Hey Tech, Im trying to learn how to make my own computer chips from cpu to ram...

OP here is a decent book to learn how computers work from a hardware perspective, which is what you want when you're trying to make your own hardware components. I'm sure you can find the PDF on library genesis.

amazon.com/Elements-Computing-Systems-Building-Principles/dp/0262640686
youtube.com/watch?v=uZsiMG7N_B8
coursera.org/learn/build-a-computer

FPGA is the place to be for this level of tinkering.

"sequential logic" will be in your search terms a lot.

I was also thinking this.

Yeah, you might want to start with a simpler goal. How about learning how flip-flops and half adders first?

Came to post this.

I never said i'm not going to learn binary, My interest is in ternary so i have to learn binary first.
I see the future of computing in ternary and if i can build my own ternary computer than I know for certain that there is no backdoor to my own hardware.

try learning how electronics and digital logic work first

Why isn't ternary used OP? Why do you think you have a magic bullet?

Ternay was not used because the Russians invented it, the americans also created a ternary computer but binary won because of politics. ternary is less costly and uses less power than a binary machine.
the russians eventually replaced their ternary computer with a binary computer that was more costly than the ternary one because binary already won out.

Binary = 2^8bits = 1Byte = 256
Ternary = 3^9.5trits = 1Tryte = 34901

Ternary is not only faster, it holds more information and is more efficient.

I was hoping you won't go to the Russians meme
Sadly, ternary is shit. Looks like the technology is unfeasible.
If you want more information per *it just stand with your finger in your ass until you get qbits and join the long list of people who failed to make ternary happen

And I'll share a personal anecdote with you - I want to implement a Lisp Machine. At least on a FPGA.
I thought it could be more efficient or faster at running Lisp.
Then I did some reading. Looks like it won't be better for running lisp.
However, I still want to do it. It looks like a fun project and an opportunity to learn new things. I'm under no illusion, however, that It'll be a technological breakthrough or that I figured something out that NOBODY IN INDUSTRY AND RESEARCH has figured out.

So I wish you best of luck with your project, but please, have no illusions.
Let me know if you want some material on electronics or on digital circuits, I can upload some books later.

The project is personal not for commercial. if it turns out good then i'll give it out.

I've been trying to convert this for hours, could anyone help me out?

I hope i got the math right.

I'm trying to see what the equivalent 28.8k modem would be if you used trinary digits instead of binary.

You have any electrical engineering experience?

Chances are, you want to make ram because a CPU would need the instruction set, be comparable with a mobo, etc. Unless you want to make a CPU that only runs the handful of instructions you give it, which is fine too. Browsing Holla Forums on your home-made potato will be difficult though, not to mention expensive.

How would a Teneray Computer work? Is there a third state you can have on any particular node?

Seems supremely redundant unless you can have more than 2 states.

...

a ternary computer works like a binary computer but instead of 2 states it has 3.

-1,0,+1
or 0,1,2

binary only has 0,1

SQL uses Ternary.

A quick search this shit has been kicked around for over 50 years but really hasn't made it big.

Seems cost and interoperability are the main issues.

It probably won't cost too much to simple IC in ternary these days but who is going to the investment to build the 4Ghz ternary CPU?

"I never trust anyone who is more excited about success than they are about the hard work required to get there"

OP is a faggot.

...

Moores law is hitting a wall, people are going to invest if they want to or not.

Check out MIT's 6.004. I'm not sure if they have online shit posted.

Alternatively, nand2tetris should have a decent amount of material.

People are going to invest but not in ternary because it's shit
There are billions of $$$ being thrown at new physical technologies, ranging from nanotubes to vacuum tubes
And again, 8>3, so why not just go to quantum computers? because you fell for a fucking meme

because quantum computers are fucking useless unless you are trying to do advanced physics and calculations for real time traffic and weather data.

the only one who fell for a meme is you, You actually think someone is actually going to sell you a nobody a quantum computer?

As soon as scale of economies happen, they will sell me a quantum computer.

Replace "quantum computer" with "personal computer" and you sound like a smug Remington Rand engineer from 1960.

