Fun fact: TI-99/8 was supposed to be a successor to the TI-99/4A computer but it got cancelled due to TI's massive financial loses. It kinda got released though as Tomy Pyuta (Tomy Tutor).
>texas Instruments takes over the computer market.
Nope, Ricoh bought out MOS would've only helped Nintendo in the long run.
>would've only helped Nintendo in the long run.
schizophrenia, everyone.
that is true. they were bootlegs that Apple commissioned WDC to make for the //e and IIgs and are still being made today for embedded use.
>bootlegs
licensed, gigaschizo.
Sega used Yamaha for the SMS/Mega Drive, Hitachi for the Saturn, and NEC for the Dreamcast chipsets. They all worked out of the box and didn't have major issues that took like a year and a half to fix.
saturn also used yamaha chip
Only Commodore were still using dies that big at that point. the trusty old 5 um process they'd been using since the Ford Administration that resulted in chips you could roast weenies on.
> schizo shit
i don't care if you found child porn in the autismleak. it's not how history unfolded. after commodore folded, wdc was licensing it out to anyone that wanted it. nothing was "bootlegged" and apple didn't invent anything. consider taking meds to contain your compulsive lying.
>after commodore folded, wdc was licensing it out to anyone that wanted it.
The patent on the 6502 expired in 1990 so even before Commodore went under they were free to license it.
the first Famicoms and SNESes had major faults with the chipsets that took some revisions to fix. it seemed Ricoh were a bit sloppy and not quite at NEC or Hitachi's level of engineering talent.
Epson manufactured the TurboGrafix-16's hardware for Hudson, not Ricoh, also Nintendo was supposed to pick up on Hudson's chip set but they rejected it because they were making their own 16-Bit hardware.
That and the 65C02 and the 65816 were Western Design Center chips, not MOS chips.
Sega used Yamaha for the SMS/Mega Drive, Hitachi for the Saturn, and NEC for the Dreamcast chipsets. They all worked out of the box and didn't have major issues that took like a year and a half to fix.
But they were responsible enough to delay shipments until the issue was fixed. Nintendo were putting known defective chips in the initial run of Famicom/SFC to meet shipping quotas. at least the Famicoms were recalled only after several months when the problems with them became apparent.
If Nintendo spent around 50 cents extra in hardware in their $75 dollars of profit per unit at that time this wouldn't of happen.
At $5 dollars of less profit they would've had component video (the hardware already supported it), much sharper video matching that of the much later SNES JRs and digital audio support as well, not just more durable chips.
6 months ago
Anonymous
no the problem was Ricoh bungled the chips much like NEC on those first Dreamcasts. it happens. any new technology tends to have problems.
6 months ago
Anonymous
Gigaleak said it was a die size problem that was causing the PPU to overheat, NEC was a whole nother story.
Same reason why early 360s had overheating issues, the die size was too big and shrinking the die fixed it.
6 months ago
Anonymous
Yes, they used a 4.5um chip for the PPU and Ricoh die skunked the PPU to 3um and later 1.5um to fix the over heating.
I think I saw some pics of de-capped early PPUs somewhere and the dies were the size of your thumb.
6 months ago
Anonymous
Only Commodore were still using dies that big at that point. the trusty old 5 um process they'd been using since the Ford Administration that resulted in chips you could roast weenies on.
6 months ago
Anonymous
those revision 0 Famicoms are very rare as most of the PCBs were swapped out for revision B boards in the recall program. not only do the PPUs get hot but there were other assorted hardware bugs and really awful video output with lots of RF interference. i saw a picture of a rev 0 board and there were some added resistors and capacitors added as field service modifications to fix various problems. i recall hearing that Rambo was one game known to glitch on rev 0 boards.
6 months ago
Anonymous
Weren't they "HVC-CPU 01" boards?
6 months ago
Anonymous
yeah the boards where the PPU has a silicone rubber heat spreader underneath it. those were the first bugged ones with the 4 um dies and Nintendo recalled them and swapped the boards for free. after they shrunk them to 3 um the chips had a metal spreader on top and the final 1 um version had no heat spreader.
6 months ago
Anonymous
Are there any pics of the 01 boards out there?
