Hello All,
For awhile I've been wanting to get a better understand on how software communicates with hardware and how the hardware works. Along with these ventures I've stumbled across a few questions. Currently I'm looking into buying IC's for NES cart building. The site I think I'm going to buy from is: BuyICNow.com (unless someone knows of a cheaper site?), so what I'm doing is searching for '27C020 ", form which I understand is the chip I need. What comes up is what i presume is a whole bunch of brands. Being a fresh college grad I don't have the cash for anything premium, would 'TI TMS27C020-15JL' work for nes carts? Also I'm currently following this turoail for learning purposes:
http://www.annahegedus.com/videogames/9 ... -nes-carts In the tutorial the user is soldering wires from the ic legs to other points. Now I have A.D.D (could have mist something in the video) so please bare with me, is she doing it because the tecmo basketball game is mapped differently? If I use an original Super Tecmo Bowl Cart to make a 2013 one, or one with the same board would I have to do that?
Thanks for any help
BTW! do these: "32 PINS DIP Socket (RoHS)" also work with nes carts?
-Jake
Newborn repromaker
TI TMS27C020-15JL should work fine, if not degraded. Rewiring nedeed, because MaskROM's are not JEDEC compatible.
80sFREAK wrote:
Newborn repromaker
TI TMS27C020-15JL should work fine, if not degraded. Rewiring nedeed, because MaskROM's are not JEDEC compatible.
lol Newborn repro maker? Honestly my day job is c# coding for an insurance company, I have a computer information systems major which translates to 0 knowledge on hardware, I'm a fresh college grad and I just want to continue learning in my field.
besides that, thanks for the answer!
I've read a bit online about the degrading of these ICs, is there a way to prevent that? How long do these last? Also is there something I can read up that explains how to determine what the wires should be coming and going to? Is it pretty much the same for every game or do they differ? My last question is if I use these:
http://www.retrousb.com/product_info.ph ... ucts_id=42 do I have to worry about doing it?
Thanks!
-Jake
For games smaller than 64k/64k you usually don't need to do any rewiring.
We heartily encourage you to use ReproPaks whenever possible; most of us frown on the destruction of original games. However, the ReproPaks are only suitable for a large subset of all games (what are called mappers 0, 2, 3, 7, 34, and 66 with some rework), and he's sold out of MMC1 ReproPaks.
As far as purchasing, Mouser usually has better prices for small numbers of parts; also using EEPROM or Flash is usually less frustrating than UV EPROMs.
EPROM bit rot is a very slow process, and if you DTRT (one of those foil perfectly opaque stickers), you should easily get tens of years of life. I'm not clear on the relative security against bit rot for all of these technologies; certainly anecdota of thirty-year-old no-loss 'PROMs abound.
EPROMs can last a *very* long time. Some arcade games contain mostly or all EPROMs and continue to work to this day despite being 20 years old. But various factors will have an effect on the life span of the chip. And if the chip has "bit rot" you can always reprogram the chip later if you didn't damage it somehow.
There's also OTP EPROMs, which are usually a little cheaper than regular EPROMs when new, and they don't even have the window to let potential UV light in. They're not erasable, but this is fine if you're making something permanent with them. I dunno if they last longer than regular EPROMs, but it seems like there's less chance of accidental erasure at least.
Actually the way EPROMs work (or fail) isn't about UV light erasing data over time. Each bit is a cell that is charged by specific UV wave lengths that an eraser produces. The charge or amount in a cell has to be a certain level for this bit to be considered a 1. If it drops below a certain level then it's a 0. So bit rot would be over time that charge has diminished and a bit that was once set, is no longer set. OTP EPROMs are charged before they are packaged. Once they've been packaged, that's it. You have no way to recharge them because you'd need to do quite alot to open the package back up to recharge the cells.
I could be wrong but I recall reading this about EPROMs. So it's not that the window letting in light somehow lets bits rot or be erased. Again I've heard that it's actually difficult to erase an EPROM without a UV light, even the sun itself isn't a reliable eraser. So it's more about how well the EPROM was constructed and erased and programmed than worrying about covering UV windows or whatever else. Again, there are older arcade games with EPROMs which are 20+ years old and still work just fine.
There's some weirdness here.
This link agrees with you;
wikipedia seems to say the opposite. And on the third claw,
this screen shot shows that the values from bit rot tend to drift to data = address. So FIIK.
I definitely had had the (mis?)conception that the UV light "relaxed" the container that held charge, and programming "burned a hole" that allowed conduction through the FET.
It could be so that the UV light removes the charge in the cells rather than builds it up, I'm no expert on that. But we do know for certain that the erasing process results in a memory with all bits sit and that programming causes them to be clear and that you cannot reverse that without erasing. I don't know that I've seen any definitive sources on exactly how an EPROM works that you can be sure is true.
lidnariq wrote:
And on the third claw,
this screen shot shows that the values from bit rot tend to drift to data = address. So FIIK.
I definitely had had the (mis?)conception that the UV light "relaxed" the container that held charge, and programming "burned a hole" that allowed conduction through the FET.
If you follow to his source it's a discussion with
kevtris. His explanation seems to agree with wiki and makes sense to me atleast. Programming the bit stores electrons in the gate of the transitor, shining UV light of the right wavelength strips electrons from the gate (which jives with my and
wiki's knowledge of the photoelectric effect.) But there are many ways for electrons to leave the gate especially since they don't like each other... Heat, time, EMF, light, eventually you won't have enough electrons on the gate to keep the bit programmed to a '0' and it'll 'bit rot' to a '1' (unprogrammed).