Modern transistor designs are almost at an endpoint, so many new designs and technologies are being explored
All your posts were painfully ignorant of the technologies and subject you're talking about. Why don't you go read a book about semiconductor physics and then read a book about digital electronics design and CPU architecture before you touch the keyboard again?
Also, when quantum computers become viable for industries and big companies, I'm working in one of the first ten industries who'd probably take advantage of them
my offer of uploading some books for you still stands if you're willing to learn instead of being an ass

Imma assume for a moment you are serious about learning any of this sh*t....

as far as making your own chips, I dunno any realistic way to do that. sorry.

the first home PCs built (during the 1970's) were assembled from ICs (mostly) from the 7000 series of chips. Search "wiki 7400 series" to read about that. pretty much all of those chips can still be bought, and none of them cost much.

the 7400 series is 5v logic, otherwise known as TTL.
The CMOS version is the 4000-series, that runs on 3.3v. Search "wiki 4000 series" for info on those chips (you can still buy pretty much all them too)

there is lots of lower-end FPGA and CPLD boards (programmable logic arrays) available very cheap from China, but there's usually no documentation and the toolchains can be difficult to get working. Plus (to learn much) it helps to have another CPU feeding them organized data anyway...
As I've heard it, probably the easiest FPGA board to set up and running is the Mojo board ($80 US) or its China clone ($40),,, just because there is gud instructions in English for it. And it already has an ATMEL chip on board too.

Don't kid yourself, quantum computers are going to be black listed and only sold to corporations and government and military only. no one is going to be walking around with quantum computers strapped to their wrists.

but a personal computer is vastly different from a quantum computer. Im just telling the truth, you cant have people with the ability to unlock any encryption just walking around buying a quantum computer for 50$

I never expected to create nanoscale chips, I want to start out like pic related.

and FPGA would be my best option if i ever wanted the flexibility and size.

to

Firstly, quantum computing won't break all encryption, just certain kinds of asymmetric encryption. For example, AES is in no danger as far as anyone knows.

Secondly, as quantum computing matures, people will find more use for it. The people selling IBM 360s couldn't conceive of a use for personal computers back in the day. If you asked one of them what a household could use a computer for, they would probably scratch their heads before saying "I suppose a woman could store her recipes on one".

Thirdly, even if quantum computing is used only by governments now, it is very likely​ that it will eventually come to the civilian market. It's happened with most major technologies before, and I don't see why it won't happen again. Hell, there are already civilian companies trying to build quantum computers today. Once they are perfected, it will only be a matter of time before they are used more widely.

I learned a lot about hardware from the later chapters of projectoberon.com/ . Also, I don't know VHDL but my first idea for a Lisp machine would be to port github.com/hanshuebner/secd to Verilog or Lola2 (which I do know a little).

mojo board is shit and has shit documentation.

altera cyclone 2 is $20 and does everything you need as a beginner.

amazon.com/RioRand-EP2C5T144-Altera-Cyclone-Development/dp/B00LEMKR92

Nice thinking there guy
That's actually kind of what I intend to do.
I debated with myself after finding the second link if I should first implement the SECD architecture in a more abstract way before porting it to HDL or just writing detailed architecture docs.
I like the idea of implementing one from scratch but I might eventually go with implementing am already simulated one.
Thanks for the first link though

I made a binary double adder.

this man has created something legendary this is so cool

calling this a computer is a huge stretch.
most of the time the homemade 4 and 8 bit "computers" you find on interweb are nothing more but ALUs, which are just a fraction of a CPU.

been there OP, and i'm not the reading kind of guy i get border easily.
if you're like me i would recommend to get your hand in the dirt and start building.
there are many calculators, z80 and 6502 based computer diagrams out there.
as you go through the building process you'll learn how they work, like it it.
i find it easier to learn that way than reading some abstract shit on a book

This is a lot more than just an ALU. Did you even see the modules?

you can have electricity and call that the 1
you can not have electricity and call that the 0
now tell me how you would get a 3rd state out of this and make a feasible computer.

draining electricity?

he could make it voltage based and have as many states as he wants.
gonna be expensive and hard as fuck to pull off but double on paper.

registers.