6 months ago
Anonymous
I had a pic I posted a while ago but I don't have it anymore. The board had mods like added caps and resistors to fix other assorted issues. one of them was a cap put between pins 47 and 48 on the cartridge slot (CHR/WR and CIRAM/CE). apparently this fixed the corrupted graphics on Rambo. it also had a resistor running from one of the power rails to ground.
6 months ago
Anonymous
They must have rushed it to market with all of those capacitors and resistors.
I wasn't thinking and should have done that to start with. There's that weird looking rubber heatsink that the guy mentioned above.
6 months ago
Anonymous
there was a site where a Japanese guy had a revision 0 Famicom.
>tested it with 50 different carts and the FDS >most of the games worked but at least two didn't one of which was Rambo as that had graphics glitches (however many of them were early period ones from <1986 and not like VRC4 stuff) >awful video quality with tons of RF interference (this was later corrected by putting a metal shield in the cartridge slot) >PPU would get scalding hot >something about "severe dropout compared to normal FC"
that last part i didn't understand. did he mean the well-known problem on rev 0 units with sprites disappearing as the PPU warmed up?
6 months ago
Anonymous
I know that the rev 01s sometimes freeze on the 2nd level of battletoads.
6 months ago
Anonymous
The added capacitor on pins 47/48 of the cartridge slot seems to have fixed the problem with Rambo. It might also fix Battletoads. Those pins have something to do with CHR data access and both Rambo and Battletoads are CHR RAM games. It was possible that there was some kind of signal cross talk or reflection across pins 47/48 that was corrected by attaching a cap to them.
6 months ago
Anonymous
I went to add this bit of info to the Wikipedia page (more detailed technical explanation of the problems with the initial Famicoms) and I found it reverted and I got this message. What happened here and how can I get rid of this?
6 months ago
Anonymous
First, get an account on Wikipedia instead of signing off with an IP like a coward. Mods tend to be much more ruthless towards anonymous edits.
Second, your edits were reverted as being unsourced. Usually one can easily republish them with proper sourcing and they will stay.
6 months ago
Anonymous
>Second, your edits were reverted as being unsourced
not him but I've had times where I had something reverted for being unsourced when I was doing the edits bit by bit and planning to add the source links once I was finished. you can't wait a day or so for me to finish and add the sources, you have to instantly revert it?
6 months ago
Anonymous
...Just add all information at once
6 months ago
Anonymous
ok
I went to add this bit of info to the Wikipedia page (more detailed technical explanation of the problems with the initial Famicoms) and I found it reverted and I got this message. What happened here and how can I get rid of this?
but first i need to get this block removed and this Ferret guy (I presume a mod?) fired. or at least also blocked because tit for tat. if he can block me i should be able to block him.
6 months ago
Anonymous
Wikipedia uses a cheesy-ass ban cookie just like Ganker where if you reset your router or something to get a different IP it still blocks you unless you clear your cookies.
6 months ago
Anonymous
So you can evade the block just by cleaning your cookies? That's pretty funny.
6 months ago
Anonymous
no it would still block your IP. the cookie is supposed to prevent you from ban evading by resetting your router/phone but you can just erase it.
6 months ago
Anonymous
just Google Image search that. a pic of one comes right up.
6 months ago
Anonymous
>NEC was a whole nother story.
From what I read it sounds like a manufacturing error of some kind. Maybe they had problems with wafer contamination.
Meanwhile Chad Sinclair was installing confirmed defective RAM chips in 48K ZX Spectrum boards and selling those computers as cheaper "beginner" 16K computers.
the fault with SHVC SNESes is less clear but the CPUs definitely had some defect and they're also buggy. the DMA controller doesn't work. however i heard guy on Reddit insist the issue with them was also a thermal one. they may too have used a larger die in the beginning that was shrunk later and reduced the operating temperatures. i've also heard it claimed the US SNES has worse case ventilation than the SFC.
6 months ago
Anonymous
The issues with that was with the voltage regulators, Nintendo cheaped out with them.
6 months ago
Anonymous
was that the issue with the dark vertical line on the screen? i have a GPM-02 SNES and i'm pretty sure it doesn't do that.
6 months ago
Anonymous
They even forgot the thermal paste
6 months ago
Anonymous
I've seen PPU faults in them as well. Then again it's sometimes hard to say if it's the chip itself or just broken traces which can be repaired. Also the SHVCs have the separate sound board and the console will not boot if it gets loose.
Yes, they used a 4.5um chip for the PPU and Ricoh die skunked the PPU to 3um and later 1.5um to fix the over heating.
6 months ago
Anonymous
4 um would be a pretty big, outdated node size, that's like 70s stuff. I'm pretty sure 3 um was the standard by 1983. My theory is they probably decapped a TMS9918 to use as a basis for the PPU (Nintendo were quite explicit about the Colecovision being the main source of inspiration for the Famicom) and the original revision of that chip required a sink because it got hot (the SG-1000 and MSX used a later TMS9918 revision with a smaller die so it didn't need the sink anymore). I guess the original one was 4 um since it had come out back in 1979 and maybe Ricoh just copypasted the original node size because they were on a timetable to get the thing finished and shipped. That's my theory anyway.
6 months ago
Anonymous
>Lateral thinking with withered technology.
6 months ago
Anonymous
proves Jay Miner's guys were that good. the Atari TIA was 5 um on all revisions of the chip except the last one used in the 2600 Jr. that switched to 3 um and they never had problems with it getting hot.
6 months ago
Anonymous
>I'm pretty sure 3 um was the standard by 1983
3 um had been out since 1978 when Intel introduced it and licensed it to other fabs. By 1983 they had 1.5 um but that was cutting edge stuff and most fabs didn't have it yet.
6 months ago
Anonymous
here's where it gets interesting. the OAM was implemented as DRAM because they could not physically fit SRAM onto the original PPU die as the cells would be too big. if they'd used 3 um or whatever from the beginning instead of a backward-ass 1976 vintage die size this might have not been an issue and would spare countless NES programmers the annoyance of having to refresh the OAM every frame or so.
6 months ago
Anonymous
I believe Famiclones use SRAM for the OAM so you didn't need to refresh. Some games designed for them won't work on a real NES for that reason.
6 months ago
Anonymous
The first Z80s from the 70s were 4 um and they did get pretty warm IIRC. The TRS-80 Model I used those first gen Z80s. They were slugs, too--only 2Mhz. By 1980 they'd been replaced with the improved die shrunk Z80A that ran at 3Mhz. Also note the RAM chips are ceramic probably because they were an early variant of 4116 while most later ones were plastic.
6 months ago
Anonymous
kek 1.5 um was just being paranoid. that was process size the 286 used and that had 130k transistors and much higher clock speeds without getting unusually warm. the PPU probably had 10k or so transistors. i'm sure it was less than VIC-II which had 13k.
6 months ago
Anonymous
This is Nintendo we're talking about, quality is their game.
6 months ago
Anonymous
pretty sure the Mega Drive and SNES chipsets were 1.5 um. for NES chipset that's like using a grain silo to store a bag of flour.
6 months ago
Anonymous
die shrinks have a practical reason as well, it lets you have higher yields. instead of 50 chips per wafer you can have 110 if the dies are smaller.
Not specifically but the first Japanese Mega Drives had a small issue with the clock crystal generating the wrong frequency for the VDP. This was corrected by adding a daughterboard with an additional crystal on it.
and the Saturn had like five boards before they managed to launch, originally it had a cooling fan, the 1994 consoles still have the slot on the side for them, and the boards have all the traces intact.
They still do them, and they've got some amazing image sensors too. Their fixed pocket cameras are second to none.
That said, its still very much an after thought to their printer business.
they had a lot of unused line capacity so they could fulfill Nintendo's orders. all the other chip fabs in Japan at that time were absolutely booked up and couldn't spare any production capacity.
the initial Famicoms had a rubber heat spreader around the PPU. the problem was the OAM RAM would start losing its contents as it warmed up and sprites would disappear (Uemura mentioned Baseball as a game where this would happen). the revision B and D chips and apparently even a few revision E PPUs had an aluminum spreader on top of the chip but this was completely removed by mid-1985.
Nintendo weren't a computer hardware company, they never would. Would've been more interesting to ask what if Atari were the one doing it and making low cost 8-bit computers and VCS successors. The gaming crash of 83 probably wouldn't have happened since the more advanced low cost systems would've replaced the 2600 completely and received higher quality games.
the initial Famicoms had a rubber heat spreader around the PPU. the problem was the OAM RAM would start losing its contents as it warmed up and sprites would disappear (Uemura mentioned Baseball as a game where this would happen). the revision B and D chips and apparently even a few revision E PPUs had an aluminum spreader on top of the chip but this was completely removed by mid-1985.
