Image Map Image Map
Results 1 to 4 of 4

Thread: Disk controller woes

  1. #1

    Default Disk controller woes

    I recently bought several PCs and clones because I wanted to test a group of hard disk drives I had lying around. The drives are (for the most part, the ones I'm testing here) known to be good.

    Sorry in advance for this wall of text, tl;dr my MFM controllers don't work please help me

    The machine is an Informate "Business Computer Desktop," an 80286 @ 16MHz. It has a full bank of 511000 chips on the board, which I believe is either 1 or 2MB of RAM. It uses an Award BIOS with an internal setup. The CMOS battery is good and it holds it's settings, even if the clock is a little fast (20 minutes fast every day).
    I couldn't get either of my ISA VGA cards to work in it (AVGA-1 and an Oak OTI067 based card from Paradise), so I'm using an ATi EGA Wonder Series 2. It's configured for CGA composite and works wonderfully.
    It's very obviously a clone of the IBM AT.

    I also bought an IBM PC 5170 AT (The 8MHz model, which came with the ST-4038 hard disk). It's power supply was beyond repair and even with a substitute supply I was only getting a power LED. No post no matter what. Even with all of it's RAM removed, there was no response from the IBM at all, so I began to part it out and put the pieces to good use in the Informate instead.

    The IBM had it's original hard disk, controller card, an AST Advantage 128 with one bank of 4164 chips (128KB), and a modem installed. The modem was put into storage since I didn't need it.

    I've configured the AST Advantage 128 per the manual's specifications for the Informate, and I've added two more banks of 41256 chips (The board supports mismatching chip types, it allows the use of either 4164s or 41256s as long as they are separated by bank. I'm using one of these configurations, specifically 128K+512K+512K). It's set to split addressing right now, which is for a 512K AT. If I put it into contiguous mode, which is for 512K or 640K ATs, the machine will hang. In split addressing, I see no memory benefit at all. The BIOS says 640K OK, 512K OK, and Enhanced mode 1152K OK. It does this even without the board installed. I'd like the extra EMS / XMS if possible, so if anybody knows what I could do, let me know.

    And now we get into my disk controller issues...
    At first, I was using the AT's original disk controller. It's a full length, full height 16-bit ISA card with 5 connectors (diskette drive, Winchester, HDD0, HDD1, activity LED) and it has 4 jumpers to set. E1 pins 2-3 are closed (Normal Floppy address), E4 pins 2-3 are closed (Normal Winchester address), E7 is closed (??), and E11 is closed (??). The card is marked (C) 1983 FIXED DISK-FLOPPY DISKETTE, ASSY 61-031099-00, PN68X3756.
    The card was working fine for a time (Although it would not illuminate the drive activity LED). It was able to see the disk, perform and pass all of SpeedStor and HDAT's tests, and the BIOS was able to see and boot from it for about 2 hours. Then it failed to boot when restarting, then FDISK reported errors reading, then DOS would not reinstall, then the BIOS displayed an Error Initializing Disk Controller 0 at startup.
    Both SpeedStor and HDAT seek the drive perfectly fine with this controller. The drive can pass read and write tests in both, and passes it's surface media scan with just 6 defects (3 of which were already labeled on the drive).
    FDISK and FORMAT will refuse to see the drive or even run in some cases. DOS 6 installer does not recognize the controller and disk, even if it's low level formatted.

    The drive is an IBM Type 0665-53, a ~44MB MFM disk drive. It uses a rotary voice coil actuator. The drive is known good, and I've installed MS-DOS and Windows 3.1 on this drive before in another machine. As stated before, the drive still passes all of it's tests perfectly well. It's BIOS type 14 on this machine, 733 cylinders, 7 heads, 17 sectors per track, and no WPC.

    I figured that the computer failing to recognize the disk controller could be the controller beginning to fail (even though these symptoms have been constant for the last 3 days now, and the controller still passes tests in SSTOR and HDAT.) and replaced it with another controller, this time a Cirrus Logic controller.
    The Cirrus Logic controller has a few main problems. First of all, my floppy cable would not reach it without disconnecting my 1.2M drive. Oh well, I only needed 1.44M drive anyways. I've ordered a longer cable on eBay and it'll be here in about 5 days, so it's not a concern right now.
    The CL card is also a 16-bit ISA card (full height, but not full length) and also has 5 connectors on it, Floppy, Winchester, HDD1, HDD2, and Activity. This controller has 5 jumpers to set, all of which are labeled: EH is closed (HDD controller enabled), EF is closed (Floppy controller is enabled), HP_HS is closed (Normal HDD address), FP_FS is closed (Normal floppy address), and ED is closed (Enable DMA). This card has two labels on it - "SM-1150 VER 3.18" and "SM-1150," but I can't find anything about this card other than a post I made a few years ago to another forum about it.
    When this card is installed, the diskette drive seek fails and the computer cannot boot from floppy, but it will automatically boot from the hard disk. Even though this card is not the card that low level formatted the drive, it booted DOS 3.31 from the disk (was testing and managed to copy a file). However, the lack of floppy support makes this card useless to me.
    I moved E4 to pins 1-2 on the IBM card to disable the hard disk controller and removed the EF shunt on the CL card to disable the floppy controller. This way, I would get the IBM card's working floppy controller and the CL's working hard disk controller, at least until I could track down a different card.
    No such luck. The system hangs at the end of POST and does not respond at all until I reset it. Remove one of the cards and it's back to "normal."

