>the OP has some nerve making a technology post on a technology board >another feminism in technology thread, that's what we need instead
Eli Jones
^This.
Easton Davis
Actually the 6502 had 4528 transistors (3500 enhancement and 1018 depletion load) so that would be more like 11 breadboards.
Connor Gonzalez
>The 6502 has 3500 transistors. You could recreate the entire thing using eight breadboards. Wrong. There's internal connections on those breadboards, they're not as flexible as putting those fets wherever you want.
Asher Hill
The incredible thing is that the whole schematic was traced on paper and they cut the chip die by hand. There was one mistake, in that the ROR instruction was busted on early units (fixed by the second half of 1976).
Jason Bell
im thinking one or two haslel CPUs could feed afrika with a lot of bread
Thomas Harris
what about wiring/bus? that takes a few points as well
Juan Ward
That's the connectors on the outside, right?
Hudson Johnson
thats just power/ground lanes. you will need extra connections for signals
Dominic Anderson
breadcubes when?
Thomas Ramirez
If you had a full sized 830 pin breadboard, it would take only five of them.
Logan Peterson
>idk what i'm talking about >i think it makes me look smart
Joshua Wood
You can't bump your own thread, OP.
James Torres
>How does that make you feel
It makes me feel like you don't understand how breadboards work, and didn't even expend the mental effort to consider how they could possibly work.
Jonathan Garcia
id like to see a modern processor laid out like that is there a youtuber that would do it? seems impractical but doable for something like a 486
Makes me feel modern CPUs are still heavily overpriced.
Jaxon Turner
Not at all because you're already talking over 1 million transistors. There's a _big_ difference between a 6502 and a 486 in terms of how complex the chip is.
Justin Walker
See
Benjamin Martinez
>You could recreate the entire thing using eight breadboards. You could make all of the connections. But it wouldn't work unless you ran it really really really slow. And then another interesting question is how much power would the discrete components use compared to the IC (well it is interesting to me).
>You could recreate the entire thing using eight breadboards. No, you couldn't.
Kayden Reyes
>id like to see a modern processor laid out like that It would be pointless, a lot of parts of the CPU have extremely tight timing tolerances that you'd have to take into account when wiring it onto the breadboard. You will fuck up everything.
>depletion load Forgive me for being an ignoramus on chip fabrication, but what does that refer to exactly?
Luke Morris
Depletion load means all oxidizers have been removed from the transistors. Enhancement means oxidizer has been added to the transistors, generally by spraying them with a water mist.
Nicholas Peterson
AFAIK a large portion of that came from how the 486 integrated the FPU onto the main chip die for the first time.
Liam Evans
MOS initially came out with a pin-compatible clone of the 6800, the 6501, but this led to them being sued by Motorola so they had to change the pinout.
Owen Allen
>manufactured in the 45th week of 1985 >a MOS 6502 and not a second source Probably taken from a 1541 drive, right?
Jack Lewis
How do you figure?
Jose Gomez
>tfw no breadboard qt3.14
Grayson Miller
The 1985 manufacture date. Straight (not customized) 6502s would have been used in Apple IIs, VIC-20s, PETs, BBC Micros, Atari 8-bits, and Commodore disk drives. Third parties didn't buy directly from MOS and their output went exclusively into Commodore's hardware. So it was most likely taken out of a 1541, although it's also possible that it was from one of those late PETs with the molded plastic case (the 8296).
Daniel Long
>400 point breadboard. The 6502 has 3500 transistors. You could recreate the entire thing using eight breadboards.
what do you think breadboard contact points and internal routing actually do you retard? you would in actuality need hundreds of these breadboards to do what you're saying.
Ryan Nguyen
Not true though. 6502s have been used in a million different embedded applications as well. That chip in the OP pic could have come from anywhere.
Nathan Anderson
Cool CPU, Ahmed.
Elijah Ward
That's not how those boards work. The one in your picture has 60 distinct connection points in addition to the power and ground planes. A transistor has three legs, so 3500x3/60 = 175 boards. Now go and do your math again.
John Edwards
>130 million boards for the Haswell Amazing.
Aiden Perry
I have the exact breadboard
Isaac Thompson
You could make a row of them all the way to the Moon.
Aaron Clark
A 6502 from discrete transistors has been made monster6502.com/ runs at 60kHz as opposed to ~1MHz of the original
Lincoln Phillips
Not really surprising to be honest. More cores will not help the fact that Moore's law stopped working in 2004ish. t. Took a CPU design course
Carson Jenkins
I work at Texas Instruments and it does take a lot of very huge, complicated, and expensive equipment to manufacture ICs. A lot can also go wrong during the manufacturing process, although Q/C is usually pretty good at removing dud chips from the line.
Owen Gomez
>You could make all of the connections. But it wouldn't work unless you ran it really really really slow
The interesting part is how Apple integrated several system functions on the Apple IIe into several custom ICs, yet during testing they found that the CPU couldn't drive the system board as fast as the II+ and its discreet logic, so they had to increase the clock speed slightly.
Noah Morales
We are also hitting the ceiling of DRAM since you can only make chip dies so dense.
Wyatt Reyes
Long circuit path.
Brandon Mitchell
Duh?
Evan Foster
You could do it in theory but it would run really slow and use a lot of power.
Kevin Ortiz
Your point? You sound like a teenage EE student.
David Myers
...
