Discussion in 'Rare and Obscure Gaming' started by Asianat0r, Jul 2, 2015.
Ow, Im gonna print this on a T-Shirt !
So is there any hardware enhancements on this iteration of the SNES CD besides the CD drive? This looks like it's inspired by the PC Engine, an independent CD drive with some PCM RAM and a cart with SRAM to load data from disc to cart, end of story. Even the name "Super Disc" is close to Hudson's "Super CDROM".
No enhancements other than RAM on the BIOS cartridge. And from OzOnE's analysis, the CD drive is linked to the EXT plug.
I dont know, but even I loved Hudson HU Cards more then SNES Cartridges
Not every instance of English online will be written in the American style.
Never thought an actual box would be found, let alone a working prototype.
Easily one of the greatest finds in video game collecting history, as melodramatic as it sounds.
Not melodramatic at all. It's quite the big deal. I don't think Dan had any idea what he had in his hands.
So it got turned on, taken to HK, opened and the "OS cartridge" is maybe broken and CD-ROM doesnt get enough power...or maybe the AC adaptor fried something....did I miss anything ?
Maybe on reddit ? Something ?
I really hope for a dumped BIOS soon, as I wont have to do strange photoshopping anymore then.
omg, that would be...imagine....
I cant stand the silence anymore >_<
They powered it up - it played SNES / Famicom? carts just fine, and the BIOS cart ran as well (so at least the ROMs in it seem fine).
But, the self-tests for the CD drive part failed, which means the controller section of the main board either isn't connected properly (quite a few "bodge" wires around there), or the MCU that controls that stuff isn't programmed (or has been corrupted somehow, but that's fairly unlikely), or they never actually got to the stage of loading games from CD (also unlikely IMHO).
From the photos of the main boards that we got a glimpse of (and LuigiBlood also analysed / labelled), it looks like they (Sony/Tendo) had been debugging the CD section or whatever, and there are multiple links that have been made using small wires.
There are also some other weird jumper wires which seem to be powering the logic chip that buffers the data between the "SNES" and "CD Controller" parts, plus some extra wires connected to the CD Controller MCU.
As others have said - the first thing would be to get the BIOS cart dumped, as I think it would be relatively easy to at least emulate the system under the MAME / MESS emulator (well, for a more talented coder than me, but I'd give it a go as well.)
That said - we obviously don't know if there were / are any prototype CDs for the system yet, so there wouldn't be a great deal of point emulating it. Would be fun to try it out though ofc.
If it turns out that no CDs can be found, the main motivation for getting the CD drive working would be to increase the sale value of the unit (if they ever decided to sell it at auction etc., which is of course up to them.)
btw, as LuigiBlood also said - it doesn't look like there are ANY real hardware enhancements over a normal SNES / Fami other than the CD drive that's been bolted on, and of course the large amount of extra RAM in the "BIOS" cart.
The BIOS cart obviously just loads large chunks of data from the CD drive, then runs it like any other cart ROM. It doesn't look much more complex than that tbh. Other than the CD-loading stuff, the rest looks like a bog-standard SNES / Famicom.
If they could ever post some higher-res photos of both sides of the main board, it wouldn't take long to figure out most of the schematic tbh.
The only really interesting part would be how the CD drive attaches, but I'm 90% sure it will simply be added as an expansion port device.
Although I can't be 100% sure - you can almost see where the CPU data bus bits are connecting to the CD section of the main board. It would then only need a few other address and enable signals to access it.
Obviously, with a BIOS dump, it would take about 10 seconds to figure out which addresses are being accessed.
That info could then be used to emulate the CD drive (for what it's worth), then figure out which format the files / folders should be in.
It's highly likely the BIOS will be reading standard ISO 9660 format CDs, but knowing Nintendo, there's a good possibility they are using a custom CD format of some type (just to stop the early PC CD-ROM drives from easily reading the files.) That's all speculation on my part though.
Can't wait to see if they dump and release the BIOS, but I have a feeling they won't be releasing it publicly (we can only hope).
With the BIOS dumped, you could even re-create a SuperDisc console using a standard SNES. lol
Again though, without any prototype / official game disks, it would be mainly for fun at that point.
But, I'm sure the SNES coder community would have tons of fun writing new games and demos for it.
