, I've heard of this POD quite a few times before. Or rather, this idea. From what I've read around here, vacuum tubes are considered a wasteful sidetrack on technological innovation, and that if we could skip over them and straight to integrated circuits things would've been fine and dandy. (Actually, the second part has never been debated upon, a side point for this thread.)

So I ask AH.commers to debate this subject in full:
Plausibility Check: Is it possible to skip the development of vacuum tubes and to the development of some types of circuits?
What sort of advancements would have to be made before a circuit could be successfully created?
Who would be the best early 20th century scientist to design such a thing?
Or would a no-name cardboard cut-out character suffice better?
What would be the estimated period in which the first circuit would be finished?
What about mass-produced?
How would this affect the world as a whole?
 
The problem with this is that ICs probably were impractical early in the century. They take seriously advanced ...everything... industrial process, chemistry, power, physics, and fine etching.

Computer tech's sadly hard to advance much, timewise, because it has huge dependencies on both several strokes of major genius and increasingly small-scale and advanced parts availability. That's why Babbage's Victorian Analytic Computer was a failure in our timeline; sophisticated automation like Hero of Alexandria's theatre effects machine and the Jacquard Loom were about the limit then.

I've been working on a timeline that would get computing noticeably early, but I do that by having a point of departure two and a half millenia ago, and it has its problems, too.
 
The early development of the transistor (forget IC's, until you have developed transistors they are not going to happen) is constrained in OTL by the Physics theories that let to an understanding of how they could work.

Transistors in the mid-30's is possible, given a flash of insight, but tubes are available then and aret going to go away just like that...

Transistors earlier will require a cludged POD, in which someone accidentally stumbles on the property of semiconductors by accident, AND realises enough of whats going on to build a working transistor. Not impossoble, but it will have to be a deus ex machina to some extent.

I'm also not sure at what point in time it was possible to make the actual semiconductor materials in acceptable purity for zone refining to work - that may well have been a stumbling block too.
 
Crystalline semiconductors (whether natural, like galena, or artificial ones, like silicon carbide) were actually pretty well-known at the turn of the century. The crystal radio (very popular in the early days of wireless transmission) uses a galena (lead sulphide) crystal as a diode.

What you need is an early and decently reliable way of doping SCs. The timepoint needs to be shifted backwards roughly two decades, around the period people had started experimenting with SiC semiconductors (historically, doping research had started on the eve of the SWW).

Although with proper design and careful SC selection, doping could be theoretically worked around.
 
Crystalline semiconductors (whether natural, like galena, or artificial ones, like silicon carbide) were actually pretty well-known at the turn of the century. The crystal radio (very popular in the early days of wireless transmission) uses a galena (lead sulphide) crystal as a diode.

What you need is an early and decently reliable way of doping SCs. The timepoint needs to be shifted backwards roughly two decades, around the period people had started experimenting with SiC semiconductors (historically, doping research had started on the eve of the SWW).

Although with proper design and careful SC selection, doping could be theoretically worked around.

There's a world of difference between semiconductor diodes and semiconductors transistors. For one, semiconductor diodes can be naturally occurring. This doesn't happen with transistors. Plus, the chemical processes and precision lithography techniques necessary for reliable production of semiconductor electronics weren't around really until after WWII.
 
One of the basic processes in semiconductor manufacture is the deposition of pure silicon layers by decomposition of silane (silicon tetrafluoride), which can each be doped selectively to produce patterns of semiconducting areas. Fluorine chemistry was virtually unknown (on account of the highly toxic chemicals produced in an age when a routine part of describing a new chemical was its taste) until the needs of the Manhattan Project required it to be developed.

The control and measurement requirements for processing semiconductors and many of the materials that go into it demand an already well-developed electronics technology to succeed. This was provided by the earlier development of vacuum tube technology.
 
Last edited:
Alright, so first up for order is the transistor.
What do we need for it to be built? Both in terms of scientific theories and elements needed to build a transistor?
 
vacuum tubes are considered a wasteful sidetrack on technological innovation
Tubes are not a wasteful sidetrack.
In Scientific Experiments where 1605 v 1606 Megahertz really matters, The Scientists go with vacuum tubes. With the best precision tubes being built in Poland.
 
Alright, so first up for order is the transistor.
What do we need for it to be built? Both in terms of scientific theories and elements needed to build a transistor?

The scientific theories probably arent in place until the 30's.

