Not micro anymore lol. It's a "PCB" version of a processor that was originally sold as a microchip. A PCB processor. PCB processors were very common in the late 60s, they are more advanced than the Walls of AA battery-sized transistors used on the TX-2 a shopping center sized computer from the 50s which was capable of full 3D graphics and it had it's own super compact Nuclear Power Plant which produced it's required 8 Megawatts of power! It used INSANE amounts of energy, literally the energy of a small town. But was already more powerful than a Super Nintendo, in fact the TX-2's graphical processing power is similar to the SNES with SA-01 and Super FX chips operating simultaneously! It was a BEAST back then. After all I am talking about a 50s computer. A 50s computer capable of 3D graphics is just amazing really. CAD was literally born on the TX-2 which makes it one of the most historically important computers. Even more amazing is the fact that the TX-2 was produced by hand, without etching techniques... It was extremely unefficient but it was an amazing construction nonetheless. Nowadays most of it was destroyed with only some small parts preserved in museums. You can see how big was a section of it's Core Memory and the transistros which were as large as AA batteries... It was really huge, really low clock machine but with an immense parallel processing capability for the time.
As a computer studies student, this thing is a work of beauty, an art piece and a memorial for an obsolete but historical and influential piece of technology. Thank you EMS for such good quality RUclips content.
imagine if he built the rest of a commodore 64 and had a working commodore 64 with that thing as the cpu in it🤣🤣🤣🤣🤣 I am guessing the gpu would also then be a work of art just like the monster 6502 is🤣🤣🤣🤣🤣
I loved the way you guys shared the voice over equally. This truly demonstrates equality in partnership and further strengthens the idea that you both had equal input into the project. Top class work.
Yet their voices compete against that unnecessary background music, whose tone and volume are distracting. The video content and narrative are great but why so many videos need such filler music when it's not needed?
Altium tip: Enable display of net name on all of your power (ground) ports, even if you expect them all to be the same. An identical graphical symbol doesn't guarantee that they're the same net. It's also not possible to catch that during review of a printed schematic. Ask me how I learned this. =) Also- it helps to keep your DRC warnings (green nets) under check. I once worked with an engineering tech that didn't realize that green meant DRC warning... Every pad on a 12 layer board was green.
@@RenX3133 Not really. We are talking about very old techniques. In fact, he's done the upscaling to PCB level from the chip. Nothing more. He did an astounding job by routing the board and getting the digital circuit done properly, but that actually is repetitive basic stuff. No offense, I really enjoy the work.
Unfortunately, I can't make it to the faire. But I would love to see a longer RUclips video of somebody demonstrating and talking about the Monster6502!
I *need* one just for visuals! It would be fascinating seeing how the LED's change according to what is running, so programming it would be hard to resist!
3000 transistors. That's so cute! Look at you! PS - Apparently the early T-800 Terminators were running 6502 code, because when the first Terminator movie came out, there was a scene where we're given a Terminator's eye view of its sensory input and decision making process, and I could have sworn it was 6502 assembler code off to the side.
Maybe there were more than one of those scenes and in one they used assembly, but I very clearly remember that the scene I saw had COBOL, because when I saw it I had just started training for a new job learning COBOL and it blew my mind.
I researched it snd my hypothesis was right, they used both: ruclips.net/video/YRnnjoiSV-U/видео.html At the beginning of my research I found only images and clips with assembly and I started to worry I became victim to the Mandela effect
"The VideoGuy" sent me and recommended this link. We think alike... I see I was here 3 years ago :) It never gets old. Wonderful project. Especially for the ones that lived the 6502 dream. But don't tell my GF. She thinks old people lived back then :oO Since this video is 2017 and it's currently 2021 (ugh) I'll forge ahead into time and see what wonders it holds. Cheers.
This monster 6502 has one of the features that computers back in the 1950s and 1960s had. This transistor scale cpu has all these indicating lights to show what it going on in the various registers. Old computers also had rows of indicating lights to indicate what was going on in the various registers.
@@evil-mad-scientist - Hows progress? - this is an awesome project, but I would swap some of the LEDs with LCD displays that show the register values & stack etc. LEDs look cool, but its all flickering too quick to understand. - Even though its down-clocked a 1/16th speed would be great to really be able to follow the clock in human time - or a debugger. - in-fact - if it had direct ASM entry & debugging - wow - it would have to be provided to every school child & it would raise computer understanding by orders of magnitude.
@@mumblic I offered to make a 20 ft tall water transistor "pocket" calculator for COSI of Columbus Ohio in 2009 and pay for it. They refused the offer. All clear plexiglass with colored water paths too. Not sure they knew what I was talking about. Oh well. I was going to use two reservoirs on a chain hoist so when a tank emptied, you could alternate the tanks to "recharge" the system. Gravity would power it. But being entirely mechanical would show people that the logic was what was important, not the electricity.
@@theodiscusgaming3909 you buy acrylic square rod, cut into blocks, drill and tap holes and use acrylic rods to push through cavities to block or allow flow. Gates.
I hope this comes to completion some day, I had convinced the heads of the CS department at one of my local universities to order one should it become available. Aside from just being a neat showpiece, the proposal was to make use of the 6502 instruction set for the CS courses that deal with these low level details, and actually allow students and instructors to run their programs on it so they can see what's actually happening
Love this retro stuff :) Good job guys, this is spot on :) Heh, my dreams were filled with this goodness when I was younger :) I very much love all the love you have put into it :) Amazing!
hi,me too but with a zilog z80 ,i used to have an apple 2 ,comodore 64 ,zilog z80 with msx computer .this proyect is great ruclips.net/video/nFIviiwPrLI/видео.html
Ugh... I just got the cold sweats when you showed that Apple II and it reminded me of grade 9 computer science class and BASIC. Those green monochrome screens were so soul sucking that I wanted to just walk out the door before it was too late.