>Nintendo/Ricoh needed it alot more then Atari
No they didn't. They couldn't possibly make a system with huge profit margin (I'm talking about about 4-5 times the production cost) like the C64 was. Tramiel also had years of experience of manufacturing calculators and computers. He knew exactly how to make computers that sell and the best guys for the task. If nintendo tried to sell computers they would've gone out of business.
>they will be selling consoles
You don't get it do you. There's no way you could successfully sell a console with 500% markup. There's no reason for Noncetendo to own an entire chip manufacturing plant, they wouldn't know what to do with it. >Ricoh would be the ones buying MOS out, not Nintendo
Why would a japanese company buy an american manufacturing plant when they could illegally rip off the same products and produce them with cheaper manpower and greater capacity in japan? Japan back then was just like how china is today. Ricoh made more money from bootleg 6502s.
Nintendo apparently considered the possibility of switching the NES chipset to CMOS at some point (like how the Genesis 2 went to CMOS chips) and programmers were warned to not use illegal 6502 opcodes as they could break on future revisions of the hardware, but they ultimately never got around to it.
the first C64 board from 1982 that had ceramic VIC-II and video output so bad it looked like the screen was coated with vasoline. also both the power jack and video port used 5 pin din plug so there were a few unfortunate accidents involving misplaced plugs before they went to the 8 pin video port on later boards.
>only have to look at a CIA chip the wrong way and it'll die.
That's not a defect in the CIA, it's that they're extremely easy to damage from static electricity as they connect to the ports on the computer and there's no anti-static protection. Also this was in an era when many American homes still had shag carpeting--from anecdotal evidence it seems zapped CIAs were less of a problem in Europe as they generally have hardwood floors in their homes.
ugh, the lack of BCD mode was a royal PITA. there is source code listing for US DW1 online and you can see the hoops they had to jump through for some game algorithms.
Commodore initially bought out MOS in 1976. At that time they were using 6 um depletion load NMOS which was ahead of Intel and Motorola who used enhancement load NMOS that required ICs to use three power rails. MOS rather quickly managed to shrink this down to 5 um. By the time C64 was out they'd been using the 5 um process for years and understood it well--the first run of VIC-II and SID had 85% yields and within six months this was gotten down to 95% yields.
so in 1981 they licensed the newer 3.5 um process from Intel and opened a separate production line for it. the 3.5 um line was mainly to make simpler, cheaper components like CPUs and SRAM that would benefit more from the smaller node size. the VIC-II and SID were made on the 5 um line as there was no perceived benefit to making them smaller. the problem was MOS had major issues figuring 3.5 um out. the first chip they made with it was the PLA which had a huge failure rate. the TED/7501 in the Plus/4 was also faulty and most of them shit themselves. the production issues were solved after they later went down to 2 um. basically, any MOS chip with a 6xxx number is 5 um, 7xxx was used for the faulty 3.5 um chips, and 8xxx was used for 2 um chips.
If you don't know, enhancement load NMOS used an additional bias voltage or something in the substrate and if that voltage was missing the transistors would form parasitic diodes and instantly self-destruct. the 8080 and 6800 used this and one of the big pluses of the 6502 was that it was depletion load NMOS so it only needed one power rail.
>The decimal mode circuitry of the 6502 was patented. To circumvent US patent laws and avoid paying Commodore Semiconductor Group royalties, this was disabled on the 2A03. The physical circuitry for the decimal mode was still present on the chip die but a laser was used to cut one trace.[2]
The patent in question expired in 1990 as someone else mentioned earlier. They could have shipped NESes with working BCD mode at that point yet still didn't.
the clone chipset that UMC produced had working BCD mode and used SRAM for the OAM table so refreshing it wasn't necessary. for that reason games explicitly designed for Famiclones often don't work on real NES
Texas Instruments takes over the computer market. Video game consoles go extinct in North America and Europe.
Nope, Ricoh bought out MOS would've only helped Nintendo in the long run.
Nobody wants a gaming console in the US since they are too busy playing their Solid State Software Command Modules on TI-99/8 computers.
Fun fact: TI-99/8 was supposed to be a successor to the TI-99/4A computer but it got cancelled due to TI's massive financial loses. It kinda got released though as Tomy Pyuta (Tomy Tutor).
