Probably saying the obvious here, but are you sure the drives can boot as they are? I mean, can you boot another machine with them?
If you can, I have no ideas about what's happening with the 286.

If you can't, there are many things to try like SYS'ing (dos 3.3 needs an empty disk in the beginning ), ensure from FDISK that partition is set to "active"...

Yes.
They are booting on everything from my 386 to an Pentium-1. Forgot to mention that I did run both "SYS C:" and "FDISK /MBR" on them.
That thing you are mentioning, about the drive. What exactly are the partition layout on that?
When running FAT-12 (Dos 3.3) I am using a single partition of 35'ish megabyte in size. (Those a little over 300 sectors)
As I am just using Fdisk to ustilize max size of the harddrive. I even tried to get it to boot with a 5mb partition on MS-Dos-6.22.

I was only referring to the fact that you can't SYS a drive using dos 3.3 unless it's empty, or there is nothing written in the beginning of the disk.
Later dos versions take care of that by moving whatever is there to make space for the system files.

But since your drives can boot when placed in other systems, there is clearly nothing wrong with them.
Last shot: have you configured the HDD in BIOS exactly the same way (same values for all parameters) as in a system that can boot from it? (preferably the 386 which has no auto-detect or LBA)

In a similar case (also with a 286) I had to use a DDO on the drive to make it boot. Even thought the drive's parameters were accepted in BIOS, the drive wouldn't boot. A smaller drive would boot just fine. Yeah I know how weird this sounds, but if you think that there was a time where 286s were around and 270MB drives didn't exist, it makes some sense.

Hi! That's a strange issue you're discussing. My 286 boxes don't behave like that (one has a 4GB CF card).

Is there any garbage data left on the drives which could confuse the BIOS ?

Make primary partition ACTIVE with FDISK. If this doesn't help (or already is) use: "FDISK /MBR"

Before this, you have to transfer the system files with command: "SYS C:" from boot diskette.

Probably BIOS bug/incompatibility with IDE drives -> needs XTIDE BIOS or similar solution

Jo22 wrote:

Hi! That's a strange issue you're discussing. My 286 boxes don't behave like that (one has a 4GB CF card).

Is there any garbage data left on the drives which could confuse the BIOS ?

Nope... The drives were all empty.

konc wrote:

I was only referring to the fact that you can't SYS a drive using dos 3.3 unless it's empty, or there is nothing written in the […]
Show full quote

I was only referring to the fact that you can't SYS a drive using dos 3.3 unless it's empty, or there is nothing written in the beginning of the disk.
Later dos versions take care of that by moving whatever is there to make space for the system files.

But since your drives can boot when placed in other systems, there is clearly nothing wrong with them.
Last shot: have you configured the HDD in BIOS exactly the same way (same values for all parameters) as in a system that can boot from it? (preferably the 386 which has no auto-detect or LBA)

In a similar case (also with a 286) I had to use a DDO on the drive to make it boot. Even thought the drive's parameters were accepted in BIOS, the drive wouldn't boot. A smaller drive would boot just fine. Yeah I know how weird this sounds, but if you think that there was a time where 286s were around and 270MB drives didn't exist, it makes some sense.

I will try testing it with a DDO. This seems like the last solution that I can resort to.

kixs wrote:

Make primary partition ACTIVE with FDISK. If this doesn't help (or already is) use: "FDISK /MBR"

Before this, you have to transfer the system files with command: "SYS C:" from boot diskette.

I have tried all this too.

h-a-l-9000 wrote:

Probably BIOS bug/incompatibility with IDE drives -> needs XTIDE BIOS or similar solution

This is actually something that has crossed my mind too.
I have no eprom burner on my hand, so I can not do this netcard-bios trick that someone have had succes with.
The strange thing, is that it does know how to handle the drives, just not booting them.
As I can use any HDD, when I am booting from the floppy.

Hmmm.... Should I try to test it with MS-Dos-5 then?

@hal That's quite possible, as it seems the Unisys 286 is running an early version of Phoenix BIOS.
I wonder, perhaps it also has some kind of MBR write-protection feature, just like AMI ?
It's been a while since I used a Phoenix BIOS. 😅

@kixs You're right, but shouldn't be this the case by default ?
If there was no partition before, a newly created partition should be active under normal circumstances (at least in PC-DOS 3.x).
Maybe the boot loader program is corrupted or experiences difficulties..

@brostenen Do you have any disk editor at hand ? Maybe you can check your boot sector.
I've attached two pictures of cylinder(track) 0/head 0/sector 1 and cylinder(track) 0/head 1/sector 1 (MS-DOS 6.2) in case you want to compare this with your hard disk or CF card.
I used two program for this, "Sector 1.0" and "DE".. I know, it's a bit overkill but maybe it can be helpful.
Both are freeware and should run on 16bit machines (8088-286).

Hmmmm....
Now we are out in the deep water, were I simply do not swim. 🙁
I will try and test with different versions of dos, these next couple of days.

The BIOS is indeed a Phoenix, version 1.7-something (or 1.6-something)
So it might be that these BIOS's are not designed for other than an MFM/RLL controller.
Or is it possible that some of the HDD access stuff are found on the MFM-controller?
You know... Needing that special controller in order to complete a succesfull boot from HDD?
Kind of like what "h-a-l-9000" is reporting.