I thought I just had a brainwave as to which chip the CD controller MCU might be, but I don't think I've found it quite yet.
I was about to go to sleep, so even had to boot up the PC again to post this. lol
I thought I would share some of my incoherent and insomniac ramblings...
One of the most important photos of the SuperDisc motherboard is this one (from a reverse-engineering point-of-view)...
If you zoom in to IC205 in the "CD-ROM" area, you can see the wire links that are straddling the yellow line.
I'm fairly certain IC205 is just a standard 74HC245 (or very similar) bidirectional buffer chip.
You can see the "74" and "5" of the part number.
With the '245 buffer chip, the signals on pins 2 to 9 get simply get buffered and passed through to pins 18 to 11 on the other side of the chip (or vice-versa, depending on whether pins 1 and 19 are High or Low)...
So, it's fairly likely that the signals on left side of the chip (pins 2 to 9) are the main CPU data bus of the SNES side, then obviously the right-hand side of the chip (pins 18 to 11) are going to the CD controller MCU (IC203), and probably also to the CXD1800Q chip, which you see in this pic...
That's why I think the CD section is simply sat on the main SNES CPU data bus, and is likely just an expansion device.
You can see that the buffer chip is only 8-bits wide, so any additional control signals from the SNES side are probably connected directly to the MCU (IC203).
What is obvious is that the CD drive section isn't likely to work at all unless that brown wire connecting to pin 10 of IC205 is actually Grounded. (Pin 10 is actually the Ground pin for the buffer chip). I have asked Dan (on YouTube) if his tech guys can test that sort of thing, but he's probably too busy to even see my message now, let alone reply to it.
It's obvious that the CD-ROM section will in all likelihood be connected to the SNES' main data bus (albeit via IC205), but to back up the theory, you can see that the jumper wires / signals on the right-hand side of IC205 also connect to vias.
Then, just above those vias (well, to the "right" more in the first photo), there's another row of vias connecting to the CXD1800 chip.
Both LuigiBlood and I looked for the datasheet for that specific part number, but we haven't yet found one.
But, as I think LB mentioned before, it's probably just a generic version of the same series, and so likely shares a similar pinout
If we look at the datasheet for the more common "CXD1186" chip, we find... pinouts.
Well, the part number is from the "1100" series instead of the "1800" that we're looking for, but some of the pins do seem to tally with what we can see on the PCB...
It looks like most of the Data bus signals match quite well (DB0 to DB7), so it gives some evidence to back up that the buffer chip is sat on the SNES CPU Data bus (apart from being fairly obvious to begin with. lol)
EDIT: Forgot about this photo - it confirms at least that the supposed data bus from the "output" of buffer IC205 does connect directly across to the CXD1800 chip.
You can also see the small white components labelled "FL..." those are normally ferrite filter beads for supplying power to each chip / section, but can sometimes be fusible as well. Each of those should be checked for +5V or other supply voltages.
I can also see how some of the pins of the LCD display / eject board ribbon connect to the MCU. I've started making a preliminary pinout for the MCU. And yes, I'm bored. lol
Anywho - as for the CD MCU (Microcontroller) chip IC203 - I originally thought it might have been a common 8051 or something like that, (and it could still be an 8051-based chip, which started becoming popular around that time, and are REALLY common to this day.)
But, because there is no obvious external ROM / RAM for IC203, it will almost certainly have a bit of internal RAM, and probably an internal Mask ROM (hence "Microcontroller" as opposed to just a "CPU").
OK, so we know the chip has a QFP48 package, but the part number has been wiped off (or was never printed)...
Pin 19 might be the /RESET pin, as it has a 10K resistor on it, and a test pad on the same side, which suggests that it could be a pull-up resistor. (/RESET signals often have a test pad next to them.)
It looks like pin 20 of the MCU is connected to pin 11 of IC205 by one of the white wires - that signal is probably either a data bit or a control signal from the main SNES CPU bus side.
The other white wire looks to be connected between pins 16 and 38 of the MCU (IC203). I'm not sure why?
Also, most of the supposed SNES CPU data bus signals are connected to pins 37-48 of the MCU (well, at least 8 of them. lol), so they are almost certainly generic IO pins on the MCU.
