I would test with another optical drive.

Chadti99 wrote on 2021-07-29, 11:56:

I would test with another optical drive.

and a cable as well; if it works try the optical drive giving issues on another pc as a possible fail.

As I stated in the first post I tried two different DVD-ROMs and two different IDE cables. A third one would make a difference? I'll try.

Also the issue started with another DVD-ROM drive, this is the second one I'm trying out. I may swap it out for a 52x CD-ROM and see if that behaves any better.

appiah4 wrote on 2021-07-29, 12:23:

As I stated in the first post I tried two different DVD-ROMs and two different IDE cables. A third one would make a difference? I'll try.

Also the issue started with another DVD-ROM drive, this is the second one I'm trying out. I may swap it out for a 52x CD-ROM and see if that behaves any better.

Yes exactly, a different type of optical drive, I’ve only tested older <24x drives on my board but not a DVD. I’ll see if I can test that though.

Have you made a point of testing with and without a disk in it? I am wondering if perhaps when booting with a disk in it it, it takes too long for a higher rotational speed DVD drive to come up to speed vs slower CD ROMs and thus it faults on not being ready.

Chadti99 wrote on 2021-07-29, 19:15:

Yes exactly, a different type of optical drive, I’ve only tested older <24x drives on my board but not a DVD. I’ll see if I can test that though.

Interesting. I don't know if I have any spare CD-ROMs left lying about though. Time to shop around for one..

BitWrangler wrote on 2021-07-29, 19:22:

Have you made a point of testing with and without a disk in it? I am wondering if perhaps when booting with a disk in it it, it takes too long for a higher rotational speed DVD drive to come up to speed vs slower CD ROMs and thus it faults on not being ready.

Umm.. I don't think there are any CDs in the drive at all. That said, the detection routine for the DVD-ROM may still be taking longer than a regular CD-ROM or something.

How onerous is the memory check on that board? If it goes Bzzzzzzt through it it might not be much bother to have it on and see if the delay helps, but if it's the slow tick tick tick type then would be a pain in the butt. Also being floppyless shortens the time it's got by the length of the blurp-blurp floppy seek. Usually there's milliseconds in it, something as stupid as having the logo on or off or using one keyboard rather than another can affect it. If network boot is an option, putting that up the priority list so it checks for the ROM and connection is an extra delay too.

I've done a shitload of troubleshooting tonight.

Things that didn't work out:

-Replace the DVD-ROM with a known working CD-ROM drive; did nothing.
-Move the CD-ROM to Secondary IDE; When set to Master, it completely messed up HDD detection for the CF-IDE on Primary Master. When set to Secondary, it was detected once, but never again.
-Try a different known working CF-IDE adapter; did nothing.
-Try three different cables, with the CD-ROM as Primary or Secondary slave: did nothing
-Play around with memory timings, change the PCI divider; did nothing
-Move PCI cards around; sometimes caused the HDD to not get detected. Strangely enough, removing the Ethernet card resulted in the hard drive not getting detected when IDE Controller was set to PCI Slot 1 (which is empty) which led me to think the PnP BIOS is fucking up some kind of resource allocation.. So:
-Play around with PCI Interrupt settings: There's something there. I mean, when set to manual and I play around with IDE Controller PCI slot # and/or IRQ allocation priorities, or INT#s I can definitely fuck up HDD detection, but I can't get the CD-ROM detected whatever I do.

So what did I do? I threw the towel and removed the CD-ROM completely. I mean, I actually do need a CD-ROM for this PC but I can live without it for a while. I know I have a loose PCI ATA controller in storage so tomorrow I will dig it up, install it and see if that makes a difference. Something's incredibly wonky about this board's IDE controller, and I can't work around it.

Have you tested with a regular HDD instead of the CF-IDE? (Or with just the optical drive on the IDE controller loading the drivers from a bootable floppy if you don't have spare HDDs to test). I've encountered several instances of CF-IDE adapters that work just fine on their own but they mess up other IDE devices and in some cases even the floppy controller.

