douglar wrote on 2023-02-14, 17:21:

I'd like to use CF cards to ferry data to the device because floppy are cumbersome.

Does anyone have any tips on how to make an IDE drive work along side an MFM drive? Is there a formula to follow or do I just have to find an IDE controller that is kind?

Normally, the XT IDE Universal BIOS can co-exist with a real MFM/RLL controller card.
It worked at least in my XT PC with a WD controller and an XT CF Lite card.

However, it may depend on which option ROM starts first in memory.
It should be the MFM/RLL controller BIOS, then XUB, afaik.
That's the default constellation, also, I believe.

A special case might be if you're running an 8-Bit MFM/RLL controller card in an AT class PC.
The option ROM on the 8-Bit card will then take full controll over the boot process, afaik.

I did such an experiment with an old Pentium PC years ago:
https://www.youtube.com/watch?v=irtDxYman88

It has to do how the memory (BIOS parameters, i/o etc) is organized, I guess.
The controller BIOS isn't detected as a HDD BIOS by the PC/AT BIOS,
it just runs the controller BIOS after finishing POST as if it was any other program, say, XUB.

-> If the MFM/RLL controller was a 16-Bit ISA model, it likely wouldn't use its own BIOS.
Rather, it would rely on the PC/AT BIOS and its HDD table stored in CMOS (remember type 47 ?).

So maybe with an 8-Bit card you can't use the on-board IDE controller for booting from an IDE HDD.
Maybe you can still access the HDD, however. Not sure if DOS still sees the HDD parameters of the PC/AT BIOS.

There are at least two ways:

- MFM drive connected to XT-style controller (8-bit, with its own BIOS)
- IDE drive connected to regular Primary IDE controller, but it can't be controlled by the main BIOS, so there must be no hard drives in the CMOS Setup - instead, use XUB

or

- MFM drive connected to AT-style controller (16-bit, without BIOS), using the same address/IRQ as Primary IDE, so make sure to remove/disable the Primary IDE
- IDE drive connected to Secondary IDE controller
Both drives must be entered in the CMOS Setup, of course no autodetection for the MFM one, must use CHS translation, PIO mode 0, and disable any other IDE-specific options like "IDE block mode"

Thanks. I can work with that info.

Did I lose the data on the MFM drive when I switched from the 8 bit controller (from an early 16kb 5 slot 5150 PC) to the 16 bit WD controller ?
The drive had not likely been powered on in 20 years.

douglar wrote on 2023-02-14, 19:25:

Did I lose the data on the MFM drive when I switched from the 8 bit controller (from an early 16kb 5 slot 5150 PC) to the 16 bit WD controller ?

No, replacing the controller doesn't affect the data.
However, as various controllers use various low-level formats, re-formatting may be necessary, and that's when the old data is gone.

Grzyb wrote on 2023-02-14, 19:02:

There are at least two ways: […]
Show full quote

There are at least two ways:

- MFM drive connected to XT-style controller (8-bit, with its own BIOS)
- IDE drive connected to regular Primary IDE controller, but it can't be controlled by the main BIOS, so there must be no hard drives in the CMOS Setup - instead, use XUB

or

- MFM drive connected to AT-style controller (16-bit, without BIOS), using the same address/IRQ as Primary IDE, so make sure to remove/disable the Primary IDE
- IDE drive connected to Secondary IDE controller
Both drives must be entered in the CMOS Setup, of course no autodetection for the MFM one, must use CHS translation, PIO mode 0, and disable any other IDE-specific options like "IDE block mode"

+1 to all this

Grzyb wrote on 2023-02-14, 19:02:

- MFM drive connected to AT-style controller (16-bit, without BIOS), using the same address/IRQ as Primary IDE, so make sure to remove/disable the Primary IDE
- IDE drive connected to Secondary IDE controller
Both drives must be entered in the CMOS Setup, of course no autodetection for the MFM one, must use CHS translation, PIO mode 0, and disable any other IDE-specific options like "IDE block mode"

For some reason this never worked for me. Depending on the IDE card, BIOS would either complain about HDC failure or would not boot even though both HDDs had system on them.
In the end I had to put SCSI card with it's own BIOS to get another disk alongside MFM for backup and surface testing. I'm not sure if the IDE cards I have are the problem, or the BIOS on the mobos I've tried is to blame (I only tried 2 mobos before I gave up and went SCSI). I also tried XTIDE (ROM in network card) but that didn't work properly either. Can't remember the details now, I think the IDE HDD was detected but then failed to read/boot anyway.

I suspect the IDE card because I could not get the MFM to work even if I had no IDE HDD connected at all, and the card was set to IDE disabled (tried both primary and secondary).

Deunan wrote on 2023-02-14, 23:57:

I suspect the IDE card

Hint: don't use a card, use onboard IDE.
With onboard IDE, you can be pretty sure the Secondary channel uses different I/O address AND different IRQ than the Primary.
And the BIOS is sure to properly operate the Secondary.

Yep. Every ISA (and even a VLB) IDE card I've been able to find that allows jumpering to the secondary port still uses interrupt 14 no matter what

So I need a Pentium mobo to run MFM along IDE? You see, I wanted to backup and test that MFM in the original 286 it was in - yes, in the end I did take it out since I decided to clean the case and mobo properly, but it was a lot of work (everything has to come out, and even some panels of the case must be removed). So that's why I prefer to use SCSI, it actually works in target system.

But it's good to know. If I ever get another MFM drive to test I will know what mobo to pick for that.