You have two pieces of software that are known to conflict with each other:
- HDD code in the AT BIOS
- BIOS extension on the XT controller

As you already know, you can disable the latter - but the AT BIOS only supports AT-style controllers.

So, the solution is:
- disable the AT BIOS HDD code, by setting all HDDs to "none" in CMOS Setup
- enable the BIOS extension on the XT card

https://www.minuszerodegrees.net/hdd/HDD%20co … versus%20AT.png

You might be able to leave your secondary controller on to have something to dump the XT drive to, but it's probably going to want to boot from it. So might wanna have some tools on a boot floppy and boot that and deal with it.

Grzyb wrote on 2026-01-11, 02:54:

You have two pieces of software that are known to conflict with each other: - HDD code in the AT BIOS - BIOS extension on the XT […]
Show full quote

You have two pieces of software that are known to conflict with each other:
- HDD code in the AT BIOS
- BIOS extension on the XT controller

As you already know, you can disable the latter - but the AT BIOS only supports AT-style controllers.

So, the solution is:
- disable the AT BIOS HDD code, by setting all HDDs to "none" in CMOS Setup
- enable the BIOS extension on the XT card

https://www.minuszerodegrees.net/hdd/HDD%20co … versus%20AT.png

Tnx, disable HDD settings in AT BIOS did the trick: after quite a long wait the drive chugged and I was presented with a C> prompt and what looks like a clean install of Compaq PC DOS 3.31 (which actually sounds like an MS-DOS version...)

BitWrangler wrote on 2026-01-11, 04:58:

You might be able to leave your secondary controller on to have something to dump the XT drive to, but it's probably going to want to boot from it. So might wanna have some tools on a boot floppy and boot that and deal with it.

If there had been anything I needed to dump, I'd have fiddled around with it - although I'd have handled booting from floppy/Gotek. As it is, there's nothing on it worth dumping; Compaq DOS 3.31 is already on Winworld.

Though being Compaq DOS, the io.sys and msdos.sys might be unique to the machine it was installed for. Not much point saving them if you don't have that machine and can't figure out what it was.

Compaq and other early clones were not quite 100% BIOS and hardware compatible with IBM PC and needed those file to patch the last 0.1% or whatever it was to be 100% DOS compatible.

BitWrangler wrote on 2026-01-11, 15:13:

Though being Compaq DOS, the io.sys and msdos.sys might be unique to the machine it was installed for. Not much point saving them if you don't have that machine and can't figure out what it was.

Compaq and other early clones were not quite 100% BIOS and hardware compatible with IBM PC and needed those file to patch the last 0.1% or whatever it was to be 100% DOS compatible.

Compaq MS-DOS 3.31 is from 1988 - I believe all those "MS-DOS compatible, but not fully IBM compatible" oddballs were already extinct by then.

I've recently experimented with Compaq MS-DOS 3.31 on a generic 486, and it worked fine.

Compaq carried some quirks through for full Compaq backward compatibility.

AT clone BIOSes got smarter, it can probably handle Compaq and IBM modes of i/o.

However you might not have tested it to where you would have seen the difference, often it was in the serial and parallel port interrupts and behavior.

edit: Though Compaq was very close, there were some MSDOS compatibles that were miles away, figuratively. Then there were some others that were tiny amounts different like Compaq.

editII: even with this, there was a point in time, the IBMs that followed on from the original AT, early PS/2, where Compaq looked more "IBM Compatible" than IBM did due to their quirks.