VOGONS


First post, by appiah4

User metadata
Rank l33t++
Rank
l33t++

Well, this is starting to become a fairly bothersome issue, so much so that I may forego building the 486 I am working on and settle for a Socket 5 Pentium instead.. But before I do that, maybe Vogons can help me out.

I am building this PC and I'm having trouble with IDE/Floppy drives every step of the way. Let me start out by listing the parts in the system giving me trouble:

UMC MB457 MB w/ OB U5SX 486-33F
16MB EDO RAM (4Mx32 with single side HM5117405S6 chips)
Goldstar 48x IDE CD-ROM

Expansion cards are:
ISA1: Acer M5105 I/O Control Card
ISA3: Trident TVGA900B 512K VGA
ISA5: Edison Gold-16 ES688FC Sound Card

The I/O card is configured with FDD enabled, IDE primary, COM1 and COM2 enabled, LPT configured as LPT1 as per thisjumper configuration diagram. Devices on the Primary IDE of this card are detected on IRQ14, at least the CD-ROM is, when MSCDEX is loaded.
The sound card is hard configured to A220/I7/D1, and the IDE controller on it is enabled as Secondary on IRQ15.

And here's the issue: IDE and Floppy is giving me all kinds of headache.

I first installed a Mitsumi Floppy Drive and a backplate Compact Flash adapter, and connected them to the I/O card's IDE header with the CF adapter as master and CD-ROM as slave, configured by jumpers. I have tried two diffretent compact flash cards, one is a no-name 256MB from China and the other is a 2GB Sandisk. The BIOS (which is a 1995 AMI BIOS with the WinBIOS interface) and LBA support, which auto-detected the compact flash cards in both cases no problem. The PC booted up, but refused to boot to the FAT16 partitions on either card, which I had prepared in a virtualmachine guest on a Win7 host PC. I said OK, that's strange, let me try to set them up on this PC because you know, old controllers can not read Hard Drives prepared by other controllers sometimes, so I went ahead and tried to boot the PC with a Windows 98 SE boot disk I had around.

The PC booted off the disk, but also spat out a couple of complaints about unknown command lines in CONFIG.SYS and AUTOEXEC.BAT. Strange, so I did a type CONFIG.SYS and type AUTOEXEC.BAT and both returned as a couple lines of garbage ASCII text. Now, at this point I would say the floppy drive is messed up, but the garbage ASCII text is consistent whenever I do a type, and it always boots up fine! So I said, ok, let's try to partition the hard drives anyhow, I launched fdisk and it hard locked up. For both CF Cards.

At this point I pull out both the CF adapter and the floppy and I install a WD Caviar 1.2GB HDD and a GOTEK Floppy Emulator. I launch an MSDOS 6.22 boot disk image, and it boots fine. It reads everything on the disk. FDISK runs fine. All is well. Except now the HDD I installed is making lots of weird clicking noises, and I remember I forgot to auto-detect it in BIOS, so I do that, and it gets detected fine and the clicking kind of goes away but not quite. I boot the DOS 6.22 Boot Disk and it even runs MSCDEX and configures the CD-ROM which is installed as a slave on the IDE chain. I am ecstatic, so I put a CD in there, and type R:

CDR101:Not ready reading drive E Abort, Retry, Fail? (I hit f) Fail on INT 24

I boot MSDOS 6.22 images, repartition the hard drive, format it and install MS-DOS 6.22. It installs fine, although there are strange hiccups - while partitioning, FDISK disk validation progress erratically resets to zero at times. Regardless, DOS boots from C: and I think to myself, phew, so I guess it was the CF adapter all along. But I decide I would do a scandisk on the HDD as I haven't used it in ages, so I run SCANDISK.

Disk Read Error

That's when I said FUCK IT and shut it off.

Granted, I missed a step here, I did not run the GOTEK alongside the CF Adapter, so I don't know if that combination would work or not, but something's wrong. Either the I/O card is dead, or both the floppy and the HDD are dead, and possibly the CF adapter is incompatible. The system runs fine with just the GOTEK on the FDD connector.