Michael Evans
>The 68000, as its name implies, has 68,000 transistors, so 170 breadboards. This was actually a coincidence and no, it did not have exactly 68000 transistors. It's name implies it's a upgrade from the 6800.
Ok so is probably 70,000 and they rounded it for marketing purposes.
Eli Wood
>posts a Wikipedia page that lacks citation >inb4 *source* from a Usenet thread 22 years ago by random neckbeards (every other document uses Wikipedia as source) Rumors != Facts, kids
>It's really unlikely the 680000 has only 68,000 transistors. More likely the name came as an upgrade of the good old Motorola "6800" series. Although there's almost no resemblance between the two architectures. They are in fact source compatible.
Kayden Morris
Probably. I'm sure they did make a big deal out of the transistor count to show how much more powerful it was than the 8086 which had under 30,000 transistors.
Andrew Young
>How does that make you feel? that it would be easier if one would use neckbreadboard instead
Sebastian Brown
noice user just be sure to not post it on Jow Forums they hate things like this
Robert Gonzalez
At the start of the 80s, everyone was fapping to the 68000 and Z8000 and the 8086 was kind of seen as a joke. IBM essentially saved it from an early demise.
Samuel Morales
Well x86 only because a think thanks to heavy marketing by Intel, IBM and Microsoft. It's like the gang of dumb guys, they have no redeeming qualities except raw marketing power.
x86 was still a joke until the late 90s.
Jason Walker
IBM didn't have a choice anyway; the other 16-bit CPUs were too expensive and no support components were yet available for them in 1981.
Ironic how the Z8000 ended up being the joke there.
Jaxson Peterson
got first job today I'm so buying bunch of these and building my own 16 bit CPU after the first paycheck
Ian Green
Some workstations used it and also military applications, but it never reached the consumer market.
Aiden Howard
nice webm, source?
Colton Ortiz
Shortly after the 6502 was introduced, MOS second-sourced it to Rockwell and Synertec. Atari did buy some chips directly from MOS, but it seems like Synertec 6502s were more common.
Second sourcing is a standard industry practice unless it's a custom ASIC designed for one application because one manufacturer alone can't supply all of the demand for a particular IC.
>I can't find the reference now, but I'm pretty sure that Frank Durda mentioned this on the comp.sys.tandy newsgroup. A lot of the Model 3s have cheap knock-off Z80s. That was the exact reason--you could never tell if you would get Zilog or the cheap ones because Tandy would stick in whichever was cheaper at the time. I remember he said that the knock-offs didn't have the alternate register set; you'd get garbage if you activated them. Durda said that when customers complained, Tandy insisted that it was not "standard" to use them! Apparently the IC fab couldn't reverse-engineer that feature correctly. (And I remember an interview with the designer of the Z80 who mentioned he deliberately put sneaky traps in the design exactly to catch those reverse-engineers!)
Colton Russell
>Apparently the IC fab couldn't reverse-engineer that feature correctly
I guess too much for whatever Taiwanese fab was doing this?
Hudson Rodriguez
I agree. CPUs are one of the most failsafe ICs there is and failures are really rare. It's usually RAM and GPUs that fail the most often.
Jose Martinez
I think FCC would have an aneurysm.
Grayson Ward
Why is that?
James Roberts
GPUs tend to get hot and DRAM is relatively fragile due to the density of the chip die.
Isaiah Gray
Apple became shit when x86 became good
Blake Lee
For example, a 4116 DRAM was 16kx1, meaning it has 16,384 bits.
If , then 819 breadboards would be needed.
Ryan Thomas
Actually no. DRAM cells consist of a capacitor+transistor pair, so the capacitor has two legs and the transistor three. Thus in a 16kx1 chip, you'd really need 32,768 distinct elements, one cap+one transistor for each physical bit. But the cap has one less leg than the transistor, so figure it.
>16,384 transistors with three legs each >16,384 capacitors with two legs each
So 546 boards.
Chase Wright
Two thirds, so 1092.
Hudson Fisher
SRAM however is just flip-flops, so a 512kx4 SRAM like the 2114s Commodore used would have four elements of 512 bits each, so 2048 total bits per chip. So 102 boards.
Aiden Long
...
Jack Richardson
Basically, it refers to whether the transistor is conducting "by default", with zero gate-to-source voltage.
Aiden Ward
>seems impractical but doable for something like a 486 It would also be enormously expensive. Discrete transistors may be cheap, but when you need a million of them, it's not so cheap anymore. I don't think anyone was expecting it to run at full clock frequency.
Owen Richardson
>Discrete transistors may be cheap, but when you need a million of them Companies get them cheap because they buy in bulk purchases and get special price discounts.
Dominic Cook
That was a stupid lawsuit anyway. MOS made the chip pin compatible so they could use 6800 testing equipment, but the chip otherwise had nothing else in common with the 6800.
Charles Ross
Yes, but when you need a million of them, it's still very expensive for a personal project, even if you get the best of corporate bulk deals.
Joseph Stewart
Unless you expose a CPU to ESD, it's really hard to kill them.
Christopher Cooper
>Z80 from a machine that was only lightly used crapped out It's rare but it could happen if a dud slipped through the Q/C tests at the factory.
Camden Diaz
Also some manufacturers won't sell direct to you if you're not a company rep. There was one case where a guy needed a 13 pin D-sub connector for an Apple II project, and it was a struggle to finally find a chink manufacturer that would produce the connectors he wanted. Most wouldn't deal with you unless you're working for IBM or something like that.