Short answer: yes. Long answer: The clock rate itself cannot go to zero; the dynamic nmos architecture sets a lower limit on the clock speed (about 20 Hz in our case). However, there is a well-known single-step circuit for the 6502 that does allow you to advance one _instruction_ at a time, even while the clock continues running.
It seems like a good (better designed?) alternative to the www.megaprocessor.com/ Mapping the memory to a little oled or other display might be an neat thing for the future. Working out equivalent scaling for some period dram dies and displaying them with a data overlay at the same scale as the monster 6502 would be awesome.
go to the MegaProcessor Project ruclips.net/video/lNa9bQRPMB8/видео.html there you can slow down, halt and single step. every single gate has an LED indicator and even hex display to make the processing more understandable
I'd love to have this hung on my wall, along with a NES with the Monster 6502 replacing the original CPU. However, I do realize the complications this would entail. It's just an idea I had. I often take motherboards and hang them on my walls as art pieces. Seeing as the Monster 6502 already looks like a beautiful piece of kinetic art, it would only make sense to hang it up.
I would buy it as a kit... i love testing my SMD skills on new projects. Just think, selling it as a kit could even help fund your future projects. To those who find such a thing frustrating... think of complicated model airplanes and cars... now imagine if you could start that car or plane when you finished... that's truly awesome guys love it...keep it coming
The c64 was a 6510 processor which was a 6502 with extra I/O pins for bank switching (amongst other things between the basic rom and ram, that you could copy by writing the same data back to the same location!)
@@evil-mad-scientist I have your discrete 555 and 741 all done and shadowboxed on my office wall. Just need your GP100 discrete version to finish off the set.
Awesome I love seeing this kind of work, interesting that making the traces slows it down especially for some one as I, who isn't so well versed in electronics. Also a fan of Applied Science channel just saying what are the odds! Great video!!
Is the speed in part a function of the size? If so does that mean in essence that it could never run at the speeds required for, say, a C64 implementation?
Yes. All transistors suffer from propagation delay, and (basically) the larger they are, the slower they can operate. Miniaturization is essential for speed. Of course, big is more fun. 8)
@@Waccoon - I suspect that the max clock speed would ALSO be limited by the sheer length of the PCB traces too. When a signal has to travel through a meter or so of PCB trace there is an inherent delay of 5 to 10 nS!
@@trevorvanbremen4718 Yes, but as long as your traces are terminated correctly, trace length only becomes an issue once you get up to hundreds of MHz. I once heard of an insanely stupid 8-bit machine where the data bus lines were pumped through a printer cable to an external peripheral. Amazingly the whole system worked... as long as the cable wasn't too close to the display or power brick. 8)
@@Waccoon... Can you remember back to the Trash-80 Model 1 with expansion box behind it? (Showing my age here... Gulp) These had a PCB edge connector at both ends of a short ribbon cable routing the address and data bus between the two boxes. (16kB of 4116 DRAM in the main keyboard and 32kB of 4116 DRAM in the rear expansion box along with a floppy controller and Centronics parallel port). They pumped out RFI rather efficiently in the AM radio band (although that RFI came more from the non-shielded memory-addressed keyboard switch matrix than from the expansion box) If memory serves me right, there was NO actual bus termination between the boxes. Simply octal buffers (74LS244 springs to mind) Can you imagine trying to get FCC-15 certification on something like that today? (Although it was a LOT 'quieter' in the RF spectrum than my S100-based system when I had the lid off [almost ALWAYS off])
I love that there's giant solder pads, as if you were going to attach 4-guage gold wire to connect it to giant pins arrayed around a large conference table. It's to scale, right?
Yes, and those pads are all functional, and in the original order of the bond pads on the original die. Giant bond wires on a coffee table is something we have talked about. :D
ExcalibursZone For all intents the 6510 is just a 6502 with an extra IO port added, and a few other minor changes. Same thing as where we often say the SNES CPU is a "65C816", but that's also slightly modified. (and the Ricoh CPU on the NES etc. etc.) Nintendo in particular liked using slightly modified CPUs, perhaps to make it harder to clone those machines? Even the MIPS R4300i in the N64 is only slightly different to the generic NEC branded VR4300, as it has two pairs of pins swapped. Sneaky. lol Love this project btw. The 6502 is easily one of the cleanest CPU designs ever, especially the instruction set / mnemonics. It would be great to see similar macroprocessors made for other popular chips eventually. Or perhaps implement a discrete version of the NES PPU, so we can all watch Mario being drawn REALLLLY SLOWWWLY. :p
It's mainly an accuracy thing. While the chips are in the 6502 family (of which there is no dispute) you cannot put a stock 6502 in a C64. You need a 6510. The Vic 20 had a 6502, the 1541 had a 6502, and the NES had a 6502. It's like saying "The Ford's engine is the same thing as that Toyota's engine with a few minor changes" you simply can't just transplant one into the other. A little simplified of an analogy, but that's basically the same thing.