I will try it with the CF cards tonight, then with a Seagate 8.4GB drive that I know for sure is fine (but good luck getting it detected with such an ancient BIOS). I used the IDE cable I currently use with another build and had no issues, but I'll also try another IDE cable, just in case. I will also try moving the CD-ROM out of the IDE chain and onto the sound card. For all I know, the floppy, HDD and CD-ROM could all indeed be dead, or the I/O card is to blame (but the COM port is fine and FDD works fine with a GOTEK) or I messed something else up in the system build but I can't find it. I'm open to all suggestions.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 1 of 14, by Jepael

User metadata
Rank Oldbie
Rank
Oldbie

Can the PC boot from the secondary IDE which is on the sound card?
If so, put the CF card adapter there and see if the problem disappears.

Also, does the problem disappear if you have only standard HDD and standard CDROM on primary IDE?

I believe the problem is that the multi-IO controller is not compatible with CF cards, the multi-IO would have to specifically take CF cards into attention so that CF card does not conflict with Floppy signals.

Actually it's the CF cards that are not 100% bus compatible with IDE hard drives. It's not the CF card adapter but CF card itself.

Reply 2 of 14, by RJDog

User metadata
Rank Member
Rank
Member
Jepael wrote:

Actually it's the CF cards that are not 100% bus compatible with IDE hard drives. It's not the CF card adapter but CF card itself.

Yeah, CF cards are ATA devices, not IDE devices. AFAIK (so stand to be corrected) the ATA standard evolved from the IDE standard, so IDE devices are ATA devices, but not necessarily the other way around. ATA also includes things like "IDE" CD-ROM drives, CF cards, etc.. if the interface card and/or BIOS is pre-mid-90s there is a chance that compatibility of this kind is the problem.

The other potential problem is that the CF card is likely an "LBA" device, where the interface card and/or BIOS may be expecting a device to identify as CHS. If you can manually enter CHS values, maybe that might work towards a resolution?

Reply 3 of 14, by appiah4

User metadata
Rank l33t++
Rank
l33t++

Thanks for the input, here's an update.. The CF card works and I managed to boot from it. One of the following did the trick, though which, I do not know:

1. Moved the CD-ROM to the sound card's IDE controller with the old ribbon cable
2. Removed the WD Caviar HDD which I now presume is bust
3. Reinstalled the CF adapter with the 2GB Sandisk CF, connected it to the IO card with a new ribbon cable
4. Booted DOS 6.22 install disk, F3'ed out of setup and did an fdisk /mbr followed by sys c:
5. Rebooted and everything worked like a charm.

Except that the CD-ROM still gives me CD101 errors. Funny thing is that it shows up fine in system diagnostic tools, and the DOS CD-ROM drivers load fine. Something must be wrong with it mechanically, I put a CD in and try to get a dir listing, but it just tries to spin the disc up and stops after a moment of trying, and keeps doing that.

Anyway, at the end of the day it would seem the floppy drive, hard drive and cd-rom drive were ALL dead. Well, at least the floppy seems to be dead anyhow, the same floppy cable and controller work fine with the GOTEK. As for the Hard Drive and CD-ROM, either they are both gone or the IDE cable is faulty. I'll try that tomorrow.

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 4 of 14, by yawetaG

User metadata
Rank Oldbie
Rank
Oldbie
appiah4 wrote:

Except that the CD-ROM still gives me CD101 errors. Funny thing is that it shows up fine in system diagnostic tools, and the DOS CD-ROM drivers load fine. Something must be wrong with it mechanically, I put a CD in and try to get a dir listing, but it just tries to spin the disc up and stops after a moment of trying, and keeps doing that.

Sounds like a reading problem, could be something as simple as a dirty lens. Also, some drives can't read CD-R's properly.

Reply 5 of 14, by appiah4

User metadata
Rank l33t++
Rank
l33t++
yawetaG wrote:
appiah4 wrote:

Except that the CD-ROM still gives me CD101 errors. Funny thing is that it shows up fine in system diagnostic tools, and the DOS CD-ROM drivers load fine. Something must be wrong with it mechanically, I put a CD in and try to get a dir listing, but it just tries to spin the disc up and stops after a moment of trying, and keeps doing that.

Sounds like a reading problem, could be something as simple as a dirty lens. Also, some drives can't read CD-R's properly.

