So...

What's next:

I have a floppy drive (unproven), a set of shrink-wrapped DOS 6.22 floppies, but no floppy cable 🙁

I also have a CF IDE adapter and an 8gb CF card but I guess that's no use until I get DOS installed ?

I also have a CD of Ultimate DOOM 😁 and rather than getting a CD drive was thinking of copying the disc contents onto the CF drive and installing it from there ?

Been a looooooong time since I played with kit of this vintage !

Btw, will the IDE ATAPI interface support any IDE CD drive, regardless of how modern it is ? (I've seen some E-IDE CD drives which I assume would not be supported by old hardware ?)

Do you need a specific driver for each model of CD drive or is there a generic driver that will work for all drives ((VIDE-CDD.SYS for example) ?

Gary.H wrote on 2022-04-10, 14:08:

Btw, will the IDE ATAPI interface support any IDE CD drive, regardless of how modern it is ? (I've seen some E-IDE CD drives which I assume would not be supported by old hardware ?)

Do you need a specific driver for each model of CD drive or is there a generic driver that will work for all drives ((VIDE-CDD.SYS for example) ?

Any standard IDE interface (i.e. everything that's not on absurdely expensive cache controllers) will accept CD drives as well as hard disks. For any standard IDE CD drive (i.e. most 4x and all 6x or faster drives), a generic driver will work fine. There is only one kind of ATAPI standard. There are a couple of early revision CD drives that can be connected to an IDE port but were manufactured before the ATAPI standard was agreed upon - these drives use some status bits differently and might be incompatible with generic drivers.

"E-IDE" never was an officially defined standard, but the enhancements (mostly faster cycles) were formalized in the ATA standard. E-IDE improvements like PIO mode 3 and 4, or LBA addressing went into the ATA standard before ATAPI was standardized, so it makes no sense to differentiate between "IDE" and "E-IDE" CD drives.

If you want high speed CD data transfer in DOS on a PCI machine, you might want to use a DMA-capable CD driver. The choice of the driver then depends on the IDE chipset you are using, not on the CD drive.

For setting up almost identical system without floppy I installed dos in virtual machine onto sd card and then used it to boot successfully via sd2ide adapter. Should work the same with CF card.

Also keep in mind that the bios likely won't detect the cdrom drive, but it will function just fine in dos/windows. Same goes for zip drives and all other atapi devices.

Some days you're the pigeon, some days you're the statue.

Today I'm the statue...

Hard to know whether it's the floppy drive or the (still sealed until this afternoon) DOS 6.22 installation floppies that are at fault ?

Bah - replacement floppy drive arrived ...

...exactly the same symptoms 🙁

New DOS 6.22 floppies next

DOS 6.2 floppies arrived today.

Hadn't noticed before but my BIOS boot screen does only show 256kb of base memory !

Gary.H wrote on 2022-05-03, 17:42:

Hadn't noticed before but my BIOS boot screen does only show 256kb of base memory !

Your POST showed correct 640KB of base memory just three weeks ago; might want to use MemTest86 4.3.7 to test your RAM.

I don't think AMI's BIOS of 1994 supports ATAPI CD-ROM startup (USB is out of question for sure) so you probably have to create a bootable floppy on another machine.

IMHO 64MB (the cacheable upper limit for 430FX, VX, and TX chipsets) on a 1994 vintage 486 is a bit too much; 16MB is more than enough and would cause fewer problems for many programs.

64 MB in a 486 with standard 256 kB of L2 cache should be used in write-through-mode (8+0).
To achieve full performance, reduce DRAM to 32 MB and set L2 cache to write-back-mode (7+1).

I took out 32mb of the RAM, leaving 32mb and now it reports 640k

Subsequent boots to the DOS 6.2 floppy 2 have resulted in 2 hangs at 'Starting MS-DOS' and one with a load of unprintable characters on the screen followed by an A:> prompt

Gary.H wrote on 2022-05-03, 18:05:

I took out 32mb of the RAM, leaving 32mb and now it reports 640k

Subsequent boots to the DOS 6.2 floppy 2 have resulted in 2 hangs at 'Starting MS-DOS' and one with a load of unprintable characters on the screen followed by an A:> prompt

Switch to another 32MB stick and see if "unprintable characters" reappear. You really need to check your RAM with MemTest86.

Faulty RAM sticks can destroy your BIOS and brick your MB. I'm not kidding; two of my Gigabyte GA-K8VM800M were bricked by a single faulty 512MB DDR stick.

Two different memory sticks 4mb and 8mb both result in hangs at Starting MS-DOS with the FDD light staying on but with no drive activity noise.

How do I test the memory if I can't even boot to DOS though ?

Have you tried both modules solo in each RAM slot?

I think by now I've tried every RAM stick in every slot 😉

My replacement DOS setup disk 1 now won't read anymore.

Managed to somehow run the setup exe off the previous DOS setup disk 😀...at which point it refused to recognise either my SD or CF IDE emulator cards as a valid IDE drive 🙁

Beginning to think this board is a bit foobar'd

Oh and most of the time the CPU now comes up as "Am (unknown)" on the POST screen 🙁

Tried a '96 Seagate 1.2gb IDE drive and although the BIOS sees it (as it did the CF and SD IDE emulator boards) the DOS setup program doesn't see the drive 🙁

(Second) floppy drive is also very temperamental, getting a lot of Abort, Retry, Failure errors.

Tbh, everything seems super-sketchy at the moment on this board .

Hmmmm

Presumably since my Dallas RTC battery is flat so I can't save any BIOS settings does that mean that when it boots (after leaving the BIOS screen) the machine won't see any of the BIOS changes I've made ? (including the successful Auto IDE setup)

Oh, well.
You may expect strange behaviour of any motherboard with an empty Dallas.

Wondering whether to go to the bother of replacing the Dallas if I have other flakey behaviour like the intermittent failure to correctly identify the cpu and increasing floppy (on board controller) read failures ?

Guess I won't know the true state of the board until the Dallas is replaced ?

An empty Dallas chip shouldn't show the symptoms you describe. As long as +5V is present, an empty Dallas chip is working just fine. It will still report the flat battery which might upset the BIOS slightly, but the symptoms would be way more gentle. With a non-defective flat Dallas, you would be able to save BIOS settings.

Your symptoms sound very much like a broken Dallas chip, though. I had it just the other day with a Dallas chip that was accidentally plugged in backwards for some seconds. I got lost CMOS settings, random CPU type display and oftentimes plain refusal to warm boot. So it likely doesn't make sense to try to attach a different battery to your chip, but replacing the complete package, as you intended anyway is the right way forward.