As the title of this post may indicate, I had the misfortune of buying an ALR Optima SL which had one of these vile abominations built into it. The motherboard appears to be an Acer V12LC, so I flashed it with the latest Acer BIOS and set the BIOS type jumper to non-OEM, and I also lifted pin 85 of the PCI0640B IDE controller chip to work around a stupid design mistake, so now hopefully simultaneous floppy and HDD I/O won't cause it to choke up at least.

The Acer V12LC motherboard used by the system has a jumper for EIDE Enable/Disable. Setting this jumper to Disabled doesn't appear to disable the CMD controller, I suspect it does nothing (though it COULD control the flawed IDE prefetch of the CMD chip... wishful thinking, I know 🙄). The only way to get reliable fast IDE support is to use a PCI IDE controller. Amazingly this board fully supports my Maxtor-branded Promise SATA/150 controller, but there's one little problem: the board only has Windows 98 drivers at the earliest, and the system only has 16MB of memory (and I just recycled all my non-EDO 72-pin simms, natch!), so Windows 98 is going to be a very painful endeavor unless I absolutely strip it to the bones using 98lite.

I'd like to run the far lighter-footed Windows 95, but I'd rather not buy a different IDE controller if I can help it. The VIA controllers have their own host of annoying quirks and gotchas, and the Windows 95-supported Promise Ultra/33 controller jumped way freakin' up in price on eBay. I ain't worried about the IDE CD-ROM drive because I can always just run that to the IDE channel on my sound card.
I suspect though once I have something that works, I could possibly just pull the sh!tty CMD IDE controller out using some removal alloy and introduce it to Mr. Hammer Time.

....Oooor maybe would it be possible to swap the PCI0640B for a PCI0646? IIRC the PCI0646 fixes the stupid design shortcomings that cause data corruption, though they themselves are far from perfect. There are a few pins that were changed, mostly relating to DMA transfers and bus-mastering, so I'm wondering if it'd be possible to leave the respective pins floating so I can just get corruption-free PIO transfers out of it at least.