The Mitsumi interface is so annoying because it's basically undocumented except for an ancient Linux driver and I need to make a Mitsumi emulator for a project :(
I had that exact PC as a hand me down from my dad. He bought it in 1992 from a small computer shop for $2499 with a 486 SX 25 soldered to the board, 1MB of 30 pin SIMMs, black 3.5" and 5.25" drives, black mechanical keyboard, black case CRT, and a small hard drive with Windows 3.1. In 1994 or 95, he put a 486 DX/2 66 in the coprocessor socket (ran at 50MHz because of the 25MHz bus), 8MB RAM, and a Soundblaster kit with a sound card with built-in IDE CD-ROM controller and a white faced 4x drive. In 1998 or so, I got it along with a Windows 95 upgrade disc and ran that. I aquired a Windows 98 disc, but it was already dog slow with 95. In 2000, I upgraded the motherboard to one with an AMD K6/2 450 that ran Win98SE just fine with some amount of RAM I don't remember as only a year later I got an ATX case with an Athlon 1000 that spelled the end of life for the black case PC.
Very disappointed there was no baby at this teardown, but I can abide this fascinating video. I have an empty CD-ROM drive box in my basement that did not include the required interface card, either.
You're kinda right about the tape drive ribbon cable. The odd cable would've come with the tape drive, but its purpose is really to add a third connector (pre-twist, mind you...) to the normal 2-drive floppy cable. Because, of course, if you had a 5-1/4" drive and a 3-1/2" drive, you would have no spare connectors to plug in the tape drive (ignoring that each "connector" often physically had both pins and a card-edge connector a couple inches apart, for compatibility with both drive types) and the last connector would be post-twist as well. It wasn't specifically to address a length issue (note that the extension connector is near the controller card end of the cable anyway.) Those tape drives would often support higher bitrates than normal PC floppy controllers, so you could install a special high-speed interface card that was designed specifically for the tape drive alone. AFAIK, that was the main difference -- it was still essentially a floppy controller under the hood, it just ran at a faster speed, and existed somewhere in the IO address space that the BIOS wouldn't look for. Ergo, the drive was compatible with either interface type because they were basically the same thing.
I was so incredibly disappointed as a teenager when I found out the LED Mhz display wasn't actually measuring the clock frequency. Also I hate those U shaped outer case panels.
Seeing that CD drive with the caddy was a CHILDHOOD MEMORY UNLOCKED moment for me. We had one of the electronic eject variants of this with an identical disc caddy in the first 'multimedia PC' that my family owned. I want to say it was a Packard Bell machine. I remember what a novelty it was for us at the time. Cheers!
The PC Audio header is an input to replace your actual PC Speaker. The four pin header on your PC is for driving an 8 ohm cone speaker (the middle two pins are not connected but I guess it stops you plugging the speaker into the reset switch header). Or, you could wire that header to this input and the audio would go through a buffer and into the sound card’s digital mixer and then on to the outputs with the rest of your audio. Useful if it annoyed you that the sound for Police Quest came from a different place to the sound for Police Quest II, and it gave you a volume control for those early PC Speaker games. I dimly recall that some cards also had a TEL header which was two mono line level channels to go to/from your ISA Voice/Fax modem. I’m now amused at the thought of dialling in to listen to someone play Doom.
The Mr. BIOS is likely the keyboard controller. Dual BIOS chips only really were a thing if the BIOS needed more than one specific size of (EP)ROM, splitting it into the high and low ROMs (except for later boards with dual BIOS ROMs for redundancy).
Yes. One of the two large ICs -- either the Mr. BIOS (not sure if they made KBCs?) or the one with the AMI BIOS sticker on it -- is definitely the keyboard controller. It would not be the first time an OEM put the BIOS sticker on the KBC chip.
Hi Gravis! I decoded those faults just for fun, and those are looking like cache or memory errors in the RAM around the 1MB area (it's in the upper memory boundary). Since it's a 486, a floppy disk version of MemTest would work, and it looks like I'm using 2.00 here (4.00 works most of the time but causes conflicts with VLB cards!) If you need a good floppy disk image of MemTest 2.00, please feel free to reach out! My DEC Venturis 466 has a weird quirk where Windows will crash with one of those errors if the IDE cable isn't 40 conductors (instead of 80 conductors), from what I was able to determine the card wasn't able to handle the noise over IDE when there were a lot of reads happening consistently. Booting DOS wouldn't really trigger this, but launching Windows definitely would! You may have a faulty controller or IDE cable too that's loading bad data into memory, and then causing Windows to crash as a result. I'd check your drivers that are being loaded too since it's possible one of those could be loading corrupted data, but honestly the easiest first step towards solving those faults is probably doing a memory test...
If it’s a cache issue, it could be related to the DLC. The cache in that thing only works reliably when some flags are set correctly. Some motherboards supported this chip better than others, and it could very well be that some settings for it in the bios got wiped.
Speaking from memory here, Matsushita Electric was the name of the company itself, and Panasonic is only a brand name. The use of the Matsushita name for the bus is largely because Matsushita used different brand names in different regions: Panasonic was the brand for most of the world but in Japan the brand name was simply “National” in the same font. That said, that didn’t last forever and now they’re Panasonic around the world including Japan.
Alright!!! This is so great. I remember my dad bringing that exact CD-ROM drive home in what must have been ‘94. I think he bought it off a colleague. It was our first CD-ROM drive and went into his 486, mainly for his NT4 install (32Mb ram baby!) and MSDN cd’s. The faceplate had this great MPC logo on it (multimedia pc), and I still think caddy’s are rad. Eventually the 486 became my pc, and I swapped out the drive for a 32x or something, as 1x is too slow to display the FMV fluidly in command&conquer. Music ran fine though. Still used the thing as a spare for years, it was a very reliable reader, even of scratched discs. Eventually in 2000 or so the laser crapped out and I chucked it. Always wanted to put one in a retro build, but I never found one on eBay.
The CPU is a 486-ish-chip-in-a-386 package. That's why it's not great, but it's one of the best options for the 386 socket. MR BIOS is generally a more-featured bios than stock, often allowing for larger drive support, but that's a keyboard controller in your case.
I was sesrching comments to see if anyone else noticed that this is indeed a 486DLC so that would make that motherboard a somewhat uncommon 386DX motherboard with vesa local bus! (CPU Galaxy did some tests on a ne a "while ago") So IMHO It would be a shame that motherboard get scraped 😢
Fun fact: The interface card for the Forte VFX-1 VR headset was an ISA card, that used a VESA feature connector to read the graphics card's framebuffer to display on the VR headset - it had no GPU itself, so it worked surprisingly well (aside from a few graphics modes it couldn't understand) with a Matrox Mystique which also had a VESA feature connector. They also used a pluggable precursor to USB for connecting the gyro in both the headset and the handheld puck controller.
Oh man... That CD-ROM drive. In high school, our auto shop had a computer with that exact drive (which I know because it had the mechanical-eject caddy) with various auto repair manual CDs. It was A Process to get to look things up on that computer. It also had a light pen for interacting with the program on a monochrome amber CRT. Must have been Hercules graphics, as it would very visibly switch between text and graphics modes. The program for loading your repair manuals was all text mode, with the light pen to navigate (or type on a keyboard with a VERY nasty waterproof barrier over it,) and once you selected your item, it would slowly load up the image to display in monochrome.
This machine was CLEARLY loved by its original owner. Black case, spray-painted drive faceplates, math co-processor, Mr. BIOS replacement, exotic sound card, even more exotic CD-ROM (at the time) .... This was not a typical mom-and-pop build. Someone really put some time and money into this thing.
Ohh, it seems so weird to me that you never came across any battery damaged 386/486 motherboard repair videos by people like Necroware, he is in Europe, but Adrian Black has also some battery bombed motherboard repair videos. Hope you can get it repaired as this is not just a standard 486 VLB motherboard but a very late 386DX motherboard designed to be paired with that weird 486DLC with only 1kb of level 1 cache Cyrix came out to compete with cheap Intel 486s, it's BTW plenty faster than any 386. Good luck with that and happy new year! 🎉
This MB is like 30 years old also, I had a lot of motherboards that fell in my hands that were a lot younger and had really dirty slots of all types cause remember everything wants to get inside a slot when there is plenty of time, moisture and varta nicds! And they will definitely stay there until somebody with a kind heart and patience (and probably an air gun, ipa, vinegar, contact cleaner spray, fiberglass pen etc) blows their shit away and make the contacts shiny again to let the data flow happilly! The CPU Cyrix set is not crap I would say it is rather interesting and rare at least for us Europeans, and I think that everything about this PC is interesting, I would love to see the Tape Drive work, the case is very cute and practical, that CDROM is certainly worth a repair or at least a try it is really unique! You are very lucky dear Sir even the HDD workswhat else of an indication do you need to be told that this machine wants to live but needs your help? Cheers from Greece, may this New Year stop a war or 2 for change (and not start new ones) health and peace. Jim
That case looks exactly like the one we had for our 486 back in the day, just beige instead of black! Would be keen to track one of those down again to rebuild the system of my childhood.