I haven't yet found an older MCU which matches the same QFP48 package, and the expected pin layout yet, but I will keep looking.
If anyone has any ideas about which early-90s MCU it could be, please let me know.
The thing that worries me about getting the CD drive working is that there's no signs of life at all - even if it wasn't connected to the SNES CPU bus, you'd expect that the MCU would still start up, and would try to zero the laser sled, or perhaps let you eject the tray / show some numbers on the LCD display etc.?
That's what makes me think the MCU either has it's power disconnected somewhere on the board / polyfuse is blown, or it's simply had it's firmware erased?
Interesting stuff though. I'd love to get better photos of the boards, especially the underside of the motherboard. #winkwinkdan
Oh yeah, forgot to say. lol...
My theory earlier was that the MCU might have been an Hitachi (now Renesas) H8, which I think were starting to crop up in the early 90s?
But, I haven't found a QFP48 version with a pinout that seems to match the track layout as yet.
A Hitachi H8 chip is also used in the 64DD as the main controller, but that was obviously a couple of years after the SuperDisc was likely developed. Both units do share a very obvious heritage and connection though.
Here's the datasheet for the H8/3664.
There are many different part numbers though, so we might get lucky soon and hit on the correct chip if we keep searching...
Although, finding the correct chip / datasheet won't necessarily help get the SuperDisc fully running of course - that depends on a lot of other factors. More basic tests with a multimeter can help rule certain things out though.
As people have already said - the very first thing they could try is just to measure the voltages around the CD-ROM section, and try the test pads for +5V next to the LCD display etc.
Thank you very much OzOnE, for your last 3 posts ='D
No problem, @ItsMeMario.
I figured out a few other things yesterday - I only just noticed that the buffer chip (IC205) is soldered on "backwards". LOL
The reason for that actually made sense after only a short while - it looks like their test firmware was outputting an inverted signal from the MCU, or maybe it had incorrect timing or something? That signal was the one that controls the "DIRection" pin on the buffer chip...
It seems like they've simply rotated the chip (and shifted it by one pin), so that an inverted "DIR" signal can then be used.
That also explains why they needed the other brown wire for hooking up the +5V supply to the chip.
The buffer chip is getting it's Ground connection by pin 10 being soldered to pad "19" on the board. This would normally have been the /Output Enable signal, so it shows that the signal was likely just tied to Ground (ie. the buffer chip is always "enabled", it's only the DIRection signal that changes.)
But, for some other unknown reason, they've also added a wire between pins 16 and pin 38 of the MCU?
I'm fairly sure pin 38 was the original intended DIRection signal for the buffer chip, but then they have connected another wire from MCU pin 20 to pin "1" of the buffer chip (the DIRection input).
In theory, as long as they were / are outputting the correct signals from the MCU, using the logic chip in that reversed way should work fine.
It still means that there are quite a few things that need to be tested on the SuperDisc to have a chance of getting the CD section working.
It may even be possible that it will spring into life as soon as the chip is soldered the normal way around, and maybe the other wire between pins 16 and 38 is removed? Again, that is pure speculation on my part, and of course couldn't possibly be attempted until ALL of the surrounding signals were tested for proper timing etc.
Personally, I still think there is a voltage missing somewhere on the board, as the drive doesn't even eject the tray etc.
Plus, the real "smoking gun" is that the audio output apparently isn't working either?
Anywho, I've been working on some preliminary pinouts for the (as yet) mystery CD MCU, and the other NEC MCU (uPD75P308) on the LCD display board.
These are only VERY rough notes for the CD MCU pinout atm, and will definitely have errors / assumptions / guesswork...
I think I've properly traced the correct order of the CPU Data bus signals though, and maybe traced the /PAWR and /PARD signals to pins 36 and 37 on the CD MCU. Although, the signals might be swapped or completely incorrect - it's something that would take mere seconds to confirm with a basic multimeter.
Again, without better photos of both sides of the main board, much of this is just educated-guesswork based on my mild experience with PCB design / repairs etc.
For example, I could see from the pinout of the CXD1100P chip (S-SMP) that it's data bus bits were in a certain order relative to the vias that are just above it. Since it's a two-layer board (or possibly with some inner power planes), you can't really have tracks that "cross-over" each-other on the underside of the board, so it's likely that those vias connect to the S-SMP chip, and then you can visually see the traces carry on to the buffer chip IC205.