    Okay, maybe that controller wasn't able to use a 1.44M diskette drive (even though all the chips on the board date to 1989), or maybe the controller was incompatible with something the Award BIOS was doing. No worries, I have more cards.

    Western Digital controller is next. This is a more basic variation of an advanced card, with a smaller ROM and only one jumper, but provides the same functionality, just without latching and caching features.
    It's a 16 bit ISA card, full height, but not full length, which again has 5 connectors - Floppy, Winchester, Drive 0, Drive 1, LED. The more advanced card has 7 jumpers to set, but mine only has one, which is W1. Only pins 7 and 8 are available and they are closed.
    This time, the machine boots and gives me a diskette drive error. I check and the BIOS diskette type has changed. I set it back to 1.44M and the error went away... To be replaced by what seemed like a hang. The computer stopped responding for several minutes until I decided to put a diskette in the drive, at which point the computer beeped and told me there was an "Error Initializing Hard Disk 0" and asked me to check setup. Still type 14, all parameters still the same. Exit and do the same thing again, but press F1 to continue this time. The machine boots into DOS 3.31 and SSTOR and HDAT are both able to perform seek tests on the drive. However, using either utility to perform read / write tests or a media surface scan results in every cylinder throwing an error. Yes, all 733 are bad across all 7 surfaces apparently. I know this is not true, like I said, the computer *just* booted from the hard disk with a different controller and the disk passed all it's self tests with the original IBM controller.

    I have another WD controller, which is the more advanced version of the card I was using. It's more or less the same, but it's a bit worse for wear. I'll try it later, but it looks like there's corrosion on the contacts, so I may have to clean it first.

    Total Hardware 99 does not have info on the IBM card or the Cirrus Logic card. The WD card is listed at https://th99.infania.net/c/C-D/20506.htm. TH99 says that W1D (pins 7 and 8, I think) are "Factory configured - do not alter," and that it should be open. I'm going to remove this jumper before switching to the more advanced card (Which does not have a jumper in this position, as TH99 specifies).

    Any thoughts on why I cannot for the life of me get this computer to work with an MFM hard disk drive? I just want to use it to test these drives and record the sounds they make, I don't even have an RGBI monitor to do anything else with!

    I'm open to any suggestions except "buy a new one" or "use another machine" - This is the only working machine I have right now, and while the IBM card was working fine at first, used examples sell for more than I paid for the entire IBM AT before shipping. I wouldn't be opposed to suggestions for good MFM / Floppy controller cards, however.

  2. #2

    Default

    Just a few thoughts...

    At this point in time no MFM drives can be considered "known good" with any real certainty. Any one of them can fail at any time. Along those lines, it's much more likely that a drive has failed than a controller.

    It's quite simple to remove the 1.2M floppy from its slot and use it from there.
    PM me if you're looking for 3" or 5" floppy disks. EMail For everything else, Take Another Step

  3. #3

    Default

    Quote Originally Posted by Stone View Post
    Just a few thoughts...

    At this point in time no MFM drives can be considered "known good" with any real certainty. Any one of them can fail at any time. Along those lines, it's much more likely that a drive has failed than a controller.

    It's quite simple to remove the 1.2M floppy from its slot and use it from there.
    The disk is working just fine mechanically. If the disk were unreadable, the system wouldn't hang and tell me the controller was bad. I have a Tandon TM252 drive that *is* bad (severe head crash, all 4 disk surfaces are destroyed), and SpeedStor and HDAT can certainly tell that the drive is not healthy. The IBM drive, a Miniscribe 3012, and my Seagate ST251-MLC-2 all seem to check out with both of those programs. Mind you, these three disks also work in other machines, albeit Pentium class machines.

    I'm thinking the IBM controller has failed, but the Award BIOS is incompatible with the other controllers I'm using. It's certainly possible some component on the IBM controller decided to lay down and die while it was running - I've seen it happen with televisions and radios.

    Yes, it is possible to remove it from it's slot and use it from there, but without leaving it in it's slot, I can't close the case. I'm just leaving it's data cable disconnected until the longer cable comes in the mail on the 20th.


    If I have a floppy diskette that reads fine in one machine, and I put it into the drive on another machine that cannot read it for whatever reason, is it more likely that the diskette has failed or that the drive or controller has a problem? Especially so if I can put the diskette back into the other machine and see it work correctly.

  4. #4

    Default

    Well, I figured out that it wasn't reading or writing to it because the data cable was plugged in upside down and the terminator resistors were not installed. Correcting those two things, the computer can now read and write to the IBM hard disk just fine, even though with many other drives (excluding the ST251) it simply hangs at boot. I'm not sure why.

    My AST Advantage still does not seem to be working correctly. The motherboard has one bank of 18 "511000" RAM chips for 1MB total RAM. The jumpers are set correctly to reflect this, but it only sees 512K without the AST card. With the AST card and 1152K installed on there, with the jumpers installed correctly, the reported total is 1152K. I've lost 1MB of RAM somewhere, half on the motherboard and half on the AST. Any thoughts? Reseating all of the RAM chips did not change anything.

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •