> what's so different in that coming from goo chemistry instead of electric bits?
Very insightful comment reply. If you don't mind, I'll respond in an analogy: What's so different about just building a bigger rocket to move more stuff into space at one time? A: You can't because you reach a finite potential in the relationship between fuel weight and lift. Extrapolating this analogy to the context of your comment reply, sure, in a theoretical sandbox environment that could happen with the silicon circuits and transistors we have today, however I find that unlikely.
Silicon based circuits and the current model and understanding we have of machines and the way they should be built, from a basic principals level is wasteful, thermally inefficient, and requires a sum of resources that if used to model the human brain would surpass our ability to house, power, construct, and allocate the precursor materials for. Given our current technology and the parameters involved, it's just not going to happen for us.
Maybe tomorrow something will get invented that will even further push moore's law into the dirt, but I just don't see that happening in this generation nor lifetime.
> Is it possible to have a cube of 1nm transistors. Like a 100 million ^ 3. The cube would have size of 10cm w.d.h.
Probably, but I suspect operating it would cause it to melt or explode :-) We could probably manufacturer 3D chips today (or more 3D than they are today), but from what I understand cooling them starts to become a major issue. (Not involved in chip design or manufacturing)
> The number of components in rockets and IC transistors are probably roughly comparable.
Unless you are counting the payload, I really doubt it.
ICs are only possible because people create some way to mass manufacture those features, and people don't even design the features by themselves.
If people created them at the traditional route of "manually design|manufacture this -> encapsulate into component -> use components to manually design|manufacture that -> encapsulate again...", it would be impossible to create chips as complex as we have now. But there still aren't better way to produce rockets.
All the problem with rockets is that they are created by joining a big number of things that don't really like to stay together (or, in a few cases, don't like to stay by themselves either). I doubt they have more bare complexity than a modern car.
Quote: "Here’s the list of the underlying technologies:
(1) Electromechanical;
(2) Vacuum Tube;
(3) Relay;
(4) Transistors; and
(5) Integrated Circuits."
Surely not in that order? Because Relay = Electromechanical. The "Relay" item is certainly out of place. In the history of computing devices, relays (and purely mechanical computing devices like the Difference Engine) predated vacuum tubes, which predated transistors, which predated integrated circuits.
> I studied microelectronics. I am aware of the technical challenges. Can you explain those challenges that are primarily non-technical?
We spent far more time buying EDA software, installing it, talking to foundries, getting the PDKs, signing NDAs, dealing with buggy EDA software, dealing with slow EDA response times, etc. than actually working on our chip.
>Things like Silicon-on-insulator, high-k dielectrics, finfets, extreme ultraviolet lithography are not innovative or new ideas?
I'm not saying they aren't, but I have noticed that the general level of openness, and following that, innovation and open-mindedness has dropped dramatically in the past decade or so, and I do have to say that the general semi industry has stayed generally innovative, and much of my criticism is directed towards the rest of the industry primarily. That being said, there is a major glacial pace.
Example of a real conversation I had with an engineer at one of the major (can't name the exact one) foundries about a device that's actually pretty close to reality:
Me: "Why don't you use this X device?"
Him: "Because it's still research"
Me: "Sure, but it's very promising, why aren't there at least any industrial research efforts to commercialize it?"
Him: "Because it's still research"
Me: -__-
SOI is innovative, but it's been held back by cost and the self-heating effect, both things that really aren't that much of a problem.
FinFETs were launched by a DARPA initiative.
High-k dielectrics I will say are the single most interesting (if not innovative) innovation in the last decade in the semi industry, although I have some bias there.
EUV is a feat to engineering no doubt, but again, my grievances aren't really focused in that area.
> Semiconductor engineering is complex, but probably not so complex that your average Math Olympiad couldn't pick it up in a month or two.
We have a chip shortage going on for months now, if all it took was two months and a bunch of smart people, those billions of dollars in chip orders would've made it happen. That tells me it has to be a bit more complex than you think.
For context: A 4090 has 76 billion transistors, and the B200 superchip that Nvidia announced has around 200 billion transistors, on an absolutely enormous package of 2 dies glued together ala the M2 Ultra. So this would suggest a minimum density increase of 5 times in the next 6 years. Moore's Law isn't dead yet folks.
> I wonder if you can make micro machines at this level? The MEMS thing.
At this size range, though state-of-the-art MEMS (mechanical vibrating frequency filters for RF receivers in phones, accelerometers) can have sub-100nm dimensions, basic accelerometers, pressure sensors, and inkjet heads are absolutely doable.