Toyota vs Ford is a rather poor analogy -- the 6510 is _extremely_ similar to the 6502 in terms of chip layout, usage, etc. It's more like the same engine with an extra exhaust pipe. If you look at the datasheet for the 6510 processor, right there on the first page it says "The internal processor architecture is identical to the MOS Technology 6502 to provide software compatibility." These chips were intentionally designed by the same people to be compatible with each other, despite their minor differences. We understand that _many_ people know the subtle differences between these devices. We glossed over certain details to explain the big picture of what our project is. We are not, for example, making a how-to video about which CPU device device models and speeds can be used in which motherboards. (Fun fact: The 6510 is _MUCH_ closer to the 6502 than a 6502 is to the Ricoh 2A03 in an NTSC NES. About 1/4 of the 2A03 die is essentially a transplanted, slightly modified 6502 core, and about 3/4 of it is not. That's quite different from the 6510, which is just a slightly modified 6502 core.)
I noticed this is an old ass video but have you considered donating this to the living computer museum? It would be an excellent exhibit for people to play around with.
Why not the new iPhone A13 CPU with 8.5 billion transistors... It would take 12 years non-stop for an industrial 78.000 CPH pick-and-place machine just to assemble the components 😎
I'm propably wrong here as it's just a wild guess. But I'm guessing it has to do with its size. It takes longer for the signals to travel on this version than on the smaller original one. It can also be that they might have used cheaper parts for this one.
AFAIK there are two main factors: size and fragmentation. Size means bigger elements, which means more current/more electrons. This means that we need more time to move all those electrons around each cycle. This CPU surely consumer much more power than the original chip. Fragmentation means that all the elements are not placed on a single chip, made from one bulk of silicon. There are many downsides to that: bigger distances between transistors' terminals, additional capacitance and inductance coming from wiring (which means bigger impedance and lower max frequency). Another downside is that transistors created on a single chip behave better working together, hence bigger speeds are achievable.
From the MOnSter 6502 website : "The primary limit to the clock speed is the gate capacitance of the MOSFETs that we are using, which is much larger than the capacitance of the MOSFETs on an original 6502 die. "
Guys, size itself cannot cause any limitation :) Current travels in conductor with such speed, that it would have to be size of thousands of kilometers to see any difference. It's considered to appear at the other end immediately at such scales, regardless. As someone noticed limitation occurs from properties of discreet elements used and frequencies they can work at.
over 20 cm it takes (1/1.5Giga)seconds for a signal to arrive, so you might run into problems in the ghz range, but I don't think that would play that big of a role at 4mhz.
May i? You'd need a few glass boxes to do it in. Instead of VFD's you could make micro sized triode and tetrode tube nodes inside almost the size of an SMD transistor. Instead of LEDs you can use fluorescent die directly on the lines you wish to illuminate. Capacitance would be a problem because of the high voltages needed to operate the vacuum valves, but embedding resistors and ceramic capacitors inside wouldn't be that hard. I can see it happen using vapor deposition and etching techniques like they use in the production of silicon devices these days. It would require quite the development process, but i think it's doable!
Howard Black Now that you mention those micro tubes, i vaguely recall seeing pictures of them - and yes, it would have been quite interesting to have been able to see to which degree the miniturisation of this technology could have gone.
if you want to make money, make a version of this you can hang on the wall as modern art with battery so it blinks when you turn it on. and maybe with pics and descriptions og the famous computers it was used in underneath! do the rest yourself
Yesss! I just love blinkenlights. If they make it easy for a noob just to run some program at one touch to have the lights blink, I'd love it on my wall in my office.
This brings smile to me :) , I had 6502 pirated Apple II back in the '80s and tried some machine language coding around it during that time for fun. Long Later I moved to MASM and TASM. What I nice memories.
Paul Frederick Indeed.... the culmination of life decisions has delivered a reality where most of my time is spent paying for commitments made long ago.
@@adisharr That is categorically untrue. _Some_ people _who were young in the late '70s/early '80s_ and were interested in programming experimented with 6502 assembly language. Most people with an interest in programming are either too dead, too old or too young to fall into that narrow window. The majority of people with an interest in programming have never used 6502 assembly language. That's not to say that it wasn't a popular processor, just that people have been interested in programming for many more years than the 6502 was relevant. I programmed in 6502 assembly language myself in the mid '80s before moving on to ARM, which _is_ relevant, and in a major way, today.
@@johnm2012 I should have rephrased that, most people that had a Commodore 64 as a kid and were interested in programming ended up using assembly on the 6502. We had clubs all over the place and everyone had their copy of MADS.
@@adisharr The BBC Micro was the big seller here at the time. It came with the excellent BBC BASIC with a built-in assembler which made it very easy to try out 6502 assembly language. All the OS entry points were documented so there was no need for PEEKING and POKING obscure memory locations and a whole 16 bytes of precious zero page was set aside for user vectors. It was a great platform to learn on. It was so good that I never bothered with more powerful processors, such as the Z80 or the 68000 but jumped straight from 6502 to ARM, which immediately seemed strikingly similar - like a 32-bit version of the 6502. Perhaps that's not surprising since it's instruction set was designed by Sophie Wilson, who knew the 6502 inside out, having written BBC BASIC and key parts of the OS, in assembly language, of course.
The MOnSter has approx 17.88 transistors per square inch. At that density, an i7-860 which has 774M transistors would be a PCB measuring 548 feet on each side...
6502 gets all the fame but most of the early work I did was Z80 based, mostly TRS-80. Wanted to work on the 8080 but they were too expensive at the time. A 4004 replica should be easy to do. I think someone did something similar for a PDP-8.