I don't recall Cardinal sound cards, but I did have a Cardinal 33.6 modem that I got in ca. 1998. As you and others have discussed on many occasions, there was a lot of overlap in the engineering (analog, DSP) between sound cards and modems.
Yeah .. when I think of Cardinal (which isn't often), my first thought is modems, too. I don't know if I would've remembered that they had sound cards now, but it seems like I knew that once upon a time. It was only a surprise in that it wouldn't have been a common choice, so fairly rare to see now.
35:00 Yeah that's a leaked battery. The board can be fixed but it needs a lot of love and work. The component next to the battery is a ferrite bead, not a fuse.
In the early days of CD ROM, retail packaging of non-Scsi drives was typically a multimedia bundle with sound card and speakers. That is probably why interface cards were not an issue at the time. If you were buying a CD ROM at retail, you probably wanted a sound card. If you were in business, you would probably get what your system builder put together…even if it was you.
@ I only made sense of my first CDROM buying experience 32 years ago just now. Because of the technical detail in your video. It was a bundle of cheap components in a colorful box for $180 on the closeout shelf in Walmart.
the laser preamp may also need recapping on that Panasonic drive; I have a similar vintage Clarion slot-loading CD player with a Panasonic drive mech, and all the electrolytics on the laser preamp are blown. it had identical symptoms: insert a disc, disc would spin, you'd hear the laser focus over and over, and it'd spit the disc back out.
also the Panasonic CD drive my dad bought in 1994 or 1995 came with the controller card, but he ended up using his soundblaster to avoid wasting an ISA slot.
and fwiw every 'dead' single speed drive I've found was due to bad caps, simple mechanical failures, or the even objective lens falling out! those lasers are a lot less powerful/over driven compared to newer drives with higher data rates.
90% sure the brand of case is 'Macase' and model ks-330. I remember my school (here in Australia) having stacks of 486 dx2 66 machines in these particular cases; circa ~1996. The school repurposed the cardboard boxes the towers came in, one box was filled with balls for PE hence that model number is somewhat permanently etched into my memory! Nice find :) I've never seen a black version of that style case either
The MKE interface. Back when I had lots of time and little money, I made an interface card from a discrete ttl printer port card, and using the Linux source as a reference. Mr. Bios chip is a keyboard controller. Went into one one of the big chips 486 and later.
Announcing Mr. BIOS, the latest brand new product from the company that brought you such fine products as Mr. Coffee, Mr. Radar, Mr. Rental and Mr. Fusion.
On the "early IDE were just on the ISA bus" - that's one reason why you had to manually specify the physical layout of the drive in the BIOS; because there were no "smarts" to determine any of that on the motherboard side early on.
No, those two things don't actually have anything to do with each other. In fact the earliest hard drives (MFM) used an interface that was very much like the Shugart floppy drive. The controller (an actual controller, on a giant card in the ISA bus) would send step commands to the stepper motor to select tracks, and head-select commands to pick which of the heads to use, which meant the software had to tell the controller the geometry, so the controller knew what signals to send to the mechanical bits on the drive. Once we had IDE, the controller moved to the drive, and the geometry was abstracted away to the point that only the drive knew what it _really_ was. But, the software interface had long since been standardized to that old MFM style, so the C/H/S parameters stuck around long after they became meaningless to the drive itself. The numbers you still had to enter (until the BIOS auto-detect did it for you) were purely fictional, and for the most part, they just served as a means to calculate the total number of blocks the drive was willing to report as accessible. Sometimes you had to lie, take from one column and multiply another proportionally (you didn't really think your drives had 64 or 255 heads, did you?), just to ensure that none of the numbers got too big to store in the fields they were originally given. The drive and/or BIOS and/or OS would translate between these schemes in a messy and fragile combination that could (and sometimes did) lead to "misunderstandings" and corrupted file systems. Thankfully, roughly around when OSes went 32-bit, we had LBA -- a much more straightforward "0 to nth sector" model that we've used ever since.
That battery hadn't gone apocalyptic yet. If you see visible crystallization on the battery can then it was leaking fast or for a long time. It gets out through the end cap of the can as the seal fails, then the electrolyte spreads over the board and up into the slots by capillary action, like a wick. The seal material isn't immune to bases so it eventually just gets eaten through. If you were lucky none of the broken traces are directly under slots or chips you'd need to desolder.
When I was a kid around 1992 or 93 I had a friend who had an old 386 with a CDR almost just like that +2 speakers built into the case more than likely that header is for something like that I think I saw a mosfet on that card. I still remember playing the DOS version of MegaMan on that thing.
The fact that the errors are inconsistent coupled with the MS DOS RAM extender error when it craps out of Windows leads me to guess that it's a RAM issue, or more specifically a problem on the motherboard side of the RAM equation - the corrosion was quite close to the memory so there may be some damage hidden under the slots, and as the BIOS RAM check is extremely crude and only does the most perfunctory of testing it's not uncommon to pass the check and then run into issues when you actually try to *use* the memory. A diagnostic ROM and/or only populating some of the slots may help narrow it down.
I went over to the used computer store, popped a plain ISA IDE controller into it, and all the problems went away, it works perfectly now. Then I put a VLB video card in it, it didn't start up the first time, I wiggled it, then it started but there was corrupted graphics. So the VLB port is definitely messed up, and I did some further thinking - the reason this card has a controller chip instead of just being a bridge over to the ISA bus is because that controller chip is actually going through the VLB, I confirmed this by looking at the design of another card I found that has both VLB and ISA IDE ports on the same card. So, since we know the VLB is messed up, and the hard drive was going through the VLB rather than the ISA, I'm pretty sure that was why it was having all these problems, it was loading corrupted data from the drive.
@@CathodeRayDudeGaiden Yeah the fact that the stack is all zeros to me indicates that the CPU is probably executing a bunch of garbage instructions. Which would happen if during load time the text section/COMDAT of the image gets corrupted. In other words, most definitely in-flight corruption during load of the CPU instructions.
@@LunaTheFoxgirl Battery corrosion can definitely affect VLB. I cleaned up a 486 board that very luckily only had slight battery leakage, but VLB doesn’t work because some traces for it run directly underneath the battery. ISA works fine because those traces aren’t doing that whole maneuver around the board like VLB tends to.
One brand of battery notorious for leaking when it is found in computers is Varta. Varta is a German brand so that surprises me somewhat. I can imagine some dirt cheap no name Chinese battery leaking more realistically but the fact is the Varta CMOS batteries were known for leaking. Is the cause of that the same as for alkaline cells that got left in things too long and leaked? I am sure that some of us had a toy we liked for a while and forgot about it while the batteries were still in it then maybe a decade later we find it and are disappointed by finding that has happened. Especially since often it does not work any more since the stuff that leaked is caustic.
There were a few DIY -NES cart-dumpers- eeprom programmers that (in single/pure DOS mode) used the DMA/IDE interface to generate (and read) the signals by accessing the relevant addresses. -- edit: I forget if later versionsgot around the clock-speed/timing issue by tapping one of the address lines ; but now I wonder if the interface card(s) bundled with so many grey-market cart/rom dumpers is(/were) a repurposed CD-ISA bridge/address decoder.
I honestly didn’t know that ATAPI hadn’t always been a thing. I used SCSI on my Amiga, and by the time I switched to PC in the late 90s, I guess all the dust had already settled.
Damn! I need that mechanical loading CD-ROM drive in my life! The CPU and FPU are also very special and so is the 386 mobo with VLB. It's a dream system. Please save it.