You can even see that the traces from those vias come from another set of vias close to the expansion port, then go up to the main SNES CPU etc. I would say with fair confidence that this is the main "SNES" data bus.
I used a combination of the other available photos to trace some of the Expansion port signals which are also connected to the cart connector. That re-affirms the idea that the CD section is just a normal expansion device, and probably doesn't even do any DMA stuff, it all looks like standard PIO stuff to me?
According to some of the CD chips that we have the datasheets for, the SuperDisc apparently has a "2x speed" CD chipset, so around 300 KB/sec peak transfer rate.) That would obviously fill up the total of 384KB of WRAM in no time though, even via PIO. hehe
(thanks again to LuigiBlood for the chip analysis etc. )
Anywho. Enough of my rambling again for one night.
Very nice observation !
I guess we cant expect anything new in 2015 anymore.
HongKong might be very exciting with such a machine in your hands... ;p
Well not to rain on anyone's parade...
But the fact that IC205 is in upside down and offset in the socket by a pin, and the fact that the board has clearly had some work done to it, makes me thing this was probably an engineering unit that got hosed and given to Olaf or whoever as a shelf piece. We fry hardware protos all the time during bring-up and they usually end up sitting on a desk or shelf somewhere. It does appear that they were pretty far along in the design process however given that a full PCB was spun and the silkscreening is in place to differentiate the CDROM area from the SFX area and what I presume was the original SFC circuitry.
There is still a ton to be learned from the hardware but I think trying to get it back to working order, especially without any media, is a pretty long shot given the perceived collector's value of the unit. Anything past passive voltage or signal capture seems to decrease the dollar value so I don't see them really going that far. It would be nice if they were willing to get some ultra high-res shots of both sides of the PCB though, and possibly indicate the layer count.
It is possible that the CD section of the unit is dead, but there are a ton of really easy and non-destructive tests that can be done yet.
It's also possible that there is still some firmware programmed into the CD MCU, since it was clearly being run at some point. The small mods that have been done do look perfectly legit for a proto unit too.
So, there's still a lot of hope for getting it running, I just think there are obvious voltage supplies missing, which would completely explain the missing audio output too (most audio opamps run off a +/- 12V-15V symmetrical supply, which could also be split, regulated, then used to run the CD section motors as well).
I'm surprised the LCD display doesn't work though, but that could also be due to missing supplies or a popped / removed pico fuse.
It would actually make sense that they purposely removed a fuse or filter bead to isolated the CD section from the SNES side too (for current measurement, or maybe there was a problem with it that they were trying to fix?)
LuigiBlood sent Dan an e-mail about two days ago enquiring about Dan sending us the higher-res pics of the main board (which he apparently already sent to somebody else?). He did try to reply, but the bulk of the message got lost for some reason. LB replied and asked again.
It may be a multi-layer board in regards to some power planes, but the signal layers seem to be just two-layers, which makes it a hell of a lot easier to reverse. I've found quite a few of the probably pinouts for the CD section MCU now, and I've identified the main CPU Data bus from the SNES side, and that's what goes to the buffer chip (as expected). It also looks like the /PARD and /PAWR signals are going directly to the MCU, which is also expected, and means that the CD section is just a standard expansion peripheral.
I'm almost certain that the audio APU isn't completely dead either, as the SNES side would never start up properly without it (I learned this while writing parts of the SNES FPGA core I'm working on.) The game code has to wait for the APU bootloader to start, then it loads it with a chunk of data for playing music and samples etc.
That's why I'm surprised that they didn't find out why the audio wasn't working - it would have taken seconds to test the audio output directly from the APU, and so ruling out the chipset itself. I know with many collector's items / antiques that certain sellers would rather have the unit intact and "as it was found", but in this case I'm sure it would be worth more if at least the audio was working, and even more if the CD section at least powered up.
If / when we do get another reply from Dan, we'll update this thread.
Back in the early 90s, CDROM tests would fail if no CD was in the unit. Just throwing this out there.
The CD tray on this unit apparently doesn't open, and nothing is displayed on the LCD.
This is too cool!
Separate names with a comma.