Very nice indeed!!!! I'm sure there'll be a market for it. Sign me up. I was actually thinking of making a 4-bit computer using simple diode logic (light-emitting for more fun), and only transistors when absolutely required. But my brain fried when I thought about it. Nice short story.
Not only very impressive, but also a work of art.
Nope
@Ali Cuntë it needs to be a monster 65816 now so they can make a monster apple 2gs
Now to implement a 6502 entirely in relays.
@The Creeper King But why not?
The Creeper King then what are we waiting for? ;)
Bill_Mccoy QUICK eat my potato
Nah, tubes would be better and less noisy. The only noisy part would be the diesel-powered generator needed to run the darn thing.
@The Creeper King I want this to happen
"Oh, yeah thats my CPU over there. The coal input is at the bottom and the starting lever is right there..."
... and that pendulum swinging back and forth... is the clock
ΗΑΗΑΗΑΗΑΗΗΑΑ
Like a real life "Minecraft" version of a enlarged microprocessor lol
@@CanadaBud23
Enlarged prostate.
its very powerfull. like 3 kilowatts
So you made a megamicroprocessor?
Yes, although we prefer the term "macroprocessor."
I would call it a "Micro-Macro-Processor" ;)
Not micro anymore lol. It's a "PCB" version of a processor that was originally sold as a microchip. A PCB processor. PCB processors were very common in the late 60s, they are more advanced than the Walls of AA battery-sized transistors used on the TX-2 a shopping center sized computer from the 50s which was capable of full 3D graphics and it had it's own super compact Nuclear Power Plant which produced it's required 8 Megawatts of power! It used INSANE amounts of energy, literally the energy of a small town. But was already more powerful than a Super Nintendo, in fact the TX-2's graphical processing power is similar to the SNES with SA-01 and Super FX chips operating simultaneously! It was a BEAST back then. After all I am talking about a 50s computer. A 50s computer capable of 3D graphics is just amazing really. CAD was literally born on the TX-2 which makes it one of the most historically important computers. Even more amazing is the fact that the TX-2 was produced by hand, without etching techniques... It was extremely unefficient but it was an amazing construction nonetheless. Nowadays most of it was destroyed with only some small parts preserved in museums. You can see how big was a section of it's Core Memory and the transistros which were as large as AA batteries... It was really huge, really low clock machine but with an immense parallel processing capability for the time.
to end the discussion, it would just be "processor"
humunguprocessor
Such an impressive project! I will see you at Maker Faire.
Awesome seeing you here. I watch your channel as well :)
Sounds like a threat.
@@boukinist3899 lol smh
Terminator!
:)
Oooh 555 likes... can't touch the magic timer number
This is what I pictured when I first heard, "I built my own computer". I was astounded then disappointed. You sir, have reastounded me.
As a computer studies student, this thing is a work of beauty, an art piece and a memorial for an obsolete but historical and influential piece of technology. Thank you EMS for such good quality RUclips content.
imagine if he built the rest of a commodore 64 and had a working commodore 64 with that thing as the cpu in it🤣🤣🤣🤣🤣
I am guessing the gpu would also then be a work of art just like the monster 6502 is🤣🤣🤣🤣🤣
I think this should be sold as a kit.
are you able to solder 8000+ SMD components? good luck with that
@@buddyryanmckendrick6601 Yes i am =-O
@@buddyryanmckendrick6601 yeah, whatever, as if that were the only way kits were sold. That's a bold strawman of yours.
@@buddyryanmckendrick6601 - I'm gonna wait until they release the kit using THT
alan sanchez do... do you want to suffer that bad?
This is absolutely stunning! What a beautiful piece of engineering.
I loved the way you guys shared the voice over equally. This truly demonstrates equality in partnership and further strengthens the idea that you both had equal input into the project. Top class work.
Yet their voices compete against that unnecessary background music, whose tone and volume are distracting. The video content and narrative are great but why so many videos need such filler music when it's not needed?
This brought tears to my eyes, such an awesome effort.
Altium tip: Enable display of net name on all of your power (ground) ports, even if you expect them all to be the same. An identical graphical symbol doesn't guarantee that they're the same net. It's also not possible to catch that during review of a printed schematic. Ask me how I learned this. =)
Also- it helps to keep your DRC warnings (green nets) under check. I once worked with an engineering tech that didn't realize that green meant DRC warning... Every pad on a 12 layer board was green.
Cool story bro... now can you translate that into English si vous plait ?
At your next electrical engineering interview, just hand this board over as your resume.
Depending on the job he wants to apply, this might actually work.
Would not work for a engineering job. But definitely a technician.
People who can do this probably dont have to interview... they're recruited.. :-)
@@woodiemarv It would. Much more impressive than any degree
@@RenX3133
Not really. We are talking about very old techniques. In fact, he's done the upscaling to PCB level from the chip. Nothing more.
He did an astounding job by routing the board and getting the digital circuit done properly, but that actually is repetitive basic stuff.
No offense, I really enjoy the work.
Unfortunately, I can't make it to the faire. But I would love to see a longer RUclips video of somebody demonstrating and talking about the Monster6502!
One of the greatest projects of all times 👍👍👍
SUPER ! Maybe 8086 is better right away ?
I feel some kind of nostalgia watching this
like I'm back in the 80S that I grew up with MS-DOS...