>texas Instruments takes over the computer market.
>would've only helped Nintendo in the long run.
schizophrenia, everyone.
>bootlegs
licensed, gigaschizo.
saturn also used yamaha chip
better yields per wafer, homosexual.
>I don't know about the Gigaleak.
> schizo shit
i don't care if you found child porn in the autismleak. it's not how history unfolded. after commodore folded, wdc was licensing it out to anyone that wanted it. nothing was "bootlegged" and apple didn't invent anything. consider taking meds to contain your compulsive lying.
>after commodore folded, wdc was licensing it out to anyone that wanted it.
The patent on the 6502 expired in 1990 so even before Commodore went under they were free to license it.
it was really just the BCD circuit that was patented and which they had to disable on the NES CPU to skirt patent laws
>The patent on the 6502
Which of the dozens of patents did youtube tell you is the "the patent" kiddo?
you have no idea what you're talking about
They would have still spun out and made low rent tech by the 90s? What boring fricking proverbial.
They would've been absorbed into Ricoh USA, plus Nintendo would've made sure everything was top notch.
the first Famicoms and SNESes had major faults with the chipsets that took some revisions to fix. it seemed Ricoh were a bit sloppy and not quite at NEC or Hitachi's level of engineering talent.
Epson manufactured the TurboGrafix-16's hardware for Hudson, not Ricoh, also Nintendo was supposed to pick up on Hudson's chip set but they rejected it because they were making their own 16-Bit hardware.
That and the 65C02 and the 65816 were Western Design Center chips, not MOS chips.
that is true. they were bootlegs that Apple commissioned WDC to make for the //e and IIgs and are still being made today for embedded use.
And Hudson and Nintendo would still use those chips.
>Epson manufactured the TurboGrafix-16's hardware for Hudson
they made the PCBs, NEC made the GPU/APU for the console.
Sega used Yamaha for the SMS/Mega Drive, Hitachi for the Saturn, and NEC for the Dreamcast chipsets. They all worked out of the box and didn't have major issues that took like a year and a half to fix.
Dreamcast had initial shortage in Japan because NEC had a huge failure rate during chip production.
But they were responsible enough to delay shipments until the issue was fixed. Nintendo were putting known defective chips in the initial run of Famicom/SFC to meet shipping quotas. at least the Famicoms were recalled only after several months when the problems with them became apparent.
If Nintendo spent around 50 cents extra in hardware in their $75 dollars of profit per unit at that time this wouldn't of happen.
At $5 dollars of less profit they would've had component video (the hardware already supported it), much sharper video matching that of the much later SNES JRs and digital audio support as well, not just more durable chips.
no the problem was Ricoh bungled the chips much like NEC on those first Dreamcasts. it happens. any new technology tends to have problems.
Gigaleak said it was a die size problem that was causing the PPU to overheat, NEC was a whole nother story.
Same reason why early 360s had overheating issues, the die size was too big and shrinking the die fixed it.
I think I saw some pics of de-capped early PPUs somewhere and the dies were the size of your thumb.
Only Commodore were still using dies that big at that point. the trusty old 5 um process they'd been using since the Ford Administration that resulted in chips you could roast weenies on.
those revision 0 Famicoms are very rare as most of the PCBs were swapped out for revision B boards in the recall program. not only do the PPUs get hot but there were other assorted hardware bugs and really awful video output with lots of RF interference. i saw a picture of a rev 0 board and there were some added resistors and capacitors added as field service modifications to fix various problems. i recall hearing that Rambo was one game known to glitch on rev 0 boards.
Weren't they "HVC-CPU 01" boards?
yeah the boards where the PPU has a silicone rubber heat spreader underneath it. those were the first bugged ones with the 4 um dies and Nintendo recalled them and swapped the boards for free. after they shrunk them to 3 um the chips had a metal spreader on top and the final 1 um version had no heat spreader.
Are there any pics of the 01 boards out there?
I had a pic I posted a while ago but I don't have it anymore. The board had mods like added caps and resistors to fix other assorted issues. one of them was a cap put between pins 47 and 48 on the cartridge slot (CHR/WR and CIRAM/CE). apparently this fixed the corrupted graphics on Rambo. it also had a resistor running from one of the power rails to ground.