As an aside, non-windowed OTP EPROMs are actually X-ray erasable EPROMs (or gamma if you want to be extreme), vice UV-erasable EPROMs.
Luckily EPROMs weren't cost effective at the time of the NES, and we got good old non-bit rotable mask ROMs in games.
You have your 1s and 0s backwards. Fully erased EPROMs are all FFs as I recall. Programming bits makes them 0.
The reason all our NES games are MaskROMs isn't because they cared about longevity or anything like that. It was about cost. MaskROMs are way cheaper than EPROMs (at the time) particularly in the quantity that games were produced in. MaskROMs can still bite the dust. But generally they tend to hold up well. Some pirate carts (besides modern ones) actually have EPROMs in them.
MottZilla wrote:
You have your 1s and 0s backwards. Fully erased EPROMs are all FFs as I recall. Programming bits makes them 0.
How is that different from what I said?...
infiniteneslives wrote:
eventually you won't have enough electrons on the gate to keep the bit programmed to a '0' and it'll 'bit rot' to a '1' (unprogrammed).
Perhaps the way I said it was just confusing. To program the bit you store electrons on the gate making it a (programmed) '0'. It bit rots when it loses the electrons which is the same thing that's done by erasing leaving the chip 0xFF or '1' for each bit (unprogrammed).
Quote:
The reason all our NES games are MaskROMs isn't because they cared about longevity or anything like that. It was about cost. MaskROMs are way cheaper than EPROMs (at the time) particularly in the quantity that games were produced in. MaskROMs can still bite the dust. But generally they tend to hold up well. Some pirate carts (besides modern ones) actually have EPROMs in them.
That's why I said we're lucky EPROMs weren't cost effective for Nintendo
How is 'bit rot' really possible with mask roms? It's possible for any chip to fail, but it'd be more likely for the address decoder or something that switches to fail than an individual bit. That and those types of fails would be induced to to wear vice sitting on a shelf and 'rotting'.
Aren't mask ROMs still more cost effective than EPROM? Not so in small quantities, obviously, but if we're talking about a Nintendo production we're not talking about small quantities.
Both of them are really a thing on the past...
Everything is flash now, or developing tech.
INL, yes all the wording gets confusing. Some of it would be that you'd expect the UV light to be charging and the programming to be emptying/losing something. You say it has to do with the opposite, that UV light depletes the cells of electrons. So then bit rot would be bits turning to 1 as they lose electrons.
MaskROMs are still and were in the past more cost effective than EPROM or EEPROM. But nowdays 3DS and PS Vita cards may just contain NAND flash memory similar to SD cards rather than MaskROMs. GBA used MaskROMs, I think Nintendo DS did too, though I think those were serial memories and not parallel. 3DS could still be MaskROMs of some sort. If anybody knows for sure that would be interesting to know. Flash has gotten so cheap so who knows.
DS used 8-bit parallel memory for the game storage. It was accessed using a block device protocol analogous to that of CompactFlash, just as both GBA and N64 had used seek-and-read protocols to reduce pin count, and the protocol was encrypted on the wire. Nintendo did invest in an OTP memory company called Matrix Semiconductor, but I don't know whether Matrix Memory chips ever ended up used in DS games.
MottZilla wrote:
INL, yes all the wording gets confusing. Some of it would be that you'd expect the UV light to be charging and the programming to be emptying/losing something. You say it has to do with the opposite, that UV light depletes the cells of electrons. So then bit rot would be bits turning to 1 as they lose electrons.
The photoelectric effect strips electrons, it doesn't deposit them. And as for the bit rot from 0 to 1, that agrees with what kevtris saw in hit rotten EPROMs.
Quote:
MaskROMs are still and were in the past more cost effective than EPROM or EEPROM. But nowdays 3DS and PS Vita cards may just contain NAND flash memory similar to SD cards rather than MaskROMs. GBA used MaskROMs, I think Nintendo DS did too, though I think those were serial memories and not parallel. 3DS could still be MaskROMs of some sort. If anybody knows for sure that would be interesting to know. Flash has gotten so cheap so who knows.
For what it's worth, I work for one of the remaining memory companies and I know we don't make mask roms. The idea of having a memory company mask out and fabricate ICs for your specific use is pretty insane really considering how DIRT cheap flash is. I can't imagine even with 3ds quantities that a custom IC could be more cost effective than flash is right now. Perhaps the small guys like macronix are still offering mask roms though, I can't be sure. If there were a non-destructive means to open my 3DS games I would.
INL, thanks for cleaning that up about the purpose of the UV. So that would be a cell with no electrons is actually going to return 1. So as you say, bit rot is actually going to cause bits to become set rather than clear. And then that makes sense that UV light exposure would potentially speed up bit rot. It just seems backwards from how I remember hearing it years ago.
Perhaps the info on DS or 3DS is out there somewhere. If either use Flash then in theory it may be possible to reprogram commercial cartridges. Though I'd guess you'd have to unlock them and know some sort of programming process to do it. Poor reliable MaskROMs left in the dust. =(
infiniteneslives wrote:
wiki's knowledge of the photoelectric effect.
Sorry I somehow double linked to the kevtris thread when I meant to link to the PEE (photoelectric effect), working now though.
As for the reprogramable 3ds thing. You never know, they could have easily used flash but then put some logic in to prevent a hacker from doing such things effectively making it OTP. Or some sort of encrypted key needed for the write command. If they did use flash I'd guess they tried to protect it to some degree.
That or use flash and cut off the write enable pin internally. Compare how SL1 works on DS and DS Lite units: if it is not bridged, most of the firmware flash is read-only. That'd even let a 3DS game use the same chip for the game and the save if the save is in the few sectors that don't require the additional write enable.