Ohhhh the Computer Memories
yeah ms-dos the last stable build of windows
Wow. I don't even know where to begin, this is beautiful on so many levels! Thanks for sharing!
The most valuable thing of that is the pleasant of LEDs
So cool.. I learned so much from programming that venerable little beauty. Fantastic homage to this milestone of computing.
I *need* one just for visuals! It would be fascinating seeing how the LED's change according to what is running, so programming it would be hard to resist!
Wow, I love it. Visually representing bits with LEDs is incredible, demystifying silicon just a little bit.
Imagine how expensive and large this would be if the transistors were replaced with vacuum tubes!
Well that's just silly. Oh hang on... Perhaps 'silly' is the whole point of the exercise?
i mean, isn't that where all this computer stuff came from originally anyway?
It would fill an average sized house and consume 100's of KW I shouldn't wonder
How have I only just come across, this is AWESOME!
3000 transistors. That's so cute! Look at you!
PS - Apparently the early T-800 Terminators were running 6502 code, because when the first Terminator movie came out, there was a scene where we're given a Terminator's eye view of its sensory input and decision making process, and I could have sworn it was 6502 assembler code off to the side.
Bender runs on a 6502 as well.
Maybe there were more than one of those scenes and in one they used assembly, but I very clearly remember that the scene I saw had COBOL, because when I saw it I had just started training for a new job learning COBOL and it blew my mind.
I researched it snd my hypothesis was right, they used both: ruclips.net/video/YRnnjoiSV-U/видео.html
At the beginning of my research I found only images and clips with assembly and I started to worry I became victim to the Mandela effect
Yeah, all they needed was the software.
"The VideoGuy" sent me and recommended this link. We think alike... I see I was here 3 years ago :) It never gets old. Wonderful project. Especially for the ones that lived the 6502 dream. But don't tell my GF. She thinks old people lived back then :oO Since this video is 2017 and it's currently 2021 (ugh) I'll forge ahead into time and see what wonders it holds. Cheers.
When's the Z80 version coming out? 😉
8500 active transistors? Good luck making few square meters PCB.
@Old LiquidThe project should be named TTL home heater.
@@adamw.8579 Your mistake is thinking in 2D... ;)
@@AttilaAsztalos Haha... huh.
This monster 6502 has one of the features that computers back in the 1950s and 1960s had. This transistor scale cpu has all these indicating lights to show what it going on in the various registers. Old computers also had rows of indicating lights to indicate what was going on in the various registers.
no "MOnSter 6502 Update #2" ...................
We sent out an update to the mailing list recently. We're (finally) close to release.
@@evil-mad-scientist very nice!
@@evil-mad-scientist Sweet. Can't wait to see it.
@@evil-mad-scientist well how bout posting that update to the channel for the rest of us peasant folk to view?
@@evil-mad-scientist - Hows progress? - this is an awesome project, but I would swap some of the LEDs with LCD displays that show the register values & stack etc. LEDs look cool, but its all flickering too quick to understand. - Even though its down-clocked a 1/16th speed would be great to really be able to follow the clock in human time - or a debugger. - in-fact - if it had direct ASM entry & debugging - wow - it would have to be provided to every school child & it would raise computer understanding by orders of magnitude.
Fantastic project, specially for teaching people how an Integrated Circuit actually works. Big respect right there.
I'm sorry... But WHY???
Wait... I don't care about why!!!! I love it!!!
Because more and more people especially the new generation has no idea how a CPU works.
It's a great educational tool.
@@mumblic I offered to make a 20 ft tall water transistor "pocket" calculator for COSI of Columbus Ohio in 2009 and pay for it. They refused the offer. All clear plexiglass with colored water paths too. Not sure they knew what I was talking about. Oh well. I was going to use two reservoirs on a chain hoist so when a tank emptied, you could alternate the tanks to "recharge" the system. Gravity would power it. But being entirely mechanical would show people that the logic was what was important, not the electricity.
@@anthonylosego wait, how do you make transistors with water?
@@theodiscusgaming3909 you buy acrylic square rod, cut into blocks, drill and tap holes and use acrylic rods to push through cavities to block or allow flow. Gates.
@@mumblic they don't wants to learn stupid processors ... All they want is fortnight & cyberpunk
I hope this comes to completion some day, I had convinced the heads of the CS department at one of my local universities to order one should it become available. Aside from just being a neat showpiece, the proposal was to make use of the 6502 instruction set for the CS courses that deal with these low level details, and actually allow students and instructors to run their programs on it so they can see what's actually happening
Amazing project, great job and thanks for the update!
wow very impressive!
So fu@k1n cool!!!
Are you guys going to kickstart this into production so we can buy one of their own?
Perhaps in the future. We're not at that stage yet.
There are at least two people who want one.
make that a 3
Thats what I've been wondering for quite some time.
Make it a 4!
Love this retro stuff :) Good job guys, this is spot on :) Heh, my dreams were filled with this goodness when I was younger :) I very much love all the love you have put into it :) Amazing!
Darned impressive ! I started out assembler programming on an Apple II way back when.
hi,me too but with a zilog z80 ,i used to have an apple 2 ,comodore 64 ,zilog z80 with msx computer .this proyect is great ruclips.net/video/nFIviiwPrLI/видео.html
I read that you were close to release a year ago. How’s update #2 coming??
Still close to a release; temporarily on hold due to COVID.