They must have rushed it to market with all of those capacitors and resistors.
I wasn't thinking and should have done that to start with. There's that weird looking rubber heatsink that the guy mentioned above.
there was a site where a Japanese guy had a revision 0 Famicom.
>tested it with 50 different carts and the FDS
>most of the games worked but at least two didn't one of which was Rambo as that had graphics glitches (however many of them were early period ones from <1986 and not like VRC4 stuff)
>awful video quality with tons of RF interference (this was later corrected by putting a metal shield in the cartridge slot)
>PPU would get scalding hot
>something about "severe dropout compared to normal FC"
that last part i didn't understand. did he mean the well-known problem on rev 0 units with sprites disappearing as the PPU warmed up?
I know that the rev 01s sometimes freeze on the 2nd level of battletoads.
The added capacitor on pins 47/48 of the cartridge slot seems to have fixed the problem with Rambo. It might also fix Battletoads. Those pins have something to do with CHR data access and both Rambo and Battletoads are CHR RAM games. It was possible that there was some kind of signal cross talk or reflection across pins 47/48 that was corrected by attaching a cap to them.
I went to add this bit of info to the Wikipedia page (more detailed technical explanation of the problems with the initial Famicoms) and I found it reverted and I got this message. What happened here and how can I get rid of this?
First, get an account on Wikipedia instead of signing off with an IP like a coward. Mods tend to be much more ruthless towards anonymous edits.
Second, your edits were reverted as being unsourced. Usually one can easily republish them with proper sourcing and they will stay.
>Second, your edits were reverted as being unsourced
not him but I've had times where I had something reverted for being unsourced when I was doing the edits bit by bit and planning to add the source links once I was finished. you can't wait a day or so for me to finish and add the sources, you have to instantly revert it?
...Just add all information at once
ok
but first i need to get this block removed and this Ferret guy (I presume a mod?) fired. or at least also blocked because tit for tat. if he can block me i should be able to block him.
Wikipedia uses a cheesy-ass ban cookie just like Ganker where if you reset your router or something to get a different IP it still blocks you unless you clear your cookies.
So you can evade the block just by cleaning your cookies? That's pretty funny.
no it would still block your IP. the cookie is supposed to prevent you from ban evading by resetting your router/phone but you can just erase it.
just Google Image search that. a pic of one comes right up.
>NEC was a whole nother story.
From what I read it sounds like a manufacturing error of some kind. Maybe they had problems with wafer contamination.
Meanwhile Chad Sinclair was installing confirmed defective RAM chips in 48K ZX Spectrum boards and selling those computers as cheaper "beginner" 16K computers.
the fault with SHVC SNESes is less clear but the CPUs definitely had some defect and they're also buggy. the DMA controller doesn't work. however i heard guy on Reddit insist the issue with them was also a thermal one. they may too have used a larger die in the beginning that was shrunk later and reduced the operating temperatures. i've also heard it claimed the US SNES has worse case ventilation than the SFC.
The issues with that was with the voltage regulators, Nintendo cheaped out with them.
was that the issue with the dark vertical line on the screen? i have a GPM-02 SNES and i'm pretty sure it doesn't do that.
They even forgot the thermal paste
I've seen PPU faults in them as well. Then again it's sometimes hard to say if it's the chip itself or just broken traces which can be repaired. Also the SHVCs have the separate sound board and the console will not boot if it gets loose.
>Nintendo were putting known defective chips in the initial run of Famicom/SFC to meet shipping quotas
Commodore did the same with the C64, and so did Microsoft with the 360 and Sony with the PS3.
For Nintendo's case it was cheaping out on minor parts.
Foe Microsoft and Sony it was the die was too large.
didn't the first Famicoms overheat or something?
Yes, they used a 4.5um chip for the PPU and Ricoh die skunked the PPU to 3um and later 1.5um to fix the over heating.
4 um would be a pretty big, outdated node size, that's like 70s stuff. I'm pretty sure 3 um was the standard by 1983. My theory is they probably decapped a TMS9918 to use as a basis for the PPU (Nintendo were quite explicit about the Colecovision being the main source of inspiration for the Famicom) and the original revision of that chip required a sink because it got hot (the SG-1000 and MSX used a later TMS9918 revision with a smaller die so it didn't need the sink anymore). I guess the original one was 4 um since it had come out back in 1979 and maybe Ricoh just copypasted the original node size because they were on a timetable to get the thing finished and shipped. That's my theory anyway.