I tried with an original AWE64 driver CD. It actually is a rather old Goldstar 48x that probably did not get used in years. How to clean the lens though?

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 6 of 14, by Mister Xiado

User metadata
Rank Member
Rank
Member

One of the old CD lens cleaner discs would work fine, if you can get it to play the disc past the brush track. Alternately, disassemble the drive and clean the lens directly. You just have to take the outer steel shell and front faceplate off. PITA, really.

b_ldnt2.gif - Where it's always 1995.
Icons, wallpapers, and typical Oldternet nonsense.

Reply 7 of 14, by CkRtech

User metadata
Rank Oldbie
Rank
Oldbie

If you do open it up, it is a great time to also do any lube work on gears and tracks. Make sure your lens can move well, make sure the tray can move well, etc.

Displaced Gamers (YouTube) - DOS Gaming Aspect Ratio - 320x200 || The History of 240p || Dithering on the Sega Genesis with Composite Video

Reply 8 of 14, by appiah4

User metadata
Rank l33t++
Rank
l33t++

Sounds like more work than I am willing I will replace with a newer 52x CDROM and be done methinks.. and if that wont work then the problem mist be the IDE cable. If I had it connected the wrong way around sysmark would not have detected it on the ide chain right?

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 9 of 14, by RJDog

User metadata
Rank Member
Rank
Member
appiah4 wrote:

If I had it connected the wrong way around sysmark would not have detected it on the ide chain right?

Correct; like most things digital, it either works or it doesn't (with some room for errors... but if the cable looks to be in good physical condition then you're likely okay)

Reply 10 of 14, by appiah4

User metadata
Rank l33t++
Rank
l33t++
RJDog wrote:
appiah4 wrote:

If I had it connected the wrong way around sysmark would not have detected it on the ide chain right?

Correct; like most things digital, it either works or it doesn't (with some room for errors... but if the cable looks to be in good physical condition then you're likely okay)

Right, so goodbye CD-ROM, hello CD-ROM, whenever I have the time.. I hope I can find a nice clean beige CD-ROM with play/stop skip etc. buttons and a volume dial on it. Much more fitting for the era..

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 11 of 14, by chinny22

User metadata
Rank l33t++
Rank
l33t++

I left my dead CD-ROM in my 486, its era fitting as its the original drive!
But with gotek floppy, installing dos/win3x is quick, and after that you can use the network to transfer files, or worst case pull the cf card out and put it in a reader on another PC.
What I'm trying to say is I don't really miss having a working CD drive in my 486

Reply 12 of 14, by TheGreatCodeholio

User metadata
Rank Oldbie
Rank
Oldbie

Assuming your version of Windows 95 is pre-Windows 95 OSR2, the partition is probably not marked as a LBA partition, not FAT32. That means that disk access is still done through C/H/S geometry. You said the BIOS is from 1995, which means it probably predates the INT 13h extensions that allow LBA access. So the most likely reasons is that the disk geometry the BIOS presents differs from the disk geometry you formatted it with. That will always make MS-DOS / Windows 95 choke at bootup.

I suggest looking at the disk geometry the BIOS is giving the CF card, and then recreating the disk image in emulation with that exact geometry and formatting DOS with it. The disk geometry will match and DOS will boot properly.

DOSBox-X project: more emulation better accuracy.
DOSLIB and DOSLIB2: Learn how to tinker and hack hardware and software from DOS.

Reply 13 of 14, by Jo22

User metadata
Rank l33t++
Rank
l33t++

This LBA/CHS issue reminds of a video I watched earlier: https://youtu.be/wNpq0sybH6Q?t=496
I guess quite a few DOS/Win 3.1 users with disk overlay software must have had a similar experience at some point, as well.
- Not knowing, that earlier versions weren't Win95-ready yet, they ended up with a broken installation.

"Time, it seems, doesn't flow. For some it's fast, for some it's slow.
In what to one race is no time at all, another race can rise and fall..." - The Minstrel

//My video channel//

Reply 14 of 14, by appiah4

User metadata
Rank l33t++
Rank
l33t++

In this case it did turn out that all three IDE devices I was using (hdd, cd, floppy) were bad, but this motherboard had much worse issues than ide so this build has since been discarded.

Retronautics: A digital gallery of my retro computers, hardware and projects.