Regarding the BIOS stuff, if you tweak the PCI settings manually you may end up with the IDE controller IRQs assigned to one of the slots, which can cause a conflict. I would leave it on AUTO with only the graphics card installed while you test. Interestingly enough, I own a Gigabyte GA-486AM/S (same chipset as your board) where the 3rd PCI slot is marked as "DISPLAY" so which slot you use seems to matter in some way.

I don't have any unused HDDs that are within the size limitations of this board, I think; so no, I did not try with a real HDD.

Actually, I do have a 534MB Seagate Medalist I can try next week... Hmm....

I should try and find a <8GB HDD for this build, maybe.

What jumpers do you have on the HDD? If you've got Master/single, slave, and CS, try putting the CD in as slave, and jumpering both master/single AND slave (Which used to be called out as slave present) on the HDD. This used to be called out in install instructions in the 90s, as sometimes needed on some installs if the controller, HDD and slave drive didn't all speak quite the same dialect of ATAPI.

appiah4 wrote on 2021-08-02, 21:54:

I don't have any unused HDDs that are within the size limitations of this board, I think; so no, I did not try with a real HDD.

Actually, I do have a 534MB Seagate Medalist I can try next week... Hmm....

I should try and find a <8GB HDD for this build, maybe.

If the drive is bigger than 8.4GB, the board will happily go along with 16383/16/63 as CHS settings and ignore the rest. It won't hang like the newer Award BIOSes with the 32GB limit.

BitWrangler wrote on 2021-08-03, 00:11:

What jumpers do you have on the HDD? If you've got Master/single, slave, and CS, try putting the CD in as slave, and jumpering both master/single AND slave (Which used to be called out as slave present) on the HDD. This used to be called out in install instructions in the 90s, as sometimes needed on some installs if the controller, HDD and slave drive didn't all speak quite the same dialect of ATAPI.

The "HDD" is a CF-IDE, it can only be set ass Master (jumper closed) or Slave (open). I have a feeling this board's native IDE controller misbehaves with CF-IDEs. I think feiopa already documented CF-IDE's not really working as intended in his manual unless you set them to PIO Mode 4. I think they don't work quite right even then. So my list of remaining troubleshooting steps:

1. Try with a HDD+CD-ROM combination (both on the same or seperate chains)
2. Try with a PCI ATA controller

I'll report back later.

appiah4 wrote on 2021-08-03, 06:24:

1. Try with a HDD+CD-ROM combination (both on the same or seperate chains)

1 is a partial success. With a Quantum 40GB Fireball as Master and the DVD-ROM as slave, I can set the capacity limit jumper and it gets detected as an 8GB LBA HDD (and if I don't set it it gets detected as an 8.4GB CHS one). This results in a POST and the DOS 6.22 boot floppy detects and sets up the DVD-ROM drive perfectly anytime.

However, partitioning the drive was a colossal failure. FDISK saw Non-DOS partitions (NTFS) that were in negative sizes, and after deleting these I tried to set up a new primary partition which caused a lock up. Upon boot, the BIOS would still detect the HDD fine, but DOS or Win9x FDISK would fail to load, giving me a "no fixed disks present" error. I guess I would need to set this drive up as CHS 8.4GB on a newer computer and do the partitions there before moving it to my 5x86.

Regardlessi out of curiosity I went back to the CF-IDE and now it also detects the DVD-ROM perfectly every time. The BIOS settings that I changed are:

PnP BIOS Aut-Config Disabled
1st Available IRQ 7
1st Available IRQ 9
1st Available IRQ 10
1st Available IRQ 11
Slot 1 Using INT# AUTO
Slot 2 Using INT# AUTO
Slot 3 Using INT# AUTO
PCI IRQ Activated By Level
PCI IDE Controller Enabled
PCI IDE IRQ Map To PCI-Auto
Primary IDE INT# A
Secondary IDE INT# B
IDE HDD Block Mode Enabled
IDE Primary Master PIO Mode 4
IDE Primary Slave PIO Auto
IDE Secondary Master PIO Auto
IDE Secondary Slave PIO Auto
Assign IRQ For VGA Enabled

With these settings it seems to be working. This is the weirdest, quirkiest IDE controller I've seen so far, probably has to do with its early PCI Bus implementation..