Yup, I think he's right. It's probably something with the BIOS.

But I wonder what it is. You said, this machine came with a MFM fixed disk, right ?
So did this MFM controller have it's own BIOS ? I mean, did you have to run the debug program from DOS if you wanted to change
the settings for the MFM drive or were the hard disk settings shown in the Phoenix BIOS (just like IDE) ?

I'm just asking, because I remember there were differences between IDE and ST506 controllers (at least the 8bit ones):
The MFM controllers had a different port address, If I'm not mistaken. But I'm not 100% sure about this.
No idea if this was related to the I/O ports of the MFM controllers themselves or the BIOS ports which are accessible by DOS.
Or maybe it was just a PC/XT thing and not really related to MFM/IDE in particular..

So if the main BIOS was designed to store parameters for MFM drives, it could be possible that it isn't aware of IDE at all.
And the reason you're able to read your HDD is because DOS is able to retrieve the CMOS parameters.
I know, this sounds strange. And it doesn't really make any sense to me - DOS relies on int13h services the BIOS provides and
IDE is a superset of ST-506/412. So maybe that Phoenix BIOS was just modified for that specific MFM controller ?

This reminds me of something, maybe I already told this story before :
Years ago I installed a File Card in my 586 PC and forgot to disable the internal IDE controller.
When I turned on my PC, the ST506 controller took over and booted an older version of DOS.
Even booting via floppy didn't work anymore. But after I removed that card, re-formated the IDE drive with DOS 3.3,
inserted the card again and booted that older version of DOS, I was able to access the IDE drive as "D". 😀
So MFM and IDE controller weren't interferring with each others. But maybe it just worked, because each of them was driven by its own BIOS.

Btw, the pictures above look more complicated than they are. They simply show the data at the "beginning" of the HDD.
The data merely consists of the boot code and some DOS/FAT information (just look at the string "MSDOS5.0", these are bytes 3-10).
Just run one or both programs on your 286 and check if your disk looks *similar* like this. There's nothing else to do. 😀

Here's an explanation of the stuff you see in the second picture. The only difference is that they talk about a floppy instead of a hard disk.
So it's a little bit different (floppies don't have a MBR, for example). And the terms are also a bit different.
On a HDD you speak of cylinders/heads/sectors (CHS) and on a floppy you say tracks/sides/sectors.
I know, that's a bit confusing. Especially if you use a floppy utility to examine a HDD or vice versa..

Here's a good example of this (do you notice, it says sector/side/track even though it is a HDD):
The difference between Floppy Disks and Hard Disk Drives

Edit: Edited something. Hope the text is more friendly now.

Edit no.2: I forgot to mention: The first picture with the boot code shows the Mast Boot Record (MBR) at CHS 0,0,1.
This disk region is also known as "Absolute Sector 0", because it's the very first physical sector on a HDD. You can read more about it here.
This thing is created when you run FDISK /MBR. And the end of the MBR also contains the Partition Table, which FDISK writes to when you make or change partitions.

Just try a smaller drive or install a DDO on one. In 5' you'll know if the PC can boot from an HDD or you should look for other things. Most probably the BIOS doesn't handle correctly the disks you defined.

brostenen wrote:

kixs wrote:

Make primary partition ACTIVE with FDISK. If this doesn't help (or already is) use: "FDISK /MBR"

Before this, you have to transfer the system files with command: "SYS C:" from boot diskette.

I have tried all this too.

Try again just to make sure. Had to do this a couple of times last week on my 286. It booted fine with a Pentium, but no-go on a 286. Used real HDD and a couple of CF cards. It finally worked only after I did everything again and made sure to use CHS all the way - used MS-DOS 6.22 boot floppy on a 286 and done everything on it (fdisk /mbr was the most important as I used the CF cards with newer systems. MBR doesn't change with reformat).

I agree with both of you. But I guess it's also a good idea to wipe the disk completely, before doing anything with it.

Okay..... Minor progress. I installed one of my 32mb CF cards as a fixed type harddrive at a size of 10mb (cant remember the number between 1 and 46).
And the machine would boot the card with a complete MS-Dos-6.22. This points to a error bug in the BIOS when running type 47. As far as I can figure it out. As of now, I have downloaded EZ Drive Overlay for WD drives, and trying it out on a Seagate drive. It's the one version that Phil have tried on CF cards, and other non WD drives. Yeah.... It booted to the drive, and I am currently installing MS-Dos-6.22, having started the installer from the DDO that resides on the drive. I thought that I might as well give it a go, using a 630mb drive, before I am testing 32mb CF cards. 🤣

I really hope this is the way to go, before hunting expensive hardware down. If this is the solution, I need to source eighter a cool running SCSI-II harddrive (perhaps a 40mb from an Apple LC-III), or I need to source an 16-bit ISA card with EIDE option. (BIOS).

SUCCESS..... It is now working. 630 glorious Megabytes at my disposal, on the Unisys Pw/2 Series 300.
Thanks to all, for helping me solving this issue.

Good to hear that it's working now. Out of curiosity, when you were choosing type 47, you were setting the correct drive parameters as values, right?