Its possible to envisage an earlier experiment finding the result of a P-N junction earlier, by accident (this gives you the semiconductor diode, the precursor to the transistor).
Now lets assume that whosever found this to work did a few experiments with P-N-P or N-P-N junctions, and got something to work (ie found by experimental accident rather than theory).

You still need a breakthroughs. Zone refining (to get you sufficient purity of materials), for example. You also need some years at least to be able to produce them in any sort of reasonable quantity.

Also, the vacuum tube isnt going to go away - its STILL in use today for high power applications! In the interwar period, power output is still a developing problem.

I could see the transistor being developed during the 30's, though, in time for WW2. Thats around 10-15 years in advance, technical advances better than this tend to fall apart due to the lack of a supporting tech base (or at least progress much more slowly)
 
Alright, so first up for order is the transistor.
What do we need for it to be built? Both in terms of scientific theories and elements needed to build a transistor?

MOSFET transistor theory was developed by Lilienfeld in 1926.
The problem was that it was impossible to build it because:
1) it was not possible to have material pure enough and cristalline enough (silane fractional distillation, but also Chocralski growth)
2) it was not possible to grow/deposit an insulating film thin enough.

trouble 1) was overcome after the war, but trouble 2) was not.
developing of the BJT transistor was due to an experimental error: Bardain-Brattain-Shockley broke down the unreliable dielectric and discovered that the device worked anyway (1947).
(problem 2) was overcome around 1959)

The main problem is that the equivalent device using vacuum tubes (a vacuum diode with a polarized grate acting as a switch) was well known in the 1800.

Since chemistry and manufacturing was not advanced enough until post WW2 to have the material, I think that semiconductor-transistors-before-tubes would be a bit ASB, I am afraid, since we would have to postulate the possibility of having samples of pure semiconductors available in nature and/or through rudimental refinment.
This could in principle be possible by re-working minimum energy of the cristalline lattices, but that would be a pre-creation POD :))) and would probably make organic life impossible to develop

On the other hand, relay based circuits could have developed.
mind you, these would not be integrated (discrete components), would be slower and quite unreliable, too
 
Last edited:
MOSFET transistor theory was developed by Lilienfeld in 1926.
The problem was that it was impossible to build it because:
1) it was not possible to have material pure enough and cristalline enough (silane fractional distillation, but also Chocralski growth)
2) it was not possible to grow/deposit an insulating film thin enough.

trouble 1) was overcome after the war, but trouble 2) was not.
developing of the BJT transistor was due to an experimental error: Bardain-Brattain-Shockley broke down the unreliable dielectric and discovered that the device worked anyway (1947).
(problem 2) was overcome around 1959)

The main problem is that the equivalent device using vacuum tubes (a vacuum diode with a polarized grate acting as a switch) was well known in the 1800.

Since chemistry and manufacturing was not advanced enough until post WW2 to have the material, I think that semiconductor-transistors-before-tubes would be a bit ASB, I am afraid, since we would have to postulate the possibility of having samples of pure semiconductors available in nature and/or through rudimental refinment.
This could in principle be possible by re-working minimum energy of the cristalline lattices, but that would be a pre-creation POD :))) and would probably make organic life impossible to develop

On the other hand, relay based circuits could have developed.
mind you, these would not be integrated (discrete components), would be slower and quite unreliable, too

From memory, wasnt the purity issue solved for Germanium at least by zone refining (I agree, silicon is too far a step, but that didnt come in until germanium transistors were well established)?
So we could have the experimental error happen a decade or so earlier. However I cant remember if zone refining was before or after the initial experiments....
 
From memory, wasnt the purity issue solved for Germanium at least by zone refining (I agree, silicon is too far a step, but that didnt come in until germanium transistors were well established)?
So we could have the experimental error happen a decade or so earlier. However I cant remember if zone refining was before or after the initial experiments....

The experimental error DID happen on a germanium device.

Germanium MOS transistors, however, were not suitable for mass production for 2 reasons.
1) Ge has a lower bandgap, thus Ge device do not work well at high temperatures (and currents dissipate heat, thus enhancing device temperature).
This could be fine for laboratory-test but not for mass market device
2) Ge has an horrible interface towards the dielectric, thus trapping charge at the interface, which shifts the MOSFET threshold voltage rather unpredictably.
Thus nominally equal devices have actually different characteristics.
This also, could be fine for laboratory-test but not for mass market device.

BJT Ge transistor were possible since problem 2 was not an issue, and degradation 1) was accepted as the price to pay.

Also, when going below mm-scale, we have to develop some lithography technique, too
 
Top