Vinegar will help neutralize alkaline battery guts, just thoroughly rinse it off within a few minutes or it will damage other components. Deoxit is safer if you have it. Do be careful cleaning slot pins. They are way thinner than you think. Unfortunately, they put the RAM near the corrosion generator. Random crashes may be damaged traces near the RAM. If it's a multi-layer PCB, fixing them may be impossible.
Unfortunately we are loosing history day by day from Varta destruction and cap plague. At some point we need to start preserving every 8080, 88, 86, 286, 386, and 486 (and into the Pentium era) regardless of how common they were during production. Working units are becoming rare now and are sliding towards extinction.
I have a very similar VGA+Multifunction combo card that uses the same chips I can confirm that for VGA it's about as fast as you can go for 486 video without having to go to PCI. It's not the fastest VLB Video card but it's pretty damn close and you're more likely to be CPU Bound instead. Your suspicions about the promise chip are correct - That is a VLB IDE controller and is much faster than an ISA IDE "controller" (That as you mentioned, really just shoves the IDE bus onto the ISA bus) I too have used it in a 386/486 combo board, although I have a different model (Although mine also came with a TI 486DLC chip, on a motherboard with a soldered on 387? These late 386/486 boards are wild. An expensive math co processor just soldered to the board, and you can override that one by slapping one into the 387 socket) These boards are kind of rare, and they can be extremely fast 386 boards or kind of fast 486 boards. Yours has 256kb of cache which is mighty impressive for a system like this. Someone cost engineered the shit out of this machine to be as much of a multimedia machine as you could get without doing the right thing and putting a 486 into it. If you pull that 387 chip you can put a 486 in that socket and get a bit more speed out of the board. You may also need to remove the TI 486DLC (Which is just a 486 core in a 386DX pin pattern, so you lose a little speed when going back to the 386 DX bus system compared to the native 486 Bus) to make this work. As for VLB stability issues. Try the other slot. These 386/486 combo boards typically only support VLB in both slots if you're running a 486 CPU, and if you're running a 386 CPU only one of the slots is officially supposed to work (Which is why you have a combo VGA/IDE board - your 386 Class CPU only runs one VLB slot so you have to make the most of it) VLB cards are wildly sensitive to signaling issues, and these 386/486 combo boards have to do some Deep Magic to make VLB run on a 386, even if it is very fast on a 386. You can also try lowering your CPU clock speed to 33mhz, which is probably safer on a board that has corrosion issues. The corrosion on this board doesn't look that bad. You could probably save it from more of an early death with a good long bath in some IPA and a bit of scrubbing with a tooth brush.
Seem to remember that the Amiga a500 cd rom (a570?) came with a mechanical eject caddy cdrom drive! Also regarding the board refurb, do watch some of Adrian digital basement… ;)
The PnP Sound Blasters still had support for the three CD-ROM protocols. I have one! A later model than mine replaces all of them with “IDE” (ATAPI?), before it goes away entirely. I guess it was still worth including for 386/486/early Pentium users who wanted to add CD-ROM. Maybe you mentioned it and I didn’t realise, VLB isn’t strictly for graphics. My 486 has a VLB I/O card, no graphics on that. It definitely uses it for the IDE ports, which I know because (luckily very minor) battery leakage damaged VLB traces that ran directly underneath the battery. MR. BIOS was very impressive, it has a very fast boot mode. I flashed it on my Pentium board and it definitely is fast. Has an early in-BIOS boot menu too. And yes, it calls itself MR. BIOS on boot!
Good to know on the PnP ones, I don't have many that have the interface so I wasn't sure. Re: VLB, yep, the card I have in this video uses it for all that stuff! I also found one at the REPC today which has a REALLY interesting layout: two IDE ports, one labeled "ISA IDE", one labeled "VL IDE", and you can see how one goes through a host controller and one doesn't!
17:38 I've recently seen those backro-named to MAin or Main Attachment and SL= Secondary Link. A little clunky, but a good way to get some un-wanted baggage out of IT terminology.
I might! I've since done some more work and gotten it to run Windows reliably, but the VLB port definitely needs to be replaced and that trace repaired for the thing to work perfectly, and I don't really need another machine that's not super reliable, so maybe it would be better if he had the board and I just built a different machine in this case
Oh interesting, I had this Case/MB/CPU combo back in the mid 90s as a hand-me-down, but with a beige case. The MB seems in bad shape, but please don't throw it away or scrap it. Maybe you noticed that the 486 is smaller than a usual 486. That's because that's a socket for a 386. Yes. This board can run either a 386/387 combo or a 486 in the socket around the 387's socket. And yes, it can use VLB with the 386 (not to its benefit of course). Beautifully weird design :) If you *really* want to throw it out, let me know and I might want to pay for shipping to my "recycling facility". I've probably had enough battery corroded PCB repairs in my life, but for this gem, I'd do it again ;)
Yeah I've been getting a lot of comments to the effect that this is a very unusual and interesting board. Unfortunately with the damage it'll never a great testbed machine for me, but I definitely don't intend to trash it, don't worry. Once I decide what to do with it I'll probably scrub through these comments again and see who gets dibs :p
I don't know, but I wonder if that ROM socket on the sound card might be sample ROM? As you said, it does have an optional SCSI interface, so it COULD be boot ROM... but the SCSI chipset footprint looks like the old NCR 5380 or a derivative thereof. These were used on the OG Mac and later on sound cards like the Pro Audio Spectrum. You COULD add a hard drive to those, but they were not fast. They were low cost and optimized for low-speed peripherals like (early) CD-ROMs. You probably wouldn't want to boot from that interface.
so yeah, about that! i have this DIM memory floating around in my head that i've seen four-drive IDE chains and i could not remember where the hell it came from. maybe that's it??
I'm not sure how that could possibly happen. The select lines only allow for two combinations. If there's any way to get four discrete drives to function on the same bus, it would have to be a purely software trick that overrides the hardware addressing conflict somehow. I don't know enough about the lowest levels of PC interface programming to say whether that's even remotely possible, but I kinda doubt it. If such a thing does actually exist, I would be very curious to see how it pulled it off. I would assume it had to augment the IDE interface to add another select bit (2 bits = 4 combinations.)
the pc speaker header is an INPUT, it mixes the pc speaker from the motherboard header through the sound card instead. You could donate the mainboard to Adrian black for restoration, i'm sure he could make a good video out of it and it would sate your curiosity.
shame that board is toast. those hybrid 386/486 boards are pretty interesting. maybe with some vinegar it can be rescued edit: ok so a bit of info, that 486 installed is actually a TI 486 SLC, that is, a 486 upgrade for a 386 motherboard. so that slot is for a 386 class cpu. what's interesting is the math copro socket, which is waay bigger than it needs to be. that can actually take a full 486 cpu. so it's a nice board, usually opti stuff is reliable, and vlb 386 is kinda interesting, while mostly useless lol
I’ve managed to get slots working by spraying contact cleaner into the slot and then just rocking the card back and forth to scrape off any crap. Adrian Black from Adrian’s Digital Basement swears by DeOxit, but that’s not available in the UK so I personally swear by WD-40 contact cleaner. (Not regular WD-40 though!!)
Would be pretty cool if you could get it rehabbed in all of its terrible glory. And then maybe max it out. Who knows, we can throw a few bucks into the beer fund.
Lime juice? that motherboard must smell nice. The fact windows is hung but CTRL+ALT+DEL works is interesting because that means it wasn't some hardware fault necessarily. The bus and memory and CPU are still running and not halted or clobbered too badly, such that a lot of systems within the software are still working. It was easy for windows 3.x to be misconfigured and just fail to boot. Could be some driver just waiting forever for some device which isn't present. Weirdly, even with all those other errors, it's still not totally hanging the system. I can't entirely imagine what they might be. Maybe bad/unreliable RAM would be my first guess. Maybe some unreliable sectors on the disk occasionally misreading a bit here and there. But could even be some deeper issue with something just being a little unreliable with memory accesses or IO with the disks or maybe some other devices which windows is trying to initialize.
It really did smell nice!!! Sometimes it WOULD hang hard, other times it would just give me a blinking cursor but C+A+D worked, etc etc - all moot now. I pulled the VLB card, put a plain ISA IDE controller in there + an ISA VGA card, and all the problems went away, machine worked perfectly. So I'm thinking it's the combo card spitting noise onto the bus when it wiggles due to the wrecked socket, causing totally stochastic failures.