I want to see this made with through hole components
You Can, at: ruclips.net/video/FaIMl8MNEHs/видео.html
Ugh... I just got the cold sweats when you showed that Apple II and it reminded me of grade 9 computer science class and BASIC. Those green monochrome screens were so soul sucking that I wanted to just walk out the door before it was too late.
you guys are legends!
nice one :)
Very impressive work. Great as educational tool. Keep up a good work guys!
Some People Rule!, Others Dominate!!, You Guys took it to a Whole Nother, LEVEL!!!))
Fascinating tech. A thing of beauty and a joy for ever.
Can you single step or massively slow the clock?
Short answer: yes. Long answer: The clock rate itself cannot go to zero; the dynamic nmos architecture sets a lower limit on the clock speed (about 20 Hz in our case). However, there is a well-known single-step circuit for the 6502 that does allow you to advance one _instruction_ at a time, even while the clock continues running.
It seems like a good (better designed?) alternative to the www.megaprocessor.com/ Mapping the memory to a little oled or other display might be an neat thing for the future. Working out equivalent scaling for some period dram dies and displaying them with a data overlay at the same scale as the monster 6502 would be awesome.
go to the MegaProcessor Project
ruclips.net/video/lNa9bQRPMB8/видео.html
there you can slow down, halt and single step. every single gate has an LED indicator and even hex display to make the processing more understandable
What a monumental undertaking! Kudos
Quite a contrast to the Altoids ELF
Any Ben Eater fans here?
Me
Unsure on the definition of "fan"... but I did subscribe to his channel.
Yup
Yessir
FANTASTIC!!! brilliant job, and you should be VERY proud of yourselves. I've loved the 6502 in its day... Thank you!!! :)
I'd love to have this hung on my wall, along with a NES with the Monster 6502 replacing the original CPU. However, I do realize the complications this would entail. It's just an idea I had. I often take motherboards and hang them on my walls as art pieces. Seeing as the Monster 6502 already looks like a beautiful piece of kinetic art, it would only make sense to hang it up.
I would buy it as a kit... i love testing my SMD skills on new projects. Just think, selling it as a kit could even help fund your future projects. To those who find such a thing frustrating... think of complicated model airplanes and cars... now imagine if you could start that car or plane when you finished... that's truly awesome guys love it...keep it coming
Wow! Impressive... So the 6502 only had about 3500 transistors. I thought it had more, but there you see. Time goes by... *LOL*
Infinite loop now new ones have over 230 million
230 million? LOL The core i7 4790k has 1.4 BILLION.
The vega gpu has 12.5B transistors.
GPUs cheat a little in my eyes.. since its the same thing copy pasted many times lol
Although either way its INSANE
The c64 was a 6510 processor which was a 6502 with extra I/O pins for bank switching (amongst other things between the basic rom and ram, that you could copy by writing the same data back to the same location!)
Yes, exactly. The VIC-20 was the Commodore with the 6502 chip
Wow! I know you are busy, but could you roll an NVidea GP100 version?
We'll get right on that!
@@evil-mad-scientist I have your discrete 555 and 741 all done and shadowboxed on my office wall. Just need your GP100 discrete version to finish off the set.
will we be able to make this later looks very cool
I'd love a painting sized computer to hang on my wall
Even the 6502 is enjoying that RGB goodness.
Best quote of the century:
People: *"ARE YOU NUTS?!"*
EMS: Probably.
AFAIK the C64 had a 6510. It was the VIC20 that had the 6502.
Damn i love those flashing Leds :D
ruclips.net/video/12LLJFSBnS4/видео.html
Awesome I love seeing this kind of work, interesting that making the traces slows it down especially for some one as I, who isn't so well versed in electronics. Also a fan of Applied Science channel just saying what are the odds! Great video!!
Is it possible to make a version that is 10 times faster than the original one?
or even 100 times faster ?
Take the arduino lol
The 6502 is still in production and the new ones are a bit more than 10 times faster.
i am designing a tiny 8 bit CPU with transistors, this video made me so happy
Is the speed in part a function of the size? If so does that mean in essence that it could never run at the speeds required for, say, a C64 implementation?
Yes. All transistors suffer from propagation delay, and (basically) the larger they are, the slower they can operate. Miniaturization is essential for speed.
Of course, big is more fun. 8)
@@Waccoon - I suspect that the max clock speed would ALSO be limited by the sheer length of the PCB traces too. When a signal has to travel through a meter or so of PCB trace there is an inherent delay of 5 to 10 nS!
@@trevorvanbremen4718 Yes, but as long as your traces are terminated correctly, trace length only becomes an issue once you get up to hundreds of MHz. I once heard of an insanely stupid 8-bit machine where the data bus lines were pumped through a printer cable to an external peripheral. Amazingly the whole system worked... as long as the cable wasn't too close to the display or power brick. 8)
@@Waccoon... Can you remember back to the Trash-80 Model 1 with expansion box behind it? (Showing my age here... Gulp)
These had a PCB edge connector at both ends of a short ribbon cable routing the address and data bus between the two boxes. (16kB of 4116 DRAM in the main keyboard and 32kB of 4116 DRAM in the rear expansion box along with a floppy controller and Centronics parallel port).
They pumped out RFI rather efficiently in the AM radio band (although that RFI came more from the non-shielded memory-addressed keyboard switch matrix than from the expansion box)
If memory serves me right, there was NO actual bus termination between the boxes. Simply octal buffers (74LS244 springs to mind)
Can you imagine trying to get FCC-15 certification on something like that today? (Although it was a LOT 'quieter' in the RF spectrum than my S100-based system when I had the lid off [almost ALWAYS off])
This is just beautiful in every way!
