You did not gave any information about your actual Conner drive, however I try to give some advice:

In case you have a CFA170A or the same series it is a difference if you pair conner drives or mixed vendors. Conner has it's own Master/Slave logic.
If you have mixed vendors you have to set jumpers differently, like explained here:
http://www.computerhope.com/hdd/hdd0003.htm
It requires the A/C jumper removed.

That might be helpful. However, mine is model CP30174E.
I actually looked up the jumper settings of this model on the same site, but if it does have a similar jumper (I haven't really looked that close), then I'll give it a try.
Alternatively, I can just try another hard drive.

What about booting the machine from floppy and seeing if you can get the CD drive to work without the Conner HD attached? That would isolate it to the HD if it works.

Yes, that is another option I had considered, but thought I'd see what other users' opinions or thoughts were.

I remember having similar problems in the past with a 486, the issue here was a jumper on the hdd set to 'Master only device' or something like that.

This avoided the slave device (cdrom) to even be recognised, so maybe checking the jumper config again might solve the problem.

Can you try connecting the CD drive to the IDE port on a sound card?

oerk wrote:

Can you try connecting the CD drive to the IDE port on a sound card?

That was also a thought. Unfortunately, this particular sound card still has the proprietary CD-ROM interfaces (on the Aztech range of cards, it was usually Panasonic, Mitsumi & Sony drives that were supported).
I could install a later model sound card that has an IDE interface, but I want to try and stay as "period correct" as possible with this particular PC (in this case around late 1992).
Unfortunately, I don't have a CD-ROM drive with a proprietary interface and the earliest drives I have are all from around 1997/1998. Older drives also tend not to be able to read CD-RW discs.

The only reason why I want to install a CD-ROM drive is to make it easier to transfer data to this PC, as opposed to having to zip everything and span it across several 1.44 MB floppy discs (this brings back memories).
I can unplug the hard drive, but that is just too much of a pain. I am also considering going the Compact Flash card route, which would make it easier to transfer data.

But, the advice given should help me to narrow down the problem. I tend to think that it is probably a case of the hard drive and CD-ROM drive not being too happy to be on the same IDE channel.

Just an update to my problem (in case someone else comes across a similar one).

I disconnected the hard drive completely and connected the CD-ROM to the IDE channel (I tried it under both master and slave).
I then used a Windows 98 startup disk (selecting CD-ROM support), but unfortunately no luck. It wouldn't recognise the drive.

I then decided to swop out my sound card for one that does have an IDE interface on the sound card (I used an Aztech Labs sound card, but any sound card with an IDE interface should do).
Finally, when loading the device drivers, it immediately recognised the drive and I can read all the contents on the CD.

I would therefore appear that the IDE controller (most likely in conjunction with the BIOS) does not support CD-ROM drives connected to the IDE channel of the controller.
I also tried a different 16-bit IDE controller (that utilises a UMC chipset) and experienced the same problem.
Either both controllers are just too old, or it could be a BIOS problem.

EDIT: This particular motherboard has a Quadtel BIOS (it's a Forcom B639 motherboard with a Headland chipset) and, with a hard drive properly configured (in the BIOS setup), automatically bypasses the A: drive and boots up straight from the hard drive (there isn't a setting in the BIOS to choose which drive to boot up from first).
If, however, I do not specify a hard drive in the BIOS setup, then it will read the A: drive first.

I'm glad it finally worked. It was probably because the controller lacked atapi compatibilty.
Controllers like this are basically just glue logic to interface the cdrom's bus with isa.
There were even eprom programmers that were meant to interface with those ide controllers instead of isa (was cheaper, no card required).

I once tried to use a generic ide/atapi drive on a multi i/o card,too)..
While it did work, it was not very stable and had some nasty glitches.
Like for example, when I changed the cd, the directory listing didn't update.
It was still the same from the previous cd. Or read errors. And multi-session discs caused problems,too.
Sometimes the drive stopped responding 'til the reset.

If its only for filetransfer you can use a serial connection and software like comdrive or FileMaven.

The boot order is stored in the cmos, don't think it was hardcoded.
Maybe it can be changed with some dos utitliy (I know one for ami bioses).

Jo22 wrote:

I'm glad it finally worked. It was probably because the controller lacked atapi compatibilty.

Maybe you could explain this in a bit more detail.

Where is the specific source that the controller requires compatibility with a AT-A based protocol?
From my point of knowledge any controller that implements AT-A is capable of running ATAPI on top of it. Here is the quote from ATAPI 2.6 standard that states in Abstract 1.1 "The ATAPI and CD-ROM interfaces described in this document are compatible with existing ATA hardware
without any changes or additional pins."

source
SFF Committee
Information Specification for
ATA Packet Interface for CD-ROMs
Rev 2.6 January 22, 1996

I think it is an urban myth that an ATA controller has to have specific additional features that enable ATAPI comaptibility on hardware level.
It is merely a BIOS support issue as known from most IDE RAID controllers.