If both the Matsushita interface and IDE are 'just ISA', does that mean it would be possible to adapt a Matsushita drive to an IDE bus? Obviously it's still not going to be speaking ATAPI so you're still going to need weird drivers for it.
"All the interfaces were proprietary." "Drives didn't come with interfaces, I suppose..." Not entirely correct. SCSI was a thing by then, and lots of early CD-ROM drives _were_ proper SCSI drives. In fact, that Matsushita 2x drive (the 562/563) actually did have a 2x SCSI variant. Lots of other ancient drives, from NEC and Ricoh and Yamaha, e.g., stuck to SCSI. The problem with that is, of course ... SCSI was super expensive for PC users. Unlike many contemporary platforms, the PC didn't have a native SCSI controller in the de-facto reference designs (originally, the IBM PC, then the Compaq 386 -- after that, everything was a gentleman's agreement), so you were required to source your own. Adaptec and Future Domain were the usual go-to options, but it was often hundreds on top of the PC and the expensive CD-ROM drive itself. The "Big Three" proprietary interfaces (which, BTW, leaves out Phillips' LMSI) were designed to be low-cost alternatives, and took the same approach as IDE: Make the controller little more than an extension of the ISA bus, and place it on the drive so it can be as minimal and purpose-built as possible. The add-in card itself was really just some address selection logic and a few buffers. Perhaps it also allowed changing the address, and selecting whether or not to map an IRQ or DMA channel. Oh, and of course it often allowed you to adapt the audio cable to RCA or 3.5mm phone jacks. This simplicity is why they were so easy to incorporate into sound cards. With a tiny bit of glue logic and some off-the-shelf buffers, you could add any or all three interfaces with very little effort. The pinout isn't even that much different -- just the same signals in a different order, mostly. The "secret sauce" is the command set they use, which is opaque to the hardware in between. When ATAPI came out, it allowed packaging SCSI commands in packets that could be transferred via the IDE interface, which took existing (SCSI) software and combined it with a trivial update to the bus interface we all already had, making it the shortest possible path to a complete solution. As to whether the drives came with interface cards -- yes, in fact, they did ... at first. Some of the original drives shipped with their bespoke controller, usually proprietary like LMSI (a very early solution.) SCSI drives often did not, since e.g., Mac (and other) users wouldn't need it, and the PC didn't have overwhelming market share until closer to 95, so that mattered. But, it wasn't unheard of to see drives in a bundle with a SCSI card targeted at PC users. Future Domain specialized in this market as a cheaper alternative to Adaptec's higher-end cards. I suspect it may have been a coping strategy to get them separately, since each would set you back a month's rent, and it was probably easier to stomach one at a time. ;-) A little later, in the first half of the 90s, the "Multimedia Kit" was a big deal, since multimedia was a huge driver to both CD-ROM and sound card purchases. These were really pricey, but would usually include the sound card, CD-ROM, some cheap desktop speakers, a mic, and a bundle of software that was definitely going to include Mavis Beacon Teaches Typing and an encyclopedia or human biology reference disc. Perhaps a kids' storybook title like Grandma and Me. Once it was potentially likely that you *had* a sound card already (but not necessarily a CD-ROM), then they started shipping with the cheapest possible ISA card -- MKE, Mitsumi, and very early ATAPI IDE being the most common. (The IDE card would often be configurable as secondary or tertiary, since this was around the time where VLB controllers and on-motherboard interfaces were beginning to ship with two IDE connector interfaces instead of just one.) The CT1860 is one such card that I happen to have on my desk, which shipped with (e.g.) the Creative Omni CD or Blaster CD 4x (another Panasonic / Matsushita) drive kits.
Okay, well, you went on to discover a lot of that yourself -- which was inevitable, I guess. :-) Oh well, forgive the lengthy narrative and just assume it makes for better storytelling and confirmation of your assumptions. As for the 34- and 40-pin query... Sony's was a very simple 8-bit interface -- which is why it can be a smaller connector. It only has D0 through D7 connected, plus corresponding grounds between the signal lines, of course, and a handful of control pins. That's it. The MKE interface is 8-bit too (not sure if there was ever a 16-bit version), but it has 8 extra lines that I haven't figured out yet. They're all connected to something, but the board that I've personally RE'd (the CT1810) has a GAL IC that I haven't yet fully mapped out. It looks like some of it is address logic, or at least depends on the selected address. Four other lines are just buffered from a pull-up resistor pack, and muxed with D3..0 when enabled by the GAL. I'unno. 🤷 I have at least one Mitsumi interface card around here somewhere (and IIRC, it's also 8-bit), but it doesn't appear that I've made a schematic of it yet. (It's much easier to trace them when you're not concerned about potentially destroying the PCB in the process by desoldering everything, and I prefer to take the opportunity when the board is a little worse for wear or doesn't work anyway.) I know it's similar to the other two, but whether it could've been a 34-pin connector too, I'm not sure. A lot of this is probably long since decoded by others, but I like reinventing wheels, so I haven't gone out of my way to spoil the puzzle for myself.
"full" For me means that I wrote a script, made sure there weren't any apparent errors, and made the diction clean enough that a decent proportion of the viewers on my main channel will want to watch it all the way through instead of losing interest a couple minutes in. If I thought it wasn't worthy of a video at all there wouldn't be a video :p But my normal standards pretty much don't allow me to make a video in less than a week, whereas this one took 3 hours
Was wondering that too. At least three popular channels do repairs on battery-bombed motherboards on the regular. Nobody should be left wondering what to do with these boards. :-)
I love those "other 90% of the computer" cards like the 6 pack. I bet they solved a lot of slot related issues back when you needed a card for everything that wasn't one of the 30 standards. if you needed more slots I bet some people just dangled the ports out an empty slot or out of one of the knockouts instead of having a ton of brackets if they needed more ports.
The Mitsumi interface is so annoying because it's basically undocumented except for an ancient Linux driver and I need to make a Mitsumi emulator for a project :(
What fun that must be needing documentation that hardly exists
I had that exact PC as a hand me down from my dad. He bought it in 1992 from a small computer shop for $2499 with a 486 SX 25 soldered to the board, 1MB of 30 pin SIMMs, black 3.5" and 5.25" drives, black mechanical keyboard, black case CRT, and a small hard drive with Windows 3.1. In 1994 or 95, he put a 486 DX/2 66 in the coprocessor socket (ran at 50MHz because of the 25MHz bus), 8MB RAM, and a Soundblaster kit with a sound card with built-in IDE CD-ROM controller and a white faced 4x drive. In 1998 or so, I got it along with a Windows 95 upgrade disc and ran that. I aquired a Windows 98 disc, but it was already dog slow with 95. In 2000, I upgraded the motherboard to one with an AMD K6/2 450 that ran Win98SE just fine with some amount of RAM I don't remember as only a year later I got an ATX case with an Athlon 1000 that spelled the end of life for the black case PC.
Mr. BIOS later earned his PhD and called himself Dr. DOS.
Very disappointed there was no baby at this teardown, but I can abide this fascinating video. I have an empty CD-ROM drive box in my basement that did not include the required interface card, either.
😂😂😂
That was my first thought when seeing the all lower case title, too.
You're kinda right about the tape drive ribbon cable. The odd cable would've come with the tape drive, but its purpose is really to add a third connector (pre-twist, mind you...) to the normal 2-drive floppy cable. Because, of course, if you had a 5-1/4" drive and a 3-1/2" drive, you would have no spare connectors to plug in the tape drive (ignoring that each "connector" often physically had both pins and a card-edge connector a couple inches apart, for compatibility with both drive types) and the last connector would be post-twist as well. It wasn't specifically to address a length issue (note that the extension connector is near the controller card end of the cable anyway.)
Those tape drives would often support higher bitrates than normal PC floppy controllers, so you could install a special high-speed interface card that was designed specifically for the tape drive alone. AFAIK, that was the main difference -- it was still essentially a floppy controller under the hood, it just ran at a faster speed, and existed somewhere in the IO address space that the BIOS wouldn't look for. Ergo, the drive was compatible with either interface type because they were basically the same thing.
i like seeing these sillier quick videos on the side channel! even if its unimportant things like "this computer is weird" its fun to see
“Quick” 47 minutes. 😂 I agree about liking these off-the-cuff vids.
Yeah, I always love videos like this
Probably "quick" because the 47 minute video took less than 2 hours to make (not counting the time spent troubleshooting the computer).
