IBM PC XT hard drive initializes twice, then boots to BASIC

[Rubix]

Hi all, I repaired this IBM PC XT today and it properly boots from floppy. However, booting from the hard drive doesn't work. The drive initializes a second time and then the computer boots into BASIC. I reconnected the cards and checked the jumper on the drive. All seems to be correct. Does anyone know what could cause this? The initialization sound seems to be OK.

 

[Plasma]

Are you sure the hard drive is good? The "initializes twice" part makes me suspect the BIOS can't read the boot sector.

 

[Chuck(G)]

In particular, when there's no boot floppy, what message does the BIOS provide?

 

[Rubix]

Are you sure the hard drive is good? The "initializes twice" part makes me suspect the BIOS can't read the boot sector.

I'm not sure. Is there an application you'd recommend to check that? Norton Utilities?

 

[Rubix]

In particular, when there's no boot floppy, what message does the BIOS provide?

It doesn't show any error code but it goes straight to BASIC.

 

[Chuck(G)]

So it's not seeing the drive at all. Could you be having an issue with your HD controller? I assume that it's not plugged into slot 8...

 

[NoPizzaTonite]

Hi all, I repaired this IBM PC XT today and it properly boots from floppy. However, booting from the hard drive doesn't work. The drive initializes a second time and then the computer boots into BASIC. I reconnected the cards and checked the jumper on the drive. All seems to be correct. Does anyone know what could cause this? The initialization sound seems to be OK.

Your system isn't recognizing the hard drive. If you boot from a floppy, can you see the hard drive or its partitions?
Was the hard drive working before? If it wasn't and unless you have something on it you really want, you might try reformatting it via the MFM card (i.e., I use g=C800:5 in Debug for my WDXT card) - look up your MFM card documentation.
Also, at least on the WDXT card BIOS needs to be 10-27-82 or later.

 

[modem7]

Hi all, I repaired this IBM PC XT today ...

The drive looks like a Seagate ST-225.
The hard disk controller is an 'IBM Fixed Disk Adapter', either of variation #2 or variation #3, as shown at [here].

Are you sure the hard drive is good? The "initializes twice" part makes me suspect the BIOS can't read the boot sector.

I'm not sure. Is there an application you'd recommend to check that? Norton Utilities?

You could try the RAYXTMFM tool, available at the web page at [here]. On that page is a link to a example screen output. In that example, take a look at 'Part 5 of 5'. You are expecting all of the sub tests in part 5 to be successful.

If it wasn't and unless you have something on it you really want, you might try reformatting it via the MFM card (i.e., I use g=C800:5 in Debug for my WDXT card) - look up your MFM card documentation.

So, a low-level format (followed by partitioning and a high-level format).
Options for performing a low-level format using the IBM Fixed Disk Adapter are included at [here].

 

[Rubix]

Wow, thank you so much! I love your website by the way! It's helped me out so much over my years of restoring vintage computers.

You're right about the ST-225. The adapter is variant #3 because it has the switch block. I already checked the settings, and they are correct.

I am going to try the RAYXTMFM tool, that looks exactly what I need to get some insights on what's going on. I wanted to write a 5.25" disk but the Dell I wanted to do that with suddenly stopped recognizing keyboards. I'll dig up another machine with a both a 3.5" and 5.25" floppy drive and I'll post the results here.

In terms of low-level formatting, I hope that I can avoid that in order to save whatever data is on the hard drive right now. I don't know what's on it but I'm super curious to find out. I read that Spinrite can do a non-destructive low-level format somehow.

So it's not seeing the drive at all. Could you be having an issue with your HD controller? I assume that it's not plugged into slot 8...

It's not in slot 8. I tried swapping the floppy controller and the HDD controller but that didn't make a difference. It could indeed be the controller, although I wouldn't expect those to go bad.

Your system isn't recognizing the hard drive. If you boot from a floppy, can you see the hard drive or its partitions?
Was the hard drive working before? If it wasn't and unless you have something on it you really want, you might try reformatting it via the MFM card (i.e., I use g=C800:5 in Debug for my WDXT card) - look up your MFM card documentation.
Also, at least on the WDXT card BIOS needs to be 10-27-82 or later.

I booted to MS-DOS 2.0 and it didn't recognize C:. It doesn't seem to find it at all.

 

[Chuck(G)]

Have you tried booting from floppy and running DEBUG to read the first sector from the hard drive? I've detailed the process several times here, but can do it again if you're at a loss.

 

[Rubix]

You could try the RAYXTMFM tool, available at the web page at [here]. On that page is a link to a example screen output. In that example, take a look at 'Part 5 of 5'. You are expecting all of the sub tests in part 5 to be successful.

I ran the application and that did show an error: "INT 0Dh is not pointing to an address within the ROM of the MFM controller at C8000." Would that mean that IRQ5 is conflicted?

Have you tried booting from floppy and running DEBUG to read the first sector from the hard drive? I've detailed the process several times here, but can do it again if you're at a loss.

I am only familiar with the command to low-level format the drive. I'll search your other posts to see how I can do this.

I also ran FDISK and that simply says: "Error reading fixed disk."