>Lateral thinking with withered technology.
proves Jay Miner's guys were that good. the Atari TIA was 5 um on all revisions of the chip except the last one used in the 2600 Jr. that switched to 3 um and they never had problems with it getting hot.
>I'm pretty sure 3 um was the standard by 1983
3 um had been out since 1978 when Intel introduced it and licensed it to other fabs. By 1983 they had 1.5 um but that was cutting edge stuff and most fabs didn't have it yet.
here's where it gets interesting. the OAM was implemented as DRAM because they could not physically fit SRAM onto the original PPU die as the cells would be too big. if they'd used 3 um or whatever from the beginning instead of a backward-ass 1976 vintage die size this might have not been an issue and would spare countless NES programmers the annoyance of having to refresh the OAM every frame or so.
I believe Famiclones use SRAM for the OAM so you didn't need to refresh. Some games designed for them won't work on a real NES for that reason.
The first Z80s from the 70s were 4 um and they did get pretty warm IIRC. The TRS-80 Model I used those first gen Z80s. They were slugs, too--only 2Mhz. By 1980 they'd been replaced with the improved die shrunk Z80A that ran at 3Mhz. Also note the RAM chips are ceramic probably because they were an early variant of 4116 while most later ones were plastic.
kek 1.5 um was just being paranoid. that was process size the 286 used and that had 130k transistors and much higher clock speeds without getting unusually warm. the PPU probably had 10k or so transistors. i'm sure it was less than VIC-II which had 13k.
This is Nintendo we're talking about, quality is their game.
pretty sure the Mega Drive and SNES chipsets were 1.5 um. for NES chipset that's like using a grain silo to store a bag of flour.
die shrinks have a practical reason as well, it lets you have higher yields. instead of 50 chips per wafer you can have 110 if the dies are smaller.
The PPU/CPUs on the early models tend to run around 55c to 65c temperature wise.
that's like VIC-II level of nuclear hot and i know the stated temperature ceiling for those was 70C (and they got pretty close when running)
Not specifically but the first Japanese Mega Drives had a small issue with the clock crystal generating the wrong frequency for the VDP. This was corrected by adding a daughterboard with an additional crystal on it.
and the Saturn had like five boards before they managed to launch, originally it had a cooling fan, the 1994 consoles still have the slot on the side for them, and the boards have all the traces intact.
Probably not much as Ricoh's big money maker was always printers and their IC business was always secondary.
They did film cameras too, my dad has one sitting around but he hasn't used it since the 90s.
They still do them, and they've got some amazing image sensors too. Their fixed pocket cameras are second to none.
That said, its still very much an after thought to their printer business.
they had a lot of unused line capacity so they could fulfill Nintendo's orders. all the other chip fabs in Japan at that time were absolutely booked up and couldn't spare any production capacity.
the initial Famicoms had a rubber heat spreader around the PPU. the problem was the OAM RAM would start losing its contents as it warmed up and sprites would disappear (Uemura mentioned Baseball as a game where this would happen). the revision B and D chips and apparently even a few revision E PPUs had an aluminum spreader on top of the chip but this was completely removed by mid-1985.
Playchoice 10 PPUs had either ceramic shells or a large finned sink on them.
Nintendo weren't a computer hardware company, they never would. Would've been more interesting to ask what if Atari were the one doing it and making low cost 8-bit computers and VCS successors. The gaming crash of 83 probably wouldn't have happened since the more advanced low cost systems would've replaced the 2600 completely and received higher quality games.
Nintendo/Ricoh needed it alot more then Atari.
Again, die shrinks fixed it.
>Nintendo/Ricoh needed it alot more then Atari
No they didn't. They couldn't possibly make a system with huge profit margin (I'm talking about about 4-5 times the production cost) like the C64 was. Tramiel also had years of experience of manufacturing calculators and computers. He knew exactly how to make computers that sell and the best guys for the task. If nintendo tried to sell computers they would've gone out of business.
They wouldn't be selling computers, they will be selling consoles and Ricoh would be the ones buying MOS out, not Nintendo.
>they will be selling consoles
You don't get it do you. There's no way you could successfully sell a console with 500% markup. There's no reason for Noncetendo to own an entire chip manufacturing plant, they wouldn't know what to do with it.