I was so incredibly disappointed as a teenager when I found out the LED Mhz display wasn't actually measuring the clock frequency. Also I hate those U shaped outer case panels.
Same on both counts! Hell is littered with those U-shaped case panels!
Seeing that CD drive with the caddy was a CHILDHOOD MEMORY UNLOCKED moment for me. We had one of the electronic eject variants of this with an identical disc caddy in the first 'multimedia PC' that my family owned. I want to say it was a Packard Bell machine. I remember what a novelty it was for us at the time. Cheers!
The PC Audio header is an input to replace your actual PC Speaker. The four pin header on your PC is for driving an 8 ohm cone speaker (the middle two pins are not connected but I guess it stops you plugging the speaker into the reset switch header). Or, you could wire that header to this input and the audio would go through a buffer and into the sound card’s digital mixer and then on to the outputs with the rest of your audio. Useful if it annoyed you that the sound for Police Quest came from a different place to the sound for Police Quest II, and it gave you a volume control for those early PC Speaker games.
I dimly recall that some cards also had a TEL header which was two mono line level channels to go to/from your ISA Voice/Fax modem. I’m now amused at the thought of dialling in to listen to someone play Doom.
Adrian’s Digital Basement is in Portland and he fixes this stuff all the time. Could be a great cross-over.
Was going to say something similar about the PC speaker header. If you check the SB16 MCD, it should have one as well (at least mine does).
The Mr. BIOS is likely the keyboard controller. Dual BIOS chips only really were a thing if the BIOS needed more than one specific size of (EP)ROM, splitting it into the high and low ROMs (except for later boards with dual BIOS ROMs for redundancy).
Yep, an 8042 microcontroller variant.
Yes. One of the two large ICs -- either the Mr. BIOS (not sure if they made KBCs?) or the one with the AMI BIOS sticker on it -- is definitely the keyboard controller. It would not be the first time an OEM put the BIOS sticker on the KBC chip.
Hi Gravis! I decoded those faults just for fun, and those are looking like cache or memory errors in the RAM around the 1MB area (it's in the upper memory boundary). Since it's a 486, a floppy disk version of MemTest would work, and it looks like I'm using 2.00 here (4.00 works most of the time but causes conflicts with VLB cards!)
If you need a good floppy disk image of MemTest 2.00, please feel free to reach out!
My DEC Venturis 466 has a weird quirk where Windows will crash with one of those errors if the IDE cable isn't 40 conductors (instead of 80 conductors), from what I was able to determine the card wasn't able to handle the noise over IDE when there were a lot of reads happening consistently. Booting DOS wouldn't really trigger this, but launching Windows definitely would! You may have a faulty controller or IDE cable too that's loading bad data into memory, and then causing Windows to crash as a result. I'd check your drivers that are being loaded too since it's possible one of those could be loading corrupted data, but honestly the easiest first step towards solving those faults is probably doing a memory test...
If it’s a cache issue, it could be related to the DLC. The cache in that thing only works reliably when some flags are set correctly. Some motherboards supported this chip better than others, and it could very well be that some settings for it in the bios got wiped.
Speaking from memory here, Matsushita Electric was the name of the company itself, and Panasonic is only a brand name. The use of the Matsushita name for the bus is largely because Matsushita used different brand names in different regions: Panasonic was the brand for most of the world but in Japan the brand name was simply “National” in the same font. That said, that didn’t last forever and now they’re Panasonic around the world including Japan.
"Call Mr BIOS, that's my name! That name again is Mr BIOS!"
Alright!!! This is so great. I remember my dad bringing that exact CD-ROM drive home in what must have been ‘94. I think he bought it off a colleague. It was our first CD-ROM drive and went into his 486, mainly for his NT4 install (32Mb ram baby!) and MSDN cd’s. The faceplate had this great MPC logo on it (multimedia pc), and I still think caddy’s are rad. Eventually the 486 became my pc, and I swapped out the drive for a 32x or something, as 1x is too slow to display the FMV fluidly in command&conquer. Music ran fine though. Still used the thing as a spare for years, it was a very reliable reader, even of scratched discs. Eventually in 2000 or so the laser crapped out and I chucked it. Always wanted to put one in a retro build, but I never found one on eBay.
The CPU is a 486-ish-chip-in-a-386 package. That's why it's not great, but it's one of the best options for the 386 socket. MR BIOS is generally a more-featured bios than stock, often allowing for larger drive support, but that's a keyboard controller in your case.
I was sesrching comments to see if anyone else noticed that this is indeed a 486DLC so that would make that motherboard a somewhat uncommon 386DX motherboard with vesa local bus! (CPU Galaxy did some tests on a ne a "while ago")
So IMHO It would be a shame that motherboard get scraped 😢
Fun fact: The interface card for the Forte VFX-1 VR headset was an ISA card, that used a VESA feature connector to read the graphics card's framebuffer to display on the VR headset - it had no GPU itself, so it worked surprisingly well (aside from a few graphics modes it couldn't understand) with a Matrox Mystique which also had a VESA feature connector. They also used a pluggable precursor to USB for connecting the gyro in both the headset and the handheld puck controller.
Oh man... That CD-ROM drive. In high school, our auto shop had a computer with that exact drive (which I know because it had the mechanical-eject caddy) with various auto repair manual CDs. It was A Process to get to look things up on that computer. It also had a light pen for interacting with the program on a monochrome amber CRT. Must have been Hercules graphics, as it would very visibly switch between text and graphics modes. The program for loading your repair manuals was all text mode, with the light pen to navigate (or type on a keyboard with a VERY nasty waterproof barrier over it,) and once you selected your item, it would slowly load up the image to display in monochrome.
The nest of crumpled IDE cables makes me barf with anti-nostalgia
Mr BIOS is right at home with Mr Coffee, Mr Radar and Mr Rental.
That's a wild looking future computer for its age for sure. that would have been a REAL Cool Guy ™ machine
This machine was CLEARLY loved by its original owner. Black case, spray-painted drive faceplates, math co-processor, Mr. BIOS replacement, exotic sound card, even more exotic CD-ROM (at the time) .... This was not a typical mom-and-pop build. Someone really put some time and money into this thing.
Huh...I did not know that. I generally figured all non-ATAPI drives were some stripe of SCSI.
Today I learned!
Ohh, it seems so weird to me that you never came across any battery damaged 386/486 motherboard repair videos by people like Necroware, he is in Europe, but Adrian Black has also some battery bombed motherboard repair videos.
Hope you can get it repaired as this is not just a standard 486 VLB motherboard but a very late 386DX motherboard designed to be paired with that weird 486DLC with only 1kb of level 1 cache Cyrix came out to compete with cheap Intel 486s, it's BTW plenty faster than any 386.
Good luck with that and happy new year! 🎉
I've been fortunate enough to have acquired a Commodore CDTV, and that uses a CD caddy, and it's also a mechanical eject!
This MB is like 30 years old also, I had a lot of motherboards that fell in my hands that were a lot younger and had really dirty slots of all types cause remember everything wants to get inside a slot when there is plenty of time, moisture and varta nicds!
And they will definitely stay there until somebody with a kind heart and patience (and probably an air gun, ipa, vinegar, contact cleaner spray, fiberglass pen etc) blows their shit away and make the contacts shiny again to let the data flow happilly!
The CPU Cyrix set is not crap I would say it is rather interesting and rare at least for us Europeans, and I think that everything about this PC is interesting, I would love to see the Tape Drive work, the case is very cute and practical, that CDROM is certainly worth a repair or at least a try it is really unique! You are very lucky dear Sir even the HDD workswhat else of an indication do you need to be told that this machine wants to live but needs your help?
Cheers from Greece, may this New Year stop a war or 2 for change (and not start new ones) health and peace. Jim
That case looks exactly like the one we had for our 486 back in the day, just beige instead of black! Would be keen to track one of those down again to rebuild the system of my childhood.
Nice video. 👏🏼
Now I’d like to see a new video with the rehab of this machine, fixing the case like new with a fresh paint.
I don't recall Cardinal sound cards, but I did have a Cardinal 33.6 modem that I got in ca. 1998. As you and others have discussed on many occasions, there was a lot of overlap in the engineering (analog, DSP) between sound cards and modems.
Yeah .. when I think of Cardinal (which isn't often), my first thought is modems, too. I don't know if I would've remembered that they had sound cards now, but it seems like I knew that once upon a time. It was only a surprise in that it wouldn't have been a common choice, so fairly rare to see now.
