The board is an Opti 895 1442G. I can find lots of information about "Opti 895", but haven't seen a BIOS that says it's specifically for 1442G. I assume I do have to find one that is specific to this model, right? It wouldn't just be any old Opti 895 bios?

Huh, okay. So now it's just acting weird. I turned off internal and external cache. Since the XT-IDE didn't work, I removed it.

Now I can't access the CompactFlash at all, it seems. BIOS recognizes it (still sporadically, and often returning different values), but it sees something. And C: drive shows up. But format.com errors at the end, saying it was unable to write BOOT. And fdisk /mbr or fdisk just lock up. Clearhdd.exe returns 'error 80' and tells me to check my Borland C++ manual for help, 🤣.

I also set the memory settings to their highest numbers. I have two 32 mb sticks that I've tried for many of these tests, and their label at least says FPM. But for the last few days I've also been running many tests using a stick of 4 mb RAM, just in case. All of my results today were attempted with both the 4 mb stick and a 32 mb stick (separately, of course). The RAM didn't seem to make any difference.

I put the XT-IDE back in to see if maybe I could use it to re-format the CF card, but after 12 reboots, it has yet to actually see the card again. Kinda sad, like it was trying it's best at first, but now it's given up. 0_o

I think your board was actually made by Atrend and is a 1442G. OPTi only made the chipset.
If you have two floppy disks and can find a RAWRITE utility to write disk images to them, I wonder if you download and try to boot SmallLinux on this thing what it would do. Just for lack of any other ideas, booting into a 32-bit protected mode OS might cause it to crash in interesting or revealing ways. Or at least let you access the hard drive without going through the BIOS. https://sourceforge.net/projects/smalllinux/f … alllinux/0.7.5/ (download and extract the tar.gz file for the disk images)

Quick aside (I will try SmallLinux either later tonight or tomorrow), but my IDE-to-USB adapter came today and . . . no news.

The mechanical drive that crashes at 45% of DOS install and causes Scandisk to hang about 6 minutes into scanning? It refuses to spin up at all connected to the adapter. I thought maybe the molex power connector it came with was bad, but the other two mechanical drives that I got yesterday both spin up when connected. However, they are not found by windows at all. I tried slave, master, and cable select jumpers, because I wasn't perfectly sure what the adapter would want. But the two that aren't detected by my 486 spin up but aren't noticed by Windows, and the one that is detected and half-ass works on my 486 won't even spin up.

They all claim to be 5 volt drives.

Oh wait, the one that won't spin up says:

5VDC ===0.41A
12VDC === 0.11A

The two that spin up and stay invisible say:

5VDC ===0.41A
12VDC === 0.21A

Unfortunately, none of that seems like it helps my 486 issue, except maybe to imply that the two drives I got in the mail yesterday are, in fact, actually dead.

Edit: Yes that is definitely the board I have. Wow, I wish I had been able to figure that out. Could have found the manual online, that way, instead of buying it like a sucker. (Just 10 bucks but still. Hah.)

Does the USB to IDE adapter have a separate power supply for the 12 V? If not, you can try powering the drive separately from regular PC power supply.

Okay, so.

I found two other motherboards today. One was a piece of crap. Had dead roaches stuffed under RAM sticks, the CPU was forced in such a way that an entire row of pins were bent perpendicular to the CPU itself (you could see them sticking out from underneath the CPU!), was filthy in general. Cleaned it up real good, but couldn't get it to boot. Probably has to do with the fact I can't quite get a CPU to go all the way into the slot. None of the pins on the screwed up CPU were missing, so I don't think that's it, but there's definitely something wrong with the socket. Really I'm just bringing that up for fun, because yeah, didn't work. Or at least I didn't work with it long enough to try and figure it out. Wouldn't know where to start trying to repair a socket anyway, even if that is what needs to be done.

The other I got for a bit under 20 bucks. Unfortunately, it's a PC Chips board with fake cache. At least I assume these toy looking things soldered to the board next to labeled but missing jumpers are fake. With the VLB I/O board I was using, I couldn't get it to recognize anything over IDE. No flash cards, not even the mechanical drive that consistently at showed up on the Opti 895. After 7 or 8 reboots trying different things, it stopped POSTing at all. Just hung there, not even activating the screen or beeping. I eventually figured out that if the VLB I/O card was installed at all, it wouldn't POST.

I checked the slightly rusty ISA I/O board I have, the one that did even worse on the Opti 895 than the VLB I/O board did, and . . . it works. It recognized the 1 GB CompactFlash card immediately, let me format it, install DOS, etc.

So from what I can tell, unless I was missing a jumper somewhere (I don't think I was), the first motherboard was bad in a way that prevented either I/O card from working correctly with IDE, and the VLB I/O card was also broken, and has now gotten worse.

Well, DOS is installed now, at least! Too bad I can't figure out how to get either of my sound cards to work. (Have a Miro FM Sound card that the internet seems to say is the same as a MAD16, but the MAD16 drivers wouldn't work. Also have an AWE64 CT4500, which I installed the PnP software for, and it claims that it's set up. And Wolf3d recognizes and autoconfigures for a Soundblaster. But there is no actual sound. Hrmph.) But I have an OS installed! Definitely been a productive evening.

Thank you all for all your help.

I'll beat my face against the concrete wall of sound cards later.