> Not with this PDK or process, no. MEMS processes are quite specialised.
But yeah, this is the problem. Although ICs and MEMS devices are made with similar tools, MEMS usually needs processing steps that don't play nicely with the steps in an IC process (e.g., etching away huge amounts of silicon to leave gaps and topography, or using processing temperatures and materials that mess up ICs). This SkyWater process cannot do MEMS.
A more general problem is that different MEMS devices often need different incompatible process steps, so a standardized process is infeasible (though http://memscap.com/products/mumps/polymumps tries).
However, there is a tiny chance that, if we get enough detail on the process steps and leeway in the design rules, a custom layout could implement a rudimentary accelerometer or something that works after post-processing (say, a dangerous HF bath), but only with intimate knowledge of said process steps (e.g., internal material stress levels) and a lot of luck.
> Sorry but that isn't a replacement for Silicon, not by a long shot. You are not getting 14nm plastic transistors anytime soon.
You missed the point. The point isn't to replace 14nm transistors.
The point is to allow engineers to do VLSI design like PCB design.
There are lots of interesting things that can even be done with 5um or 10um transistors, but we can't get there because the non-recurring expense is too high.
Here is an example: guitar pedals--specifically the analog delay ones. These pedals all use an ancient MN3205 bucket brigade CCD chip. The chip is dead simple to make, but since the volume is too low to offset the NRE, nobody is willing to make it.
If, however, you could design that chip with an NRE of $1,000 instead of $100,000, you could make a tidy profit. In addition, you would probably pull all of the other functions of the pedal into the chip as well.
There are also other nice benefits to using older and larger transistors. One of them is voltage tolerance. Modern transistors can't take voltages at even 3.3V in many cases, while the old 10um transistors could go to 18V (old school 4000 series CMOS was specified from 3V-18V for most chips).
The point isn't to replace silicon transistors. It's to create a completely separate market at a different volume point.
> I never understood what made vacuum tubes good transistors.
They generally aren't. Vacuum as-in the medium may be technically superior, but vacuum tubes as practically implemented weren't just displaced because solid state was easier to manufacture, but also because solid state quickly outperformed valves in many areas. (There are areas where valves held out for much longer and are still used today in some cases, e.g. transmitter output stages).
That being said a great thing about valves is that they're very hard to kill. They can take huge momentary overloads (a bit like magnetics, but with much shorter time constants), which would literally detonate similarly specced solid state output stages. And because they can run much hotter than any known solid state tech they can have huge energy densities and can be cooled very efficiently (boiling water cooling): https://upload.wikimedia.org/wikipedia/commons/8/80/Rs2041_s... (Siemens RS2041V, 600 kW output tube, diameter of the copper anode dissipating about 240 kW is 239 mm, was still made in 1999)
Quote: “In the low-power regime, the way these power converters work, it’s not based on a continuous flow of energy,” Paidimarri says. “It’s based on these packets of energy. You have these switches, and an inductor, and a capacitor in the power converter, and you basically turn on and off these switches.”
To me that's a failed attempt to create a layperson-accessible explanation. In my NASA Space Shuttle work in decades past I also created efficient power supplies and I used similar methods -- but I think I can explain it more effectively.
The trick to making a modern power supply efficient is to control the phase relationship between voltage and current. This is normally performed in reactive elements like an inductor, a capacitor, or both.
Put another way, instead of changing a voltage level with a resistance (which would waste power), you need only manipulate a reactive element so the phase angle between voltage and current is something other than zero degrees. For example, at a phase angle of 90 degrees, you can have substantial voltage as well as current, but no power dissipation (except where it's needed).
That's the secret to power supply efficiency -- change the voltage without using any power-dissipating elements.
Very insightful comment reply. If you don't mind, I'll respond in an analogy: What's so different about just building a bigger rocket to move more stuff into space at one time? A: You can't because you reach a finite potential in the relationship between fuel weight and lift. Extrapolating this analogy to the context of your comment reply, sure, in a theoretical sandbox environment that could happen with the silicon circuits and transistors we have today, however I find that unlikely.
Silicon based circuits and the current model and understanding we have of machines and the way they should be built, from a basic principals level is wasteful, thermally inefficient, and requires a sum of resources that if used to model the human brain would surpass our ability to house, power, construct, and allocate the precursor materials for. Given our current technology and the parameters involved, it's just not going to happen for us.
Maybe tomorrow something will get invented that will even further push moore's law into the dirt, but I just don't see that happening in this generation nor lifetime.
reply