For the case of my first 386 circa 1992 or so I absolutely spray painted a big AT case black.
35:00 Yeah that's a leaked battery. The board can be fixed but it needs a lot of love and work. The component next to the battery is a ferrite bead, not a fuse.
Might be worth saving since since it looks like a 386/486 combo board (486 would plug where the co-pro is at the moment). VLB on a 386 is neat!
I would have called that "combo card" a SuperIO card back in the day. That's an AWESOME card!
In the early days of CD ROM, retail packaging of non-Scsi drives was typically a multimedia bundle with sound card and speakers. That is probably why interface cards were not an issue at the time. If you were buying a CD ROM at retail, you probably wanted a sound card. If you were in business, you would probably get what your system builder put together…even if it was you.
oh, good point! That's probably a huge part of it
@ I only made sense of my first CDROM buying experience 32 years ago just now. Because of the technical detail in your video. It was a bundle of cheap components in a colorful box for $180 on the closeout shelf in Walmart.
Don't forget the Philips LMSI CDROM interface that was also on some pro audio spectrum cards
Creative Labs sold retail 2X CD-ROM kits that bundled the proprietary controllers. Only example I really remember for OEM bundling.
the laser preamp may also need recapping on that Panasonic drive; I have a similar vintage Clarion slot-loading CD player with a Panasonic drive mech, and all the electrolytics on the laser preamp are blown. it had identical symptoms: insert a disc, disc would spin, you'd hear the laser focus over and over, and it'd spit the disc back out.
also the Panasonic CD drive my dad bought in 1994 or 1995 came with the controller card, but he ended up using his soundblaster to avoid wasting an ISA slot.
and fwiw every 'dead' single speed drive I've found was due to bad caps, simple mechanical failures, or the even objective lens falling out! those lasers are a lot less powerful/over driven compared to newer drives with higher data rates.
The few pre-IDE CD-ROMs I recall purchasing new did come with interface cards.
90% sure the brand of case is 'Macase' and model ks-330.
I remember my school (here in Australia) having stacks of 486 dx2 66 machines in these particular cases; circa ~1996. The school repurposed the cardboard boxes the towers came in, one box was filled with balls for PE hence that model number is somewhat permanently etched into my memory!
Nice find :) I've never seen a black version of that style case either
The MKE interface. Back when I had lots of time and little money, I made an interface card from a discrete ttl printer port card, and using the Linux source as a reference.
Mr. Bios chip is a keyboard controller. Went into one one of the big chips 486 and later.
Announcing Mr. BIOS, the latest brand new product from the company that brought you such fine products as Mr. Coffee, Mr. Radar, Mr. Rental and Mr. Fusion.
On the "early IDE were just on the ISA bus" - that's one reason why you had to manually specify the physical layout of the drive in the BIOS; because there were no "smarts" to determine any of that on the motherboard side early on.
Ahhhhhh, that makes sense.
No, those two things don't actually have anything to do with each other.
In fact the earliest hard drives (MFM) used an interface that was very much like the Shugart floppy drive. The controller (an actual controller, on a giant card in the ISA bus) would send step commands to the stepper motor to select tracks, and head-select commands to pick which of the heads to use, which meant the software had to tell the controller the geometry, so the controller knew what signals to send to the mechanical bits on the drive.
Once we had IDE, the controller moved to the drive, and the geometry was abstracted away to the point that only the drive knew what it _really_ was. But, the software interface had long since been standardized to that old MFM style, so the C/H/S parameters stuck around long after they became meaningless to the drive itself. The numbers you still had to enter (until the BIOS auto-detect did it for you) were purely fictional, and for the most part, they just served as a means to calculate the total number of blocks the drive was willing to report as accessible. Sometimes you had to lie, take from one column and multiply another proportionally (you didn't really think your drives had 64 or 255 heads, did you?), just to ensure that none of the numbers got too big to store in the fields they were originally given. The drive and/or BIOS and/or OS would translate between these schemes in a messy and fragile combination that could (and sometimes did) lead to "misunderstandings" and corrupted file systems. Thankfully, roughly around when OSes went 32-bit, we had LBA -- a much more straightforward "0 to nth sector" model that we've used ever since.
its adorable
Adrian Black has a PC test bios for memory tests.
I have usually wired these speed displays to show Hi/Lo rather than actual speed in MHz.
Synthesizers/keyboards/cnc mills had black floppies. Thats why you can get black goteks.
Black floppy drives definitely existed on PCs in the 2000s, silver too!
That Cardinal sound card needs all the SMD caps replaced, the top right corner is all covered in cap juice
That battery hadn't gone apocalyptic yet. If you see visible crystallization on the battery can then it was leaking fast or for a long time. It gets out through the end cap of the can as the seal fails, then the electrolyte spreads over the board and up into the slots by capillary action, like a wick. The seal material isn't immune to bases so it eventually just gets eaten through. If you were lucky none of the broken traces are directly under slots or chips you'd need to desolder.
When I was a kid around 1992 or 93 I had a friend who had an old 386 with a CDR almost just like that +2 speakers built into the case more than likely that header is for something like that I think I saw a mosfet on that card. I still remember playing the DOS version of MegaMan on that thing.
Back in the 486 era, I had three VLB cards in my computer: A dedicated VGA card, a "Super IO" card like you have (but without VGA), and a SCSI card.
The fact that the errors are inconsistent coupled with the MS DOS RAM extender error when it craps out of Windows leads me to guess that it's a RAM issue, or more specifically a problem on the motherboard side of the RAM equation - the corrosion was quite close to the memory so there may be some damage hidden under the slots, and as the BIOS RAM check is extremely crude and only does the most perfunctory of testing it's not uncommon to pass the check and then run into issues when you actually try to *use* the memory. A diagnostic ROM and/or only populating some of the slots may help narrow it down.
I went over to the used computer store, popped a plain ISA IDE controller into it, and all the problems went away, it works perfectly now. Then I put a VLB video card in it, it didn't start up the first time, I wiggled it, then it started but there was corrupted graphics. So the VLB port is definitely messed up, and I did some further thinking - the reason this card has a controller chip instead of just being a bridge over to the ISA bus is because that controller chip is actually going through the VLB, I confirmed this by looking at the design of another card I found that has both VLB and ISA IDE ports on the same card. So, since we know the VLB is messed up, and the hard drive was going through the VLB rather than the ISA, I'm pretty sure that was why it was having all these problems, it was loading corrupted data from the drive.
@@CathodeRayDudeGaiden Yeah the fact that the stack is all zeros to me indicates that the CPU is probably executing a bunch of garbage instructions. Which would happen if during load time the text section/COMDAT of the image gets corrupted.
In other words, most definitely in-flight corruption during load of the CPU instructions.
@@LunaTheFoxgirl Battery corrosion can definitely affect VLB. I cleaned up a 486 board that very luckily only had slight battery leakage, but VLB doesn’t work because some traces for it run directly underneath the battery. ISA works fine because those traces aren’t doing that whole maneuver around the board like VLB tends to.
One brand of battery notorious for leaking when it is found in computers is Varta. Varta is a German brand so that surprises me somewhat. I can imagine some dirt cheap no name Chinese battery leaking more realistically but the fact is the Varta CMOS batteries were known for leaking. Is the cause of that the same as for alkaline cells that got left in things too long and leaked? I am sure that some of us had a toy we liked for a while and forgot about it while the batteries were still in it then maybe a decade later we find it and are disappointed by finding that has happened. Especially since often it does not work any more since the stuff that leaked is caustic.
There is definitely more life in this machine.
There were a few DIY -NES cart-dumpers- eeprom programmers that (in single/pure DOS mode) used the DMA/IDE interface to generate (and read) the signals by accessing the relevant addresses. --
edit: I forget if later versionsgot around the clock-speed/timing issue by tapping one of the address lines ; but now I wonder if the interface card(s) bundled with so many grey-market cart/rom dumpers is(/were) a repurposed CD-ISA bridge/address decoder.
Matsushita is the name of the company and Panasonic is just the North American/international brand name if I remember correctly
happy new year 2 the both of you i hope you have a fantastic year
I honestly didn’t know that ATAPI hadn’t always been a thing. I used SCSI on my Amiga, and by the time I switched to PC in the late 90s, I guess all the dust had already settled.
when the PCAT is bby
Good job getting her working to a certain point!
