500 MB, that's the BIOS limit of that machine likely.

To fix the booting, try fdisk /mbr from a DOS boot floppy!

Yes. Could be the drive is trying to configure itself as having Cyl > 1024 and Sect > 63 as for ATA-5 and the BIOS is truncating what's presented to DOS (or something along those lines) according to old Int 13h extended. What does the drive auto configure with for geometry? Maybe the drives work better for you with a higher head count and lower cylinder and sector counts?

Is LBA mode set by BIOS autodetection? If not, try set LBA manually.

If those doms work just fine in another board chances are the board you want to use doesn't like them which is odd but as peklop suggested give manual installation a try.

stamasd wrote:

So I got myself 2 industrial DOMs to see how well they would work in my retro systems. One 4GB and one 8GB. The only system that I have out of the box right now is a 486 that I was working on a little while ago. So I plugged them in one by one, and here's what happens.
The BIOS autodetects them with the correct capacity (3996MB for one, 8046MB for the other) and assigns each a geometry. Then I booted DOS 6.22 from floppy, but the DOS fdisk only partitions them in 507MB partitions. I can format those partitions, transfer DOS to them, but they don't boot. The same system works fine with regular HDDs and with CF cards in adapter, I don't have the problem with the 507MB partition and with the not booting.

What am I missing?

LBA support? The fact that BIOS can read cylinder count and calculate size for human consumption does not mean it can actually use disks larger than ~500MB. But OK, regular HDDs and CF cards work so it should have LBA support, but what if the DOM does not, and it must be used in "LARGE" translation mode instead of "LBA"?

Would you reveal us the DOM make/model or have a link to spec sheet so we can take a look at what their specs are?
Would you reveal us the geometry BIOS detects (physical), and the geometry seen by DOS (logical) by using MSD.EXE on the boot floppy?

I'd erase all partitions with fdisk, reboot, use "fdisk /MBR", reboot, check if now can generate over 500MB partitions, reboot, install DOS from floppy. At least this needs to be done when the geometry/translation mode changes.

OK here's more information.

The motherboard is an Acer VI15G rev.1.3 http://arvutimuuseum.ee/th99/m/A-B/32306.htm ; SiS85C471 chipset, AMI BIOS 12/15/93, 486DX/2-66 with 16MB RAM. This motherboard is a bit weird, it has a problem with rebooting (ctrl-alt-del doesn't reboot it, instead it locks and needs to be power cycled; same thing happens with programs that do a soft reboot such as exiting from fdisk; the CMOS battery is new so that's not the issue)

The first DOM is an Hyperdisk Industrial 4GB, model DMV340H4-004-M http://www.memorypartner.com/images/v/201301/13590103610.jpg
The BIOS autodetection gives it the following parameters:
type 47
cyl 7476
head 16
WP 65535
LZ 7476
sec 63
size 3859

I tried fdisk/mbr then reboot, fdisk again to delete the existing partition, reboot again; fdisk still tells me the total disk size is 504MB (not 507, that was a typo in my first post)

The second DOM is a Kingspec Industrial; no model mumber, just a series of ckeckboxes on the back; the "8GB" box is checked https://images-na.ssl-images-amazon.com/image … KL._SL1024_.jpg
(edit) seems to be this one: http://www.kingspec.com/products_detail/productId=34.html
Autodetection gives me the following:
type 47
cyl 15538
head 16
WP 65525
LZ 15538
sec 63
size 8020

The exact same thing happens as with the first DOM after fdisk/mbr, reboot etc. Size seen by DOS fdisk is 504MB.

(edit) for the 4GB DOM, msd.exe gives me the following info:
fixed disk type 47
1024 cylinders, 16 heads, 63 sectors/track
CMOS parameters 7476 cylinders, 16 heads, 63 sectors/track

Have you tried upgrading the BIOS? There's a version (2.3) from 1995. I haven't tried it myself, but it may solve your issue.

Can you use > 16 heads (and the reduce cylinder count). As you're using a DOM, geometry will be remapped anyway.

See here:
http://www.pcguide.com/ref/hdd/bios/sizeHandling-c.html
Particularly, this part:

"Ignorance": Some BIOSes will report the true number of logical cylinders that the drive has, making you think your system supports the full size of the hard disk. Really, the BIOS just has no clue what it is seeing. When you go to partition and format the hard disk, you will be stuck with the same limit (which can be both confusing and frustrating). This is usually seen with older BIOSes and the 1,024 cylinder limitation.

stamasd wrote:

The exact same thing happens as with the first DOM after fdisk/mbr, reboot etc. Size seen by DOS fdisk is 504MB.

But does it boot now?

And are you sure that you didn't partition and format, the drives that show larger capacities, in another computer or using a different method?

PhilsComputerLab wrote:

stamasd wrote:

The exact same thing happens as with the first DOM after fdisk/mbr, reboot etc. Size seen by DOS fdisk is 504MB.

But does it boot now?

No.

And are you sure that you didn't partition and format, the drives that show larger capacities, in another computer or using a different method?