>Ricoh would be the ones buying MOS out, not Nintendo
Why would a japanese company buy an american manufacturing plant when they could illegally rip off the same products and produce them with cheaper manpower and greater capacity in japan? Japan back then was just like how china is today. Ricoh made more money from bootleg 6502s.
Nintendo apparently considered the possibility of switching the NES chipset to CMOS at some point (like how the Genesis 2 went to CMOS chips) and programmers were warned to not use illegal 6502 opcodes as they could break on future revisions of the hardware, but they ultimately never got around to it.
>date code shows a November '85 DOM
>standard 6502 and not a 6510 etc.
This chip probably came out of a 1541, didn't it?
hkl
the first C64 board from 1982 that had ceramic VIC-II and video output so bad it looked like the screen was coated with vasoline. also both the power jack and video port used 5 pin din plug so there were a few unfortunate accidents involving misplaced plugs before they went to the 8 pin video port on later boards.
amazing how successful they were considering the early problems. only have to look at a CIA chip the wrong way and it'll die.
>only have to look at a CIA chip the wrong way and it'll die.
That's not a defect in the CIA, it's that they're extremely easy to damage from static electricity as they connect to the ports on the computer and there's no anti-static protection. Also this was in an era when many American homes still had shag carpeting--from anecdotal evidence it seems zapped CIAs were less of a problem in Europe as they generally have hardwood floors in their homes.
i've heard in general that PAL C64s are less prone to malfunctions than NTSC ones and why is anyone's guess
The NES would be allowed to keep it's decimal mode.
ugh, the lack of BCD mode was a royal PITA. there is source code listing for US DW1 online and you can see the hoops they had to jump through for some game algorithms.
>ugh, the lack of BCD mode was a royal PITA.
lol. no. Tell me you can't code without telling me.
Ricoh buying MOS instead of Commodore would've fix that.
>Ricoh buying MOS would have learned me to code
lol. no.
Commodore initially bought out MOS in 1976. At that time they were using 6 um depletion load NMOS which was ahead of Intel and Motorola who used enhancement load NMOS that required ICs to use three power rails. MOS rather quickly managed to shrink this down to 5 um. By the time C64 was out they'd been using the 5 um process for years and understood it well--the first run of VIC-II and SID had 85% yields and within six months this was gotten down to 95% yields.
so in 1981 they licensed the newer 3.5 um process from Intel and opened a separate production line for it. the 3.5 um line was mainly to make simpler, cheaper components like CPUs and SRAM that would benefit more from the smaller node size. the VIC-II and SID were made on the 5 um line as there was no perceived benefit to making them smaller. the problem was MOS had major issues figuring 3.5 um out. the first chip they made with it was the PLA which had a huge failure rate. the TED/7501 in the Plus/4 was also faulty and most of them shit themselves. the production issues were solved after they later went down to 2 um. basically, any MOS chip with a 6xxx number is 5 um, 7xxx was used for the faulty 3.5 um chips, and 8xxx was used for 2 um chips.
didn't the Intellivision also use sinks?
It did because the chips made with an ancient enhancement load NMOS process that used triple voltages.
If you don't know, enhancement load NMOS used an additional bias voltage or something in the substrate and if that voltage was missing the transistors would form parasitic diodes and instantly self-destruct. the 8080 and 6800 used this and one of the big pluses of the 6502 was that it was depletion load NMOS so it only needed one power rail.
>The decimal mode circuitry of the 6502 was patented. To circumvent US patent laws and avoid paying Commodore Semiconductor Group royalties, this was disabled on the 2A03. The physical circuitry for the decimal mode was still present on the chip die but a laser was used to cut one trace.[2]
This would've been averted of Ricoh bought MOS instead of Commodore.
The patent in question expired in 1990 as someone else mentioned earlier. They could have shipped NESes with working BCD mode at that point yet still didn't.
Some games like Battletoads used the lack of BCD as a piracy check so that bootlegs like the Dendy couldn't play them.
the clone chipset that UMC produced had working BCD mode and used SRAM for the OAM table so refreshing it wasn't necessary. for that reason games explicitly designed for Famiclones often don't work on real NES
some Wikipedia pages you can edit all day long with nothing happening and others it gets reverted instantly. i've never been able to figure it out.