Many early drives did come with controllers. Sound cards added interfaces to save a slot.
I too am a baby at teardown & repair
Damn! I need that mechanical loading CD-ROM drive in my life! The CPU and FPU are also very special and so is the 386 mobo with VLB. It's a dream system. Please save it.
Vinegar will help neutralize alkaline battery guts, just thoroughly rinse it off within a few minutes or it will damage other components. Deoxit is safer if you have it. Do be careful cleaning slot pins. They are way thinner than you think. Unfortunately, they put the RAM near the corrosion generator. Random crashes may be damaged traces near the RAM. If it's a multi-layer PCB, fixing them may be impossible.
Unfortunately we are loosing history day by day from Varta destruction and cap plague. At some point we need to start preserving every 8080, 88, 86, 286, 386, and 486 (and into the Pentium era) regardless of how common they were during production. Working units are becoming rare now and are sliding towards extinction.
What’s “Varta destruction”? My Google-jutsu it’s failing me today and can’t find anything useful.
I have a very similar VGA+Multifunction combo card that uses the same chips I can confirm that for VGA it's about as fast as you can go for 486 video without having to go to PCI. It's not the fastest VLB Video card but it's pretty damn close and you're more likely to be CPU Bound instead. Your suspicions about the promise chip are correct - That is a VLB IDE controller and is much faster than an ISA IDE "controller" (That as you mentioned, really just shoves the IDE bus onto the ISA bus)
I too have used it in a 386/486 combo board, although I have a different model (Although mine also came with a TI 486DLC chip, on a motherboard with a soldered on 387? These late 386/486 boards are wild. An expensive math co processor just soldered to the board, and you can override that one by slapping one into the 387 socket) These boards are kind of rare, and they can be extremely fast 386 boards or kind of fast 486 boards. Yours has 256kb of cache which is mighty impressive for a system like this. Someone cost engineered the shit out of this machine to be as much of a multimedia machine as you could get without doing the right thing and putting a 486 into it.
If you pull that 387 chip you can put a 486 in that socket and get a bit more speed out of the board. You may also need to remove the TI 486DLC (Which is just a 486 core in a 386DX pin pattern, so you lose a little speed when going back to the 386 DX bus system compared to the native 486 Bus) to make this work.
As for VLB stability issues. Try the other slot. These 386/486 combo boards typically only support VLB in both slots if you're running a 486 CPU, and if you're running a 386 CPU only one of the slots is officially supposed to work (Which is why you have a combo VGA/IDE board - your 386 Class CPU only runs one VLB slot so you have to make the most of it) VLB cards are wildly sensitive to signaling issues, and these 386/486 combo boards have to do some Deep Magic to make VLB run on a 386, even if it is very fast on a 386. You can also try lowering your CPU clock speed to 33mhz, which is probably safer on a board that has corrosion issues.
The corrosion on this board doesn't look that bad. You could probably save it from more of an early death with a good long bath in some IPA and a bit of scrubbing with a tooth brush.
Seem to remember that the Amiga a500 cd rom (a570?) came with a mechanical eject caddy cdrom drive! Also regarding the board refurb, do watch some of Adrian digital basement… ;)
The PnP Sound Blasters still had support for the three CD-ROM protocols. I have one! A later model than mine replaces all of them with “IDE” (ATAPI?), before it goes away entirely. I guess it was still worth including for 386/486/early Pentium users who wanted to add CD-ROM.
Maybe you mentioned it and I didn’t realise, VLB isn’t strictly for graphics. My 486 has a VLB I/O card, no graphics on that. It definitely uses it for the IDE ports, which I know because (luckily very minor) battery leakage damaged VLB traces that ran directly underneath the battery.
MR. BIOS was very impressive, it has a very fast boot mode. I flashed it on my Pentium board and it definitely is fast. Has an early in-BIOS boot menu too. And yes, it calls itself MR. BIOS on boot!
Good to know on the PnP ones, I don't have many that have the interface so I wasn't sure. Re: VLB, yep, the card I have in this video uses it for all that stuff! I also found one at the REPC today which has a REALLY interesting layout: two IDE ports, one labeled "ISA IDE", one labeled "VL IDE", and you can see how one goes through a host controller and one doesn't!
We called those "multi-io" cards.
Nice! A caddy cd-rom drive is my white whale device, they are just hopeless to find here in the nordics.
17:38 I've recently seen those backro-named to MAin or Main Attachment and SL= Secondary Link. A little clunky, but a good way to get some un-wanted baggage out of IT terminology.
Cool. If your not going to rehab the board. Maybe see Adrian if interested in it
I might! I've since done some more work and gotten it to run Windows reliably, but the VLB port definitely needs to be replaced and that trace repaired for the thing to work perfectly, and I don't really need another machine that's not super reliable, so maybe it would be better if he had the board and I just built a different machine in this case
Oh interesting, I had this Case/MB/CPU combo back in the mid 90s as a hand-me-down, but with a beige case. The MB seems in bad shape, but please don't throw it away or scrap it. Maybe you noticed that the 486 is smaller than a usual 486. That's because that's a socket for a 386. Yes. This board can run either a 386/387 combo or a 486 in the socket around the 387's socket. And yes, it can use VLB with the 386 (not to its benefit of course). Beautifully weird design :)
If you *really* want to throw it out, let me know and I might want to pay for shipping to my "recycling facility". I've probably had enough battery corroded PCB repairs in my life, but for this gem, I'd do it again ;)
Yeah I've been getting a lot of comments to the effect that this is a very unusual and interesting board. Unfortunately with the damage it'll never a great testbed machine for me, but I definitely don't intend to trash it, don't worry. Once I decide what to do with it I'll probably scrub through these comments again and see who gets dibs :p
@@CathodeRayDudeGaiden Understandable. Sounds like a good plan! Thanks for taking the extra effort! :)
I don't know, but I wonder if that ROM socket on the sound card might be sample ROM? As you said, it does have an optional SCSI interface, so it COULD be boot ROM... but the SCSI chipset footprint looks like the old NCR 5380 or a derivative thereof. These were used on the OG Mac and later on sound cards like the Pro Audio Spectrum. You COULD add a hard drive to those, but they were not fast. They were low cost and optimized for low-speed peripherals like (early) CD-ROMs. You probably wouldn't want to boot from that interface.
requisite terrifying Amiga fact: you can, in fact, put four CDROMs on one IDE interface. There's a kludge board for that. Don't do it.
so yeah, about that! i have this DIM memory floating around in my head that i've seen four-drive IDE chains and i could not remember where the hell it came from. maybe that's it??
I'm not sure how that could possibly happen. The select lines only allow for two combinations. If there's any way to get four discrete drives to function on the same bus, it would have to be a purely software trick that overrides the hardware addressing conflict somehow. I don't know enough about the lowest levels of PC interface programming to say whether that's even remotely possible, but I kinda doubt it.
If such a thing does actually exist, I would be very curious to see how it pulled it off. I would assume it had to augment the IDE interface to add another select bit (2 bits = 4 combinations.)
the pc speaker header is an INPUT, it mixes the pc speaker from the motherboard header through the sound card instead. You could donate the mainboard to Adrian black for restoration, i'm sure he could make a good video out of it and it would sate your curiosity.
OH DUH. i should have known that.
shame that board is toast. those hybrid 386/486 boards are pretty interesting. maybe with some vinegar it can be rescued
edit: ok so a bit of info, that 486 installed is actually a TI 486 SLC, that is, a 486 upgrade for a 386 motherboard. so that slot is for a 386 class cpu. what's interesting is the math copro socket, which is waay bigger than it needs to be. that can actually take a full 486 cpu. so it's a nice board, usually opti stuff is reliable, and vlb 386 is kinda interesting, while mostly useless lol
I’ve managed to get slots working by spraying contact cleaner into the slot and then just rocking the card back and forth to scrape off any crap.
Adrian Black from Adrian’s Digital Basement swears by DeOxit, but that’s not available in the UK so I personally swear by WD-40 contact cleaner. (Not regular WD-40 though!!)
Thanks Gravis ❤
That's a nice COCKWELL HARD MODEM you got there!
Would be pretty cool if you could get it rehabbed in all of its terrible glory. And then maybe max it out. Who knows, we can throw a few bucks into the beer fund.
That Cardex Video/IO combo cards were not all that common. That is dram for video. 386 board with VLB is rare. The board is salvageable with work.