Please re-create my entire brain using vacuum tubes so I can become an independent planet living off pure sunlight.
Version 2 video coming?
Very impressive! The Atari 2600 used a cheaper 6502 version btw. (the 6507 if I'm not mistaken)
I love that there's giant solder pads, as if you were going to attach 4-guage gold wire to connect it to giant pins arrayed around a large conference table. It's to scale, right?
Yes, and those pads are all functional, and in the original order of the bond pads on the original die. Giant bond wires on a coffee table is something we have talked about. :D
@@evil-mad-scientist That would be something to show off at the next VCF!
WIRE IT INTO A COMMODORE
but it uses a slightly different cpu called the 6510 it removed some pins to save costs
@@AnnoyedArt1256 i mean the c64
Ivan204 i was talking about the c64 already
Beautiful!
- A great project in PCB design, soldering and probably fault finding.
What's the cost of components? (All SMD I assume?)
- PCB pricey?
Interesting project. Though, the C64 did not run on a 6502, it was the 6510. A derivative, but not the same chip.
We did gloss over some details of that nature; a couple of the other systems that we named are also variations based on the same core.
ExcalibursZone
For all intents the 6510 is just a 6502 with an extra IO port added, and a few other minor changes.
Same thing as where we often say the SNES CPU is a "65C816", but that's also slightly modified.
(and the Ricoh CPU on the NES etc. etc.)
Nintendo in particular liked using slightly modified CPUs, perhaps to make it harder to clone those machines?
Even the MIPS R4300i in the N64 is only slightly different to the generic NEC branded VR4300, as it has two pairs of pins swapped. Sneaky. lol
Love this project btw.
The 6502 is easily one of the cleanest CPU designs ever, especially the instruction set / mnemonics.
It would be great to see similar macroprocessors made for other popular chips eventually.
Or perhaps implement a discrete version of the NES PPU, so we can all watch Mario being drawn REALLLLY SLOWWWLY. :p
It's mainly an accuracy thing. While the chips are in the 6502 family (of which there is no dispute) you cannot put a stock 6502 in a C64. You need a 6510. The Vic 20 had a 6502, the 1541 had a 6502, and the NES had a 6502. It's like saying "The Ford's engine is the same thing as that Toyota's engine with a few minor changes" you simply can't just transplant one into the other. A little simplified of an analogy, but that's basically the same thing.
ExcalibursZone
Yep, fair enough.
Usually it's me who's the pedantic one when it comes to the technical stuff, so I'm trying to break the habit. lol :p
Toyota vs Ford is a rather poor analogy -- the 6510 is _extremely_ similar to the 6502 in terms of chip layout, usage, etc. It's more like the same engine with an extra exhaust pipe.
If you look at the datasheet for the 6510 processor, right there on the first page it says "The internal processor architecture is identical to the MOS Technology 6502 to provide software compatibility." These chips were intentionally designed by the same people to be compatible with each other, despite their minor differences.
We understand that _many_ people know the subtle differences between these devices. We glossed over certain details to explain the big picture of what our project is. We are not, for example, making a how-to video about which CPU device device models and speeds can be used in which motherboards.
(Fun fact: The 6510 is _MUCH_ closer to the 6502 than a 6502 is to the Ricoh 2A03 in an NTSC NES. About 1/4 of the 2A03 die is essentially a transplanted, slightly modified 6502 core, and about 3/4 of it is not. That's quite different from the 6510, which is just a slightly modified 6502 core.)
I noticed this is an old ass video but have you considered donating this to the living computer museum? It would be an excellent exhibit for people to play around with.
my mind is blown sir you deserv the most respect for this
I want to buy this so bad, it looks soo cool!!!
Me, one year ago
come on do a 386 or 486!
Why not the new iPhone A13 CPU with 8.5 billion transistors... It would take 12 years non-stop for an industrial 78.000 CPH pick-and-place machine just to assemble the components 😎
@@StephanBuchin yea, but duke3d or Quake2 on something like that would be pretty cool right?
Good thing you have all the LEDs on there or I never would have believed it was doing anything. :) Seriously impressive work.
What is the limitation for speed?
I'm propably wrong here as it's just a wild guess. But I'm guessing it has to do with its size. It takes longer for the signals to travel on this version than on the smaller original one. It can also be that they might have used cheaper parts for this one.
AFAIK there are two main factors: size and fragmentation.
Size means bigger elements, which means more current/more electrons. This means that we need more time to move all those electrons around each cycle. This CPU surely consumer much more power than the original chip.
Fragmentation means that all the elements are not placed on a single chip, made from one bulk of silicon. There are many downsides to that: bigger distances between transistors' terminals, additional capacitance and inductance coming from wiring (which means bigger impedance and lower max frequency). Another downside is that transistors created on a single chip behave better working together, hence bigger speeds are achievable.
From the MOnSter 6502 website : "The primary limit to the clock speed is the gate capacitance of the MOSFETs that we are using, which is much larger than the capacitance of the MOSFETs on an original 6502 die. "
Guys, size itself cannot cause any limitation :) Current travels in conductor with such speed, that it would have to be size of thousands of kilometers to see any difference. It's considered to appear at the other end immediately at such scales, regardless. As someone noticed limitation occurs from properties of discreet elements used and frequencies they can work at.
over 20 cm it takes (1/1.5Giga)seconds for a signal to arrive, so you might run into problems in the ghz range, but I don't think that would play that big of a role at 4mhz.