As the flamer from previous posts allow me to offer a non butt-fucking-retarded implementation - use optic waveguides instead of electrical ones, transmit three distinct wavelengths. Use the photoelectric effect to create three distinct voltage levels at the receiver.
Now use that arbitrary component to implement ternary algebra (I better remember this, it would make a good question in job interviews for EE).
Implement NOT, AND, OR, NEG, MAX, MIN, and maybe a full adder.
Prove that your full adder works.
Show some chops if you want to be taken seriously
faggots

im not being very seriously. i hardly ever know what im talking about and the only reason i browse Holla Forums is because im a paranoid schizo and you guys seem to know the best way to keep me from being watched.

i'dd like to read some material about non binary computers, seems like a giant pain in the ass to program.

That's not how it works.
Binary is not a natural system for reality.

Electricity naturally has 3 states, Neutral, Positive and Negative.

for a ternary computer you need a power supply can can separate these 3.

Negative for - or -1
Neutral for 0 or Null / both
Positive for + or +1

You could use +5v or -5v

You could even have -4V, -3V,... 0V,...1V, 2V, 3V, 4V,... etc for whatever system you would design. Designing that system to be efficient, accurate and moderately easy-to-program is another matter in itself.

But then again, simplicity is what all the geniuses praise and I can't say whether this would give enough merit to counter the complexity. If anyone has any *recent* whitepapers to link I would be interested.

Binary only has 16 universal Logic gates

While Ternary has 19683 universal logic gates

I'm having hard time finding much of anything on ternary.

youtube.com/watch?v=BXJR8TIzrho
youtube.com/watch?v=TfJxAb0owj8
youtube.com/watch?v=pBSf2bLwAoQ

...

twistedoakstudios.com/blog/Post7878_exploring-universal-ternary-gates
here's a good site too.

So you're saying ternary is the Perl 6 of logic gates?

The ternary system seems interesting, but the logic tables must be pretty fucking big compared to binary.

I don't know anything about perl 6 code sorry but maybe, I do know that SQL for example uses ternary logic.

yes but you actually don't even need to bother with most of them. I guess if you are looking for efficiency you might look for better combinations.

homepage.divms.uiowa.edu/~jones/ternary/arith.shtml

Now see, this is a feasible computer.

fascinating stuff, thanks

What do you mean by "universal logic gate"
as opposed to any logical combinatory circuit?

oh
Are you talking about electric charge, or voltage?
because I'll remind you that voltage (what's required to move electrons about) is always defined relatively to something.
The way digital circuits operate is by having two states:
So in terms of representing states, there is no natural preference for binary or ternary, that's only user bias.
In terms of KISS, binary wins

allaboutcircuits.com/technical-articles/universal-logic-gates/

Electricity works by splitting Electrons from Atoms, Atoms are Positive Neutral and Negative.

Proton is Positive , Electron is Negative, Neutron is Neutral

Electricity is generated when an electron is moved to another atom, this creates a void or an atom with too much electrons where another electron takes its place. This creates current and generates voltage.

Dude its on breadboards, this thing won't even do a Mhz.

Its a learning tool and a toy. Get off your high horse.

This thread made me remember something. Why are media like CDs, DVDs and blu-rays binary? If they introduced more patterns that can be burnt and read from the discs, storage would increase extremely and the only downside would be having an additional processing step to translate things from and to binary which today's tech is irrelevant.

which with today's tech*

thats still not a computer and shouldn't be called that.
frequency has nothing to do with it, computers can work on sub Hz.

electrical current can flow through neutral mediums
Look, I can't be sure if you're retarded or pretending anymore.
I quit

If you cant understand basic physics than I'm not going to force you. but don't insult me and think you know better.

I did rather well at physics in university, thank you
and in electronics
and semiconductor devices
end electromagnetics
Are you sure you want to go there?
The reason I'm being so pedantic about it is BECAUSE I studied it and rather liked it. So when you say stuff like "electricity is generated" it makes me want to euthanise people who use that phrase as a eugenics program.
You can create a voltage difference, or induce current, but your entire representation of the physics of the problem is lacking.
Having any charge concentration in any system is temporary and very localized, and the entire system will always stay neutral.
I have to ask, what's your background?

I got a PhP in internet trolling.

so you're getting butthurt over semantics?

First thing you start with is an ISA. The easiest to get into is MIPS. Start with r-type and i-type instructions, from integer math. Design an ALU, and a controller. Start small with a simple integer calculator, then work on adding a register, then pipe lining, then add additions for other instructions.

I would also recommend using a HDL like verilog, and a simulator, rather then spending on an FPGA

I hope you like field effect saturation voodoo.