6:42 _«…Now, in _*_this_*_ case…»_ explaining a particular case for a baby AT case, pointing at the case… 😅
(All the best for 2025…!)
Lime juice? that motherboard must smell nice.
The fact windows is hung but CTRL+ALT+DEL works is interesting because that means it wasn't some hardware fault necessarily. The bus and memory and CPU are still running and not halted or clobbered too badly, such that a lot of systems within the software are still working. It was easy for windows 3.x to be misconfigured and just fail to boot. Could be some driver just waiting forever for some device which isn't present.
Weirdly, even with all those other errors, it's still not totally hanging the system. I can't entirely imagine what they might be. Maybe bad/unreliable RAM would be my first guess. Maybe some unreliable sectors on the disk occasionally misreading a bit here and there. But could even be some deeper issue with something just being a little unreliable with memory accesses or IO with the disks or maybe some other devices which windows is trying to initialize.
It really did smell nice!!!
Sometimes it WOULD hang hard, other times it would just give me a blinking cursor but C+A+D worked, etc etc - all moot now. I pulled the VLB card, put a plain ISA IDE controller in there + an ISA VGA card, and all the problems went away, machine worked perfectly. So I'm thinking it's the combo card spitting noise onto the bus when it wiggles due to the wrecked socket, causing totally stochastic failures.
Try to boot the Cardinal board with a Msystems DiskOnChip inserted, see what it does.
2 CHAINS!!!!! LMFAO
I always hated fixing BabyAT computers. So many wires.
If both the Matsushita interface and IDE are 'just ISA', does that mean it would be possible to adapt a Matsushita drive to an IDE bus? Obviously it's still not going to be speaking ATAPI so you're still going to need weird drivers for it.
"All the interfaces were proprietary."
"Drives didn't come with interfaces, I suppose..."
Not entirely correct. SCSI was a thing by then, and lots of early CD-ROM drives _were_ proper SCSI drives. In fact, that Matsushita 2x drive (the 562/563) actually did have a 2x SCSI variant. Lots of other ancient drives, from NEC and Ricoh and Yamaha, e.g., stuck to SCSI.
The problem with that is, of course ... SCSI was super expensive for PC users. Unlike many contemporary platforms, the PC didn't have a native SCSI controller in the de-facto reference designs (originally, the IBM PC, then the Compaq 386 -- after that, everything was a gentleman's agreement), so you were required to source your own. Adaptec and Future Domain were the usual go-to options, but it was often hundreds on top of the PC and the expensive CD-ROM drive itself.
The "Big Three" proprietary interfaces (which, BTW, leaves out Phillips' LMSI) were designed to be low-cost alternatives, and took the same approach as IDE: Make the controller little more than an extension of the ISA bus, and place it on the drive so it can be as minimal and purpose-built as possible. The add-in card itself was really just some address selection logic and a few buffers. Perhaps it also allowed changing the address, and selecting whether or not to map an IRQ or DMA channel. Oh, and of course it often allowed you to adapt the audio cable to RCA or 3.5mm phone jacks.
This simplicity is why they were so easy to incorporate into sound cards. With a tiny bit of glue logic and some off-the-shelf buffers, you could add any or all three interfaces with very little effort. The pinout isn't even that much different -- just the same signals in a different order, mostly. The "secret sauce" is the command set they use, which is opaque to the hardware in between.
When ATAPI came out, it allowed packaging SCSI commands in packets that could be transferred via the IDE interface, which took existing (SCSI) software and combined it with a trivial update to the bus interface we all already had, making it the shortest possible path to a complete solution.
As to whether the drives came with interface cards -- yes, in fact, they did ... at first. Some of the original drives shipped with their bespoke controller, usually proprietary like LMSI (a very early solution.) SCSI drives often did not, since e.g., Mac (and other) users wouldn't need it, and the PC didn't have overwhelming market share until closer to 95, so that mattered. But, it wasn't unheard of to see drives in a bundle with a SCSI card targeted at PC users. Future Domain specialized in this market as a cheaper alternative to Adaptec's higher-end cards. I suspect it may have been a coping strategy to get them separately, since each would set you back a month's rent, and it was probably easier to stomach one at a time. ;-)
A little later, in the first half of the 90s, the "Multimedia Kit" was a big deal, since multimedia was a huge driver to both CD-ROM and sound card purchases. These were really pricey, but would usually include the sound card, CD-ROM, some cheap desktop speakers, a mic, and a bundle of software that was definitely going to include Mavis Beacon Teaches Typing and an encyclopedia or human biology reference disc. Perhaps a kids' storybook title like Grandma and Me.
Once it was potentially likely that you *had* a sound card already (but not necessarily a CD-ROM), then they started shipping with the cheapest possible ISA card -- MKE, Mitsumi, and very early ATAPI IDE being the most common. (The IDE card would often be configurable as secondary or tertiary, since this was around the time where VLB controllers and on-motherboard interfaces were beginning to ship with two IDE connector interfaces instead of just one.) The CT1860 is one such card that I happen to have on my desk, which shipped with (e.g.) the Creative Omni CD or Blaster CD 4x (another Panasonic / Matsushita) drive kits.
Okay, well, you went on to discover a lot of that yourself -- which was inevitable, I guess. :-) Oh well, forgive the lengthy narrative and just assume it makes for better storytelling and confirmation of your assumptions.
As for the 34- and 40-pin query...
Sony's was a very simple 8-bit interface -- which is why it can be a smaller connector. It only has D0 through D7 connected, plus corresponding grounds between the signal lines, of course, and a handful of control pins. That's it.
The MKE interface is 8-bit too (not sure if there was ever a 16-bit version), but it has 8 extra lines that I haven't figured out yet. They're all connected to something, but the board that I've personally RE'd (the CT1810) has a GAL IC that I haven't yet fully mapped out. It looks like some of it is address logic, or at least depends on the selected address. Four other lines are just buffered from a pull-up resistor pack, and muxed with D3..0 when enabled by the GAL. I'unno. 🤷
I have at least one Mitsumi interface card around here somewhere (and IIRC, it's also 8-bit), but it doesn't appear that I've made a schematic of it yet. (It's much easier to trace them when you're not concerned about potentially destroying the PCB in the process by desoldering everything, and I prefer to take the opportunity when the board is a little worse for wear or doesn't work anyway.) I know it's similar to the other two, but whether it could've been a 34-pin connector too, I'm not sure.
A lot of this is probably long since decoded by others, but I like reinventing wheels, so I haven't gone out of my way to spoil the puzzle for myself.
"isn't worthy of a full video"
that seems like a full video
"full" For me means that I wrote a script, made sure there weren't any apparent errors, and made the diction clean enough that a decent proportion of the viewers on my main channel will want to watch it all the way through instead of losing interest a couple minutes in. If I thought it wasn't worthy of a video at all there wouldn't be a video :p But my normal standards pretty much don't allow me to make a video in less than a week, whereas this one took 3 hours
My money's on RAM. Possibly contact with the sockets.
Happy New Year
Ive never seen that case in black. ive owned it in beige.
If I could ask, could you please back up that Mr Bios rom? There doesn't seem to be an entry for this board in the Mr Bios archive.
It's not actually a bios, turns out it's just a keyboard controller
Oh, I just realized that's a keyboard controller.
If you don’t want to repair it, another TH-camr called necroware might be interested in giving it a go
Yep that is a very late 386DX motherboard with VLB made for these Cyrix/Ti 486DLC CPUs, not as common as a garden variety 486 motherboard.
_Kotobuki_ (寿) means “long life”.
39:05 Does Gravis not watch other TH-cam creators?
Was wondering that too. At least three popular channels do repairs on battery-bombed motherboards on the regular. Nobody should be left wondering what to do with these boards. :-)
Hark! The 'Ray Dude angels sing:
Glory to the reborn thing!
Peace on earth and reboots wild
God and Crockwell reconciled!
Coup de Gracie... 😂
the at in the video title should have been in caps AT
Yes we all know this if we’re here on the side channel watching
Fiberglass pens are trash. I use a tungsten pencil now
I love those "other 90% of the computer" cards like the 6 pack. I bet they solved a lot of slot related issues back when you needed a card for everything that wasn't one of the 30 standards. if you needed more slots I bet some people just dangled the ports out an empty slot or out of one of the knockouts instead of having a ton of brackets if they needed more ports.
Ct1750 ❤