Wild guys, completely wild !!!! Great project !
Next step should be electron tubes ;)
I'm intrigued, please tell us more about how you see this implemented.
May i? You'd need a few glass boxes to do it in. Instead of VFD's you could make micro sized triode and tetrode tube nodes inside almost the size of an SMD transistor. Instead of LEDs you can use fluorescent die directly on the lines you wish to illuminate. Capacitance would be a problem because of the high voltages needed to operate the vacuum valves, but embedding resistors and ceramic capacitors inside wouldn't be that hard. I can see it happen using vapor deposition and etching techniques like they use in the production of silicon devices these days. It would require quite the development process, but i think it's doable!
Howard Black
Now that you mention those micro tubes, i vaguely recall seeing pictures of them - and yes, it would have been quite interesting to have been able to see to which degree the miniturisation of this technology could have gone.
Seeing this makes me happy.
if you want to make money, make a version of this you can hang on the wall as modern art with battery so it blinks when you turn it on.
and maybe with pics and descriptions og the famous computers it was used in underneath!
do the rest yourself
Yesss! I just love blinkenlights. If they make it easy for a noob just to run some program at one touch to have the lights blink, I'd love it on my wall in my office.
This brings smile to me :) , I had 6502 pirated Apple II back in the '80s and tried some machine language coding around it during that time for fun. Long Later I moved to MASM and TASM. What I nice memories.
Stunned.....this is outrageously cool. Wishing I had even a tiny fraction of the time that has gone into this project.
You have just as much time as anyone else does. You simply choose to spend your time differently.
Paul Frederick Indeed.... the culmination of life decisions has delivered a reality where most of my time is spent paying for commitments made long ago.
like children ?
Any update on the new version?
I was writing assembly code at 15 yo for this processor in late 1979.
Most people interested in programmed experimented with Assembly on the 6502 at a young age.
@@adisharr That is categorically untrue. _Some_ people _who were young in the late '70s/early '80s_ and were interested in programming experimented with 6502 assembly language. Most people with an interest in programming are either too dead, too old or too young to fall into that narrow window. The majority of people with an interest in programming have never used 6502 assembly language. That's not to say that it wasn't a popular processor, just that people have been interested in programming for many more years than the 6502 was relevant. I programmed in 6502 assembly language myself in the mid '80s before moving on to ARM, which _is_ relevant, and in a major way, today.
@@johnm2012 I should have rephrased that, most people that had a Commodore 64 as a kid and were interested in programming ended up using assembly on the 6502. We had clubs all over the place and everyone had their copy of MADS.
@@adisharr The BBC Micro was the big seller here at the time. It came with the excellent BBC BASIC with a built-in assembler which made it very easy to try out 6502 assembly language. All the OS entry points were documented so there was no need for PEEKING and POKING obscure memory locations and a whole 16 bytes of precious zero page was set aside for user vectors. It was a great platform to learn on. It was so good that I never bothered with more powerful processors, such as the Z80 or the 68000 but jumped straight from 6502 to ARM, which immediately seemed strikingly similar - like a 32-bit version of the 6502. Perhaps that's not surprising since it's instruction set was designed by Sophie Wilson, who knew the 6502 inside out, having written BBC BASIC and key parts of the OS, in assembly language, of course.
Please make an i7 next.
The MOnSter has approx 17.88 transistors per square inch. At that density, an i7-860 which has 774M transistors would be a PCB measuring 548 feet on each side...
@@Zadster I know. Imagine how cool that would look. Like a giant i7 billboard.
@@VidarrKerr That would be epic! Maybe cover the entire side of a skyscraper.
Now build a scale DIP around it!
Projects like these are great. They demonstrate engineering by using creativity.
But....
will it run Crysis?
No, buy Intel CPU.
buy ryzen cpu
I saw one playing Darude - Sandstorm.flac
bal.sac
Yes it will run Crysis. Your grandkids will get to see the first pixel drawn.
Looking forward to seeing Update #2!
This is cool. The "music" is not cool, it is hard to hear what you are saying!
I had no problems understanding the narrator over the music
mad respect for the mad scientist.
you can secretely install an i7 underneath and achieve better results
Evi1M4chine yeah I definitely felt like I was dealing with the mob whenever I buy an Intel processor it's not like just buying anything else
Excelent proyect!!!
Saludos desde Ushuaia Argentina!!!
Where's the monster Z80
build yourself, it roughly has about twice the transistor count 8,5k
give me a few years.
6502 gets all the fame but most of the early work I did was Z80 based, mostly TRS-80. Wanted to work on the 8080 but they were too expensive at the time. A 4004 replica should be easy to do. I think someone did something similar for a PDP-8.
Finally we are getting closer to all the scifi movie computers from the 80s
all they need is to connect it to a projector so everything on the screen is projected onto their face while using it
c-can it run minecraft?
If you could port Linux and Java to 6502, then yes. It may have already been done, but I don't know.
well, gl getting even linux to run on a 1mhz processor
Linux and java to 6502 xDDD Waaaat You realize we are talking about a 8bit cpu, right ?
Very nice indeed!!!! I'm sure there'll be a market for it. Sign me up.
I was actually thinking of making a 4-bit computer using simple diode logic (light-emitting for more fun), and only transistors when absolutely required. But my brain fried when I thought about it. Nice short story.
next project ryzen 3950x megamicroprocessor :P
I would buy multiple of these boards just because they look so cool.