I tend to think it's most likely just a BIOS issue, as opposed to it being the controller.
I have a very similar controller in a 386DX-40 PC of mine and on that motherboard (which has a later AMIBIOS), the CD-ROM drive runs just fine on the controller's IDE channel (as a slave).

This particular Quadtel BIOS does pick up the controller (it has a very primitive "system information" screen) which identifies the COM & LPT ports (amongst other things like the CPU & memory).
Under the Setup screen, you have to specify the hard drive parameters. If you don't, then it won't boot to the hard drive.
There is also no "auto-detect" feature in this BIOS (you can either select from a pre-defined list of hard drive parameters, "User" or "None") - perhaps this is why the CD-ROM drive is not recognised?

What is interesting is why the CD-ROM does work on the sound card's IDE channel, but then again it could be that the sound card's IDE controller has its own logic built in to "look" for a CD-ROM drive (I'm just speculating here, I could be wrong).

Good question. And thank you for the specs, elianda. Unfortunately, I can't provide an exact or satisfying explanation to both of you.
But I recall that there were said to be differences between early ide cards (ide,ata0) and the ide interface on soundcards. Thats at least what media said back then.
They encouraged to use the soundcards interface and not the normal ide one.
I don't know if that's true, maybe its wrong. I can only tell from my experience:
The latter worked just fine with cd-rom drives, while the former didn't in a reliable way for some reason.
Especially the "directory listing not updating" thing is something I noticed a few times while using standard at-bus controllers.
So it was just a guess it's related to atapi somehow. I mean, if I'm not mistaken, the main feature of atapi was to allow non-fixed media such as streamers and cdroms to work.
If its just a myth and a pure BIOS issue, then I'm quite confused now. I don't think my early CT1740 has any kind of option rom or firmware for atapi.

But wait, this gives me an idea. Maybe we're just heading in the wrong direction..
What if the issues are not related to the ide circuit itself, but the wiring,
the electrical levels (missing buffers, etc.) or the fact that on an early ide controller card,
the cdrom always has to share the controller with the hard disk ?
Maybe that's why they did recommend the soundcard instead ? So eventhough it's still not having the ability to properly
share hdd and cdrom (because there's no BIOS support for atapi), it is having at least a second ide interface on a different
port address (0x01F0-0x01F7) where they won't conflict.
Accoding to ATA specifications, ATA multichannel support wasn't introduced until ATA-2.
And before that it wasn't common or always possible to add a second ide controller, as they had no jumpers
and were thus fixed to one port address (0x0170-0x0177). So the presence of a jumper/second ide port makes
soundcards interfaces in some, albeit very limited way, conform to ata-2 ? Yeah, that sounds odd.
Still, I'd love to check this now. If I only still had an old ide drive to test.
So may I ask, has anyone of you successfully used both a hdd and a cdrom on a soundcard in an old computer (286-486) before ?

ATA Multichannel Support, PC Hardware in a Nutshell
https://www.safaribooksonline.com/library/vie … ch13s01s03.html

IDE / ATA Parallel Bus Description
(ATA-2: "With a maximum of 4 devices on the bus.")
http://www.interfacebus.com/IDE_HardDrive_Int … escription.html

ATAPI (AT Attachment Packet Interface)
("ATAPI is part of the Enhanced IDE (EIDE) interface (also known as ATA-2).")
http://whatis.techtarget.com/definition/ATAPI … acket-Interface

Another update (hopefully the final one).

I did some digging and noticed that my hard drive has two additional jumpers, labelled "E1" & "CAM".
Depsite there being nothing in the documentation of what the purpose of these other two jumpers are, I decided to short the "CAM" jumper as well.
Once I rebooted, it successfully recognised and initialised the CD-ROM drive on the same IDE channel (as the slave drive).

So, it would appear that it wasn't the BIOS after all, but just a conflict with the hard drive.

Strangely enough, if I try to boot up with a Windows 98SE startup disk and select CD-ROM support, then the system just hangs.
If, however, I just use a standard DOS 6.22 boot disk (with the CD-ROM drivers copied to the disk and I load them at start up), then it does recognise the CD-ROM drive and intialises it.

So, now I've learned something as well about old Conner hard drives.
Sadly, the hard drive does seem to be on its way out, since its having trouble spinning up when I shut it down for a longer period.

Aha! http://stason.org/TULARC/pc/hard-drives-hdd/c … -SL-IDE-AT.html
Mentions the drive supports two Master/slave configurations...
ISA/Connor
ATA/CAM (Common Attachment Method)

Some other Connor drives (and other old era drives) have a different variant, where there is a setting of DSP (Drive Slave Present)

For drives that predate the unified implementation of Master/Slave, it does get tricky