 

[Rubix]

I ran Check!It and it reports the HD controller to be on IRQ 5, as it should be. It also reports a 20 MB hard disk. Spinrite can't access the hard disk at all.

 

[modem7]

I read that Spinrite can do a non-destructive low-level format somehow.

Spinrite can only do that if it can read the data currently in the sectors.

I ran the application and that did show an error: "INT 0Dh is not pointing to an address within the ROM of the MFM controller at C8000."

The CRC of the IBM Fixed Disk Adapter's BIOS ROM is okay. Well, at least the first 4 KB, which is what the ROM indicates is in use within the 8 KB sized ROM.

One of the things that happens at power-on is that the motherboard POST executes the 'initilisation' code within the BIOS ROM of the IBM Fixed Disk Adapter. That initilisation code performs some Power-On Diagnostics (shown at [here]) and does some other things, like adjust various interrupt vectors.

One interrupt vector that is changed is the 0Dh one, which gets pointed to a routine on the BIOS ROM of the IBM Fixed Disk Adapter. That is not happening for you.

If you had earlier wrote that you are seeing a 1701 error during the POST, that would explain why the initilisation code didn't get to the point of altering the 0Dh interrupt vector, but you are not seeing a 1701 error !!!

I ran Check!It and it reports the HD controller to be on IRQ 5, as it should be.

Conflicting with the RAYTEST tool. I'll have to have a think about that.

Maybe it's time to try a low-level format.

 

[Rubix]

Even Norton Disk Doctor couldn't repair it... so out of options, I did a low-level format. I hate that now I can never find out what was on it. But the good news is, that the hard drive works great! The XT boots from it into PC DOS with no issues. Only two bad sectors reported by both NU and SpinRite. Thanks for all the help!

 

[Rubix]

Conflicting with the RAYTEST tool. I'll have to have a think about that.

It might be that that info is based on the controller type and setting and not on the drive itself.

 

[Chuck(G)]

Stupid question, so please forgive me...
Was the hard drive originally paired with this particular controller? That is, did this exact pair (hard drive+controller) have a history of working together?

 

[Rubix]

Stupid question, so please forgive me...
Was the hard drive originally paired with this particular controller? That is, did this exact pair (hard drive+controller) have a history of working together?

Not a stupid question at all. I got the XT about a year ago after I asked a gentleman at a thrift store if he could keep any really old computers for me. Normally, they recycle these but he kept this one apart for me. Unfortunately, I don't know the history of it but the configuration you see in the first picture I posted is how I got it. I didn't change anything about the hardware, except for repairing the motherboard. Since the HDD controller is a Xebec/IBM controller that can be specifically configured for the 20 MB ST-225, and the hard disk has an IBM sticker on it, I'm assuming that both are original to this computer.

This also means that I don't know if they ever worked together, but I think it's very likely they did. I don't know what happened to it that made it unreadable but everything works fine again after the low level format. It's just that looking at super old hard drives is like archeology to me. Seeing what they used the computer for. And it has a modem and RTC installed. It would be nice to have drivers and applications for those.


That NEC Multisync is from the same year as the XT btw (1986) and came with the computer. The video card is an EGA card.

 

[twolazy]

Personally on XT's I use Spinrite. Just throwing that out there... And are you allowing the drive to warm up before you try to test it? Let the drive run idle at least an hour before you try anything. I've had that work for me with older 225s. Sometimes they just need to get hot after they sat a while.

 

[Chuck(G)]

What had occurred to me was that the hard drive had its origins on a different controller. Eight bit MFM controllers in particular are notoriously incompatible with regards to encoding methods between manufacturers and even different models from the same manufacturer.

If it were the case, it would account for the strange behavior you experienced.

Just thinking aloud.

 

[Rubix]

Personally on XT's I use Spinrite. Just throwing that out there... And are you allowing the drive to warm up before you try to test it? Let the drive run idle at least an hour before you try anything. I've had that work for me with older 225s. Sometimes they just need to get hot after they sat a while.

Which version of Spinrite do you use? I tried v2 but that didn't do much. Is 3 much more advanced? Maybe I didn't know how to use it but it didn't seem to do much.
In terms of heating up the drive... I did have it on for a while. Not for an hour, though. I think after about 10 minutes I did a soft reset and it still didn't read properly.

What had occurred to me was that the hard drive had its origins on a different controller. Eight bit MFM controllers in particular are notoriously incompatible with regards to encoding methods between manufacturers and even different models from the same manufacturer.

If it were the case, it would account for the strange behavior you experienced.

Just thinking aloud.

You're right that it could be that either the controller or the drive has been replaced sometime in the past. I got it as a complete machine, though. It's hard for me to imagine that someone would replace parts with genuine IBM parts and never got it working and then left it at that. I don't know the history of the PC though, so I can't rule that out. Seeing how many steps a low-level format takes, running DEBUG and typing in line after line, it is possible that someone just gave up after having replaced either the drive or the controller and put the PC away, eventually donating the whole thing to that thrift store. I'm glad I asked the gentleman there to hold those old machines for me, otherwise this one would've been recycled.