AHC and WI: Earlier Transistors

Here's the idea -- point contact transitors by 1920, preferably no PoDs before 1880. How does it happen and what are the effects on consumer products, et el?

Understand I am a complete layman...


... though yes, I am aware of the field effect transistor in 1925 that didn't get attention

John Fredrick Parker said:

Here's the idea -- point contact transitors by 1920, preferably no PoDs before 1880. How does it happen and what are the effects on consumer products, et el?

Understand I am a complete layman...


... though yes, I am aware of the field effect transistor in 1925 that didn't get attention

Click to expand...

Purity of the material is a huge problem. Understanding the quantum physics involved is another.

People used crystals (as in crystal radios) for years before anyone could figure out why they worked, let alone try to repeat the effect artificially.

Semiconductor tech is actually non-trivial.

OK -- suppose the field effect transistor (FCT) is invented a few years earlier and gets noticed; can it be followed by the revolution in electronics that followed the PCT in 1947?

a quick note from a man doodling with them:
the mosfet theory was understood by 1913, but before the czochralski process for putification an the hidration process to saturate unsaturate si-sio2 bonds, it is not possible to do anything worthwile.
bjt is a bit mose easier, but before ww2 is unthinkable.
having integrated devices earlier would not really make tech progress (or electronic revolution) be faster, since lithography issues would make them not much smaller than discrete devices.
moore laws is all about scaling, but if you cannot scale down significantly respect to discrete devices, it does not work

mailinutile2 said:

the mosfet theory was understood by 1913, but before the czochralski process for putification an the hidration process to saturate unsaturate si-sio2 bonds, it is not possible to do anything worthwile.

Click to expand...

Wait, wasn't that first one developed OTL in 1916? FWIG, it was another case of a happy accident, so I'm wondering if the discovery can be moved up a few years now, myself.

bjt is a bit mose easier, but before ww2 is unthinkable.

Click to expand...

Alright, I'm accepting this (though, for the hell of it, could you explain some? I really am a layman at this )

having integrated devices earlier would not really make tech progress (or electronic revolution) be faster, since lithography issues would make them not much smaller than discrete devices.
moore laws is all about scaling, but if you cannot scale down significantly respect to discrete devices, it does not work

Click to expand...

I think I'm getting why scaling down is necessary, but I'm not sure I understand why scaling down in the 1910's and 20's is implausible? (Remember, PoDs can be as early as 1880...)

John Fredrick Parker said:

Wait, wasn't that first one developed OTL in 1916?

Click to expand...

the process in itself was theorized, but level of cristallinity was not enough.
anyway, maybe just with c-z process, bjt could have a chance (you don't need hydration of interface states), but not mosfet
first mosfet was made in 1949 (and it worked really bad, that's the reason for bjt being used until the '70)

John Fredrick Parker said:

Alright, I'm accepting this (though, for the hell of it, could you explain some? I really am a layman at this )

I think I'm getting why scaling down is necessary, but I'm not sure I understand why scaling down in the 1910's and 20's is implausible? (Remember, PoDs can be as early as 1880...)

Click to expand...

because you already have devices that perform the same functions (valves). the good of going solid-state is having a device that act as a valve, but is smaller.
the more you progress, the more you can make the darn thingies smaller, by means of refining a process called photolitography.
using pre-ww1 lens, optics and chemistry, you'd obtain solid state devices roughly big as valves, and a bit worse in performance. thus there would not a push to use them (an to develop the manufactury process.)

Also, take into consideration that electronic development need a market, and you have not it in early '900.

mailinutile2 said:

anyway, maybe just with c-z process, bjt could have a chance (you don't need hydration of interface states), but not mosfet

Click to expand...

BJT has a chance? Cautious optimism

Also, take into consideration that electronic development need a market, and you have not it in early '900.

Click to expand...

Oh, don't worry about that -- I'm actually thinking the 1920's for this to take off, and I'm thinking incorporating this in a very different TL anyway

the more you progress, the more you can make the darn thingies smaller, by means of refining a process called photolitography.
using pre-ww1 lens, optics and chemistry, you'd obtain solid state devices roughly big as valves, and a bit worse in performance. thus there would not a push to use them (an to develop the manufactury process.)

Click to expand...

So photolitography is the key problem? Is there anyway you can think of to bring optics and chemistry up to speed, with a PoD after 1880 (or 1870 if need be) just enough to create transistors smaller than valves and better (even slightly) in performance?

Any more thoughts?

Remember those two earlier threads? The one about "steampunk" nuclear reactors and the one about an earlier discovery of nuclear fission?

Remember how it was explained that both those advances didn't rely on a single turning point? Or even a small number of turning points?

Remember how it was explained that both those advances relied on a host of theories, tools, techniques, and technologies? And how all those theories, tools, techniques, and technologies had a myriad of other applications beyond nuclear fission and nuclear reactors? And how all those other applications would quickly change a time line which achieved them earlier out of all recognition?

It's the same with earlier transistors.

Dathi and mailinutile2 have already explained it all in their posts, explained it correctly, succinctly, and in terms accessible to a lay person. An earlier transistor which would be recognized as a "no-brainer" replacement for valves would require advances in very fundamental theories, tools, techniques, and technologies and those fundamental theories, tools, techniques, and technologies would effect far more many things than just transistors.

John Fredrick Parker said:

So photolitography is the key problem?

Click to expand...

From the tech point of view I think that the key problem is cristallinity and purity of material.

From the motivational/economical point of view I think that the key problem is that valves and relais were already there and there was still space for their development, thus did only little sense to abandon them for good and turn to a totally different way to make switches (that's what transistor fundamentally are).
photolithography (and thus solid-state devices and their scaling) only kicks in when you cannot make significative progress on what already is there (or at least when this progress is severely slowed down).

mailinutile2 said:

<snip>

Click to expand...

Thanks for all the help

mailinutile2 said:

...
because you already have devices that perform the same functions (valves). the good of going solid-state is having a device that act as a valve, but is smaller....

Click to expand...

Aren't there other benefits of solid state though? Suppose you had large transistors, no lighter than an equivalent vacuum tube. Would they require more power to operate, or less? If less, then you have less waste heat in the circuit--you can pack them in tighter. I believe a transistor that weighs as much as a comparable vacuum tube would be much more compact. Less power required would also be a direct benefit of course.

Vacuum tubes needed to "heat up" to start working; a transistor device of the same circuit plan would turn on much more quickly.

The most important benefit I can think of is that vacuum tubes were fairly easy to break; they tended to fail for various reasons. I think a transistor-based equivalent circuit would be much more physically robust.

So, none of this makes a big difference for say a home radio set--but might make a big one for radios in automobiles, or more importantly earlier on, airplanes.

Even if the early transistors cost a lot more than the equivalent tube, if they could prove themselves longer-lived and rugged enough, I'd think the Army Air Corps and Naval aviation would prove a substantial if limited niche market initially--as would airlines if the military was not in a position to classify the technology.

In the 1920s and early '30s the US military budget was extremely low compared to WWII and postwar standards of course. And in the USA there wouldn't be much support for the idea of top secret technologies being held close by the government.

The British military establishment had its own severe constraints and issues to contend with.

Still, I'd think that even if transistors were little more than an expensive lab curiosity as late as the mid-1930s, as WWII loomed all these establishments would take another look at them.

Unfortunately, if the basic technology were published and patented before 1939, doubtless the Germans would be in the forefront of developing it, and after Hitler took over doubtless it would be both funded and kept largely secret. The Anglo-Americans would be playing catch-up.

Bipolar Junction Transistors are the most likely candidate for early invention/discovery.

Thank you Shevek ; I particularly like getting a clear idea of the results of early introduction, given its limits; besides radios in automobiles and planes, do any other implications come to mind?

Shevek23 said:

So, none of this makes a big difference for say a home radio set--but might make a big one for radios in automobiles, or more importantly earlier on, airplanes.

Even if the early transistors cost a lot more than the equivalent tube, if they could prove themselves longer-lived and rugged enough, I'd think the Army Air Corps and Naval aviation would prove a substantial if limited niche market initially--as would airlines if the military was not in a position to classify the technology.

In the 1920s and early '30s the US military budget was extremely low compared to WWII and postwar standards of course. And in the USA there wouldn't be much support for the idea of top secret technologies being held close by the government.

Click to expand...

John Fredrick Parker said:

Thank you Shevek ; I particularly like getting a clear idea of the results of early introduction, given its limits; besides radios in automobiles and planes, do any other implications come to mind?

Click to expand...

Mind you we still need to hear from someone who knows their transistors. I think they are less power-hungry than tubes, but the opposite might be true for really early primitive ones. They would almost certainly not be cheap--they are now due to mass production and the fact that we have microminiaturized them, but initially they'd use exotic substances at the very limits of the state of the art to produce.

Early transistors were made based on germanium rather than silicon; the latter is much more common but apparently harder to make work until you have a big investment in purifying it. Germanium on the other hand is rare I believe, and no doubt even it required very sophisticated handling to turn out even a rather crude transistor.

I just find it hard to believe that anyone says the only benefit of solid-state electronics is the potential of mass-produced printed circuitry. That certainly wasn't what the market was producing in the early days! Initially, transistors and diodes and the like were just substituted into circuits to replace tubes; the elements were still wired together in the same way.

I figured that was the kind of technology you were talking about.

Without solid-state devices there can be no microcomputers, true. It doesn't follow that solid-state=microcomputers.

They are less power hungry than tubes, almost from the beginning. Unlike vacuum tubes, they don't have to heat up to operate- that fact alone means they use less power than tubes.

Really, you'd only need to push earlier semiconductor diode technology- it is from the combinations of duodiodes and so on that we came to understand transistors (BJTs are just NPN or PNP layers and JFETs are just a layering of NPN or PNP materials in a certain manner- all properties that can be found by experimenting with diodes.)

Diodes are most useful in power supplies and early transistors will be most useful for their amplification- these functions don't require a lot of the later technology to be useful.

EDIT: I'm a layman compared to anyone with an engineering or history of the sciences background, I'll admit.

Shevek23 said:

I just find it hard to believe that anyone says the only benefit of solid-state electronics is the potential of mass-produced printed circuitry. That certainly wasn't what the market was producing in the early days! Initially, transistors and diodes and the like were just substituted into circuits to replace tubes; the elements were still wired together in the same way.

I figured that was the kind of technology you were talking about.

Without solid-state devices there can be no microcomputers, true. It doesn't follow that solid-state=microcomputers.

Click to expand...

Well that may bring me to the next question -- how much does this kind of stuff (mass produced printed circutry, microcomputers, etc) get moved up if we have solid state by 1920? I imagine a lot of other things still need to happen before you can even conceive of these things...

I do want to stress that I feel 1920 is very early, and that the earlier some form of transistor is actually made, the cruder it will be and the slower the progress (relative to the rate OTL after a comparable invention had been demonstrated) must be. That is, they are ahead of us OTL, but they don't forge ahead as fast until the general tech level has a chance to catch up.

Printed circuits and true microelectronics would come earlier than OTL but not as much earlier as the transistor itself. Exactly how much earlier would depend on the details of circuit fabrication. Some technology would be accelerated by the pressure of demand for a new and profitable industry, other aspects would just have to wait until the general state of the art advances under various constraints.

Hm, so fair to say use of transistors this much earlier would actually have minimal effects on computer history? As paradoxical as that is, it actually makes sense

What technologies, then, are affected in a major way? You mentioned airplanes -- what about automobiles? Is it plausible the electric starter is also delayed, leading to more development of electric vehicles?

To push along developments in semiconductor components, push distilled water and have a chemist suggest they use it when building their components. A problem well into the 60s was the presence of calcium ions from- get this- tap water being used by the fabrication companies.