I personally didn't attach the DOMs to any computer apart from the one listed above. When I received them they did have some "non-DOS" partitions as seen by the DOS fdisk which were listed at 4GB and 8GB respectively, which I deleted. It was funny how fdisk reported on the same screen that the drives had 4 and 8GB partitions respectively, all the while telling me that the disk size is 504MB.

TheMobRules wrote:

Have you tried upgrading the BIOS? There's a version (2.3) from 1995. I haven't tried it myself, but it may solve your issue.

I can't find that one. I found some mention of it on a few "driver update" sites that only try and make me download their "driver fixer" utilities instead.

stamasd wrote:

I can't find that one. I found some mention of it on a few "driver update" sites that only try and make me download their "driver fixer" utilities instead.

Here: http://www.filewatcher.com/b/ftp/ftp.aopen.co … os/vi15g-0.html

vi15gr23.exe is a self-extracting file with the 64K .rom inside. If you have a programmer and a spare chip it might be worth a shot.

TheMobRules wrote:

stamasd wrote:

I can't find that one. I found some mention of it on a few "driver update" sites that only try and make me download their "driver fixer" utilities instead.

Here: http://www.filewatcher.com/b/ftp/ftp.aopen.co … os/vi15g-0.html

vi15gr23.exe is a self-extracting file with the 64K .rom inside. If you have a programmer and a spare chip it might be worth a shot.

I do have a programmer and EPROM chips. However - I have downloaded and extracted it, and it appears to be the same BIOS I have. At least the date inside the file is the same (12/15/93). But I'll give it a try.

(edit) I did write the BIOS in a chip (used a Winbond 27C512 EEPROM, not an EPROM but it works all the same) and installed it in the motherboard. It does display a date of Jun 29th, 1995 at boot.

However it did not solve the problem. I still can create partitions of 504MB only on the DOM.

stamasd wrote:

I personally didn't attach the DOMs to any computer apart from the one listed above. When I received them they did have some "non-DOS" partitions as seen by the DOS fdisk which were listed at 4GB and 8GB respectively, which I deleted. It was funny how fdisk reported on the same screen that the drives had 4 and 8GB partitions respectively, all the while telling me that the disk size is 504MB.

Most likely they were FAT32 partitions, which are not supported by DOS 6.22. DOS 6.22 supports only FAT16 partitions up to 2GB.

Fdisk only reports what is on the partition table, it does not know the whole partition is accessible.

MSD confirms it, your BIOS does not support those disks as LBA so you can only use the first 504 MB.
Since the physical and logical geometry has 16 heads and 63 sectors, I don't understand why creating a 500MB partition does not work.

Try creating say 100MB partition, run "FDISK /MBR", reboot again from floppy, run "format C: /S" or instead of the /S parameter transfer the system files with "sys C:" command, reboot without floppy, and it really just should boot from the DOM.

Jepael wrote:

Since the physical and logical geometry has 16 heads and 63 sectors, I don't understand why creating a 500MB partition does not work.

It does work. I can create a 504MB partition, format it and transfer the system to it. After reboot, all the files are there. But the system doesn't want to boot from it. Yes, it is marked as bootable.
Also see above. Updating the BIOS to the latest revision didn't do anything to solve this problem.

Yeah, it's clearly not a problem with LBA support since you mentioned the same board is working fine with HDDs and even CF cards > 500MB.

I know you said that currently you don't have another PC to connect these DOMs, but maybe if you have one of those IDE --> USB adapters you can plug them into a modern machine and create the partitions there?

Another option: if you manage to boot some kind of Linux distribution on your 486 you can try the Linux version of fdisk and see if it helps (it would have to be a somewhat minimal Linux to fit in a floppy... but there's probably something like that out there).

At this point I think there has to be something wrong with the DOMs, maybe in the partition table or in the way they respond to the identify command from the controller...

TheMobRules wrote:

Yeah, it's clearly not a problem with LBA support since you mentioned the same board is working fine with HDDs and even CF cards […]
Show full quote

Yeah, it's clearly not a problem with LBA support since you mentioned the same board is working fine with HDDs and even CF cards > 500MB.

I know you said that currently you don't have another PC to connect these DOMs, but maybe if you have one of those IDE --> USB adapters you can plug them into a modern machine and create the partitions there?

Another option: if you manage to boot some kind of Linux distribution on your 486 you can try the Linux version of fdisk and see if it helps (it would have to be a somewhat minimal Linux to fit in a floppy... but there's probably something like that out there).

At this point I think there has to be something wrong with the DOMs, maybe in the partition table or in the way they respond to the identify command from the controller...

I have thought of using a USB/IDE adapter, I have a couple of those... but there's a problem. The DOMs have female connectors, they are meant to be plugged directly into the IDE port on the motherboard (or adapter card), not a male connector like a HDD. The USB adapters also have a female connector on them. I can't plug them in directly. I have thought of using a straight double-pin header as an interposer, but I don't have one that's long enough (2x20 pins).

I could use an old Slackware bootdisk set, doesn't fit on a floppy but two (one has the kernel, one a minimal root system). Don't remember if it has fdisk on it, but it might. I'll see if I can do it. Linux BTW ignores anything the BIOS tells it about hardware and accesses it directly.

One option might be to try one of the drive overlay softwares, then try installing DOS again?