Can't read any floppy disk on my XT

[Deksor]

Hello everyone.

Here's my problem : my XT can't read floppy disks anymore ... that computer is a clone and uses that motherboard https://picclick.com/VIP-TMX-10-11-Motherboard-w-AMD-8088-1-Processor-192357656416.html
When I first got it 1.5 years ago, it came with a 360K floppy disk drive and a 20MB hdd. I didn't have 360k floppy disk back then, so I installed a 1.44MB floppy disk drive in it and set up the 360K one as B.

I could boot on 360k formatted 1.44MB floppy disks although sometimes it didn't boot on them and I couldn't figure out why ... but now it has gotten worse and it doesn't want to boot on any floppy disk anymore ... I tested everything :
- tested the floppy disk drives in other computers, the drive were OK.
- tested with other cables, the computer is still acting like there was no disk
- tested with another IO card, still no changement in the behaviour ...

So I guess the mobo has some issue, but I can't figure out what's wrong ... The computer still boots fine on the 20MB MFM HDD (at least before I screwed up autoexec.bat/config.sys so now it just crashes everytime I boot, so I need to boot from another source to fix them as it's MS-DOS 3.3 so I don't think I can skip these like in later iterations of MS-DOS).

When I could still use MS-DOS 3.3, I tried to format disks on A and B, but it won't work, it just says "general failure" or "bad track 0".

And when it tries to boot on floppy disks, the led lights up, but the disk doesn't spin at all, and I don't think the heads are moving either ... Obviously I tested with other PSUs as well, but that didn't change a thing.



Can anyone help me ?

 

[sergey]

Not sure how it worked before

But here is what I remember:

3.5" 1.44 MB floppy drives use a hole on the bottom right side of the diskette (or a lack of it), to determine whether the diskette is 1.44 MB (high density) or 720 MB (double density).



Also the BIOS of an IBM PC/XT/clones knows nothing about high density diskettes. It assumes that everything is double density (8 or 9 sectors per track), and doesn't care about the number of tracks.

And so, if you were to format a 1.44 MB diskette as 360 KB one in an XT, I am pretty much sure it would fail completely, or work very unreliably. That is because the drive will have wrong settings - the auto-detected disk density will not match the settings BIOS is trying to use.

You can try making it work by covering that bottom right hole with a piece of a sticky tape, and fooling the drive into thinking that you're using 720 KB / double density diskette. You will need to re-format your diskettes. That usually works for me, but I've heard that some folks had reliability issues, likely caused by the difference in the magnetic coating material between 720 KB and 1.44 MB diskettes.

I think you can format the floppy with a covered HD hole either as 720 KB or 360 KB (but that would be a waste of half of the storage space )

 

[Chuck(G)]

And when it tries to boot on floppy disks, the led lights up, but the disk doesn't spin at all, and I don't think the heads are moving either ... Obviously I tested with other PSUs as well, but that didn't change a thing.

Power the system on without a disk in the floppy drive. Ignore whatever appears on the screen. Watch the floppy drive as you insert a disk. The motor should come on momentarily to "seat" the disk as you insert it. If it doesn't, you probably have a problem with your drive electronics. If it does spin, at least the motor circuitry works and perhaps we can go from there.

 

[Ruud]

Power the system on without a disk in the floppy drive. Ignore whatever appears on the screen. Watch the floppy drive as you insert a disk. The motor should come on momentarily to "seat" the disk as you insert it. If it doesn't, you probably have a problem with your drive electronics. If it does spin, at least the motor circuitry works and perhaps we can go from there.

He tested the drives in another system and there they worked fine.

My first idea was the 8237 DMA IC. But that one is needed for the HDD as well. But what is the definition of "crashing"? If it is a corrupt autoexec.bat and/or config.sys, they can be avoided by pressing F8 during booting. DOS then skips these two files and should come with a prompt. If not, then it could be the 8237 again.

The part the FDD and HDD don't have in common is INT 13h in the BIOS. If the (EP)ROM is partly damaged, I know from experience that one wrong bit is enough, then the PC can boot from the HDD but gets stuck when booting from FDD.

Kind regards, Ruud

 

[sergey]

My first idea was the 8237 DMA IC. But that one is needed for the HDD as well. But what is the definition of "crashing"? If it is a corrupt autoexec.bat and/or config.sys, they can be avoided by pressing F8 during booting. DOS then skips these two files and should come with a prompt. If not, then it could be the 8237 again.

The part the FDD and HDD don't have in common is INT 13h in the BIOS. If the (EP)ROM is partly damaged, I know from experience that one wrong bit is enough, then the PC can boot from the HDD but gets stuck when booting from FDD.

8237 DMAC is also used for memory refresh. So most likely the DMA controller is working properly. (It is very unlikely that only the DMA channel 2 is broken)

I think IBM BIOS checks its own checksum. If the ROM is corrupted it won't boot at all.

I am wondering, when you try to format a disk, do you specify the /T and /N parameters? E.g. "FORMAT A: /N:9 /T:40?

And another thing - the terminator on the 5.25" drive. If you previously had the only one 5.25" 360 KB drive in your system it is likely it has a terminator, usually a low resistance one, e.g. 150 ohm.
Apparently you've moved your 5.25" drive to the middle of the cable, and it no longer needs the terminator. The terminator should be removed or it might interfere with the operation.
The 3.5" 1.44 MB (and 5.25" 1.2 MB) drives normally have a soldered on / non-removable 1 kohm terminator, so the floppy signals still will be terminated properly.
See this diagram

 

[SomeGuy]

The described problem doesn't really sound like a motherboard issue.

A few things to double check:
- Exactly what model of floppy controller(s) are you using?
- Are you using a proper "twisted" floppy cable?
- What are you using to connect the 3.5" drive? An adapter, or a cable with both types of plugs?
- Does your secondary 360k drive have a terminator resistor pack installed?
- Have you tried with just the 1.44mb drive attached?
- Are the XT motherboard switch DIP settings correct?
- Are you covering the density holes on your 3.5" disks?
- What kind of system, drive, software, and source images are you using to make the 3.5" boot disks?

Drives not spinning usually indicate a cabling issue.

You can test for head movement by carefully manually moving the drive heads while powered off, and seeing if the drive seeks track zero at boot.

Also if you were trying to write "360k" disk images to a 720k 3.5" disk, some disk tools might incorrectly try to double-step. ImageDisk will give you more control, however you should be able to write and boot from 720k disk images just fine with less hassle.

If you can get other software on the hard drive again, you might try a different disk formatter. Many third party disk formatters will format 720k disks fine even when the DOS formatter gets confused.

 

[Chuck(G)]

He tested the drives in another system and there they worked fine.

He said:

And when it tries to boot on floppy disks, the led lights up, but the disk doesn't spin at all, and I don't think the heads are moving either

So the first question is "is the drive getting the DC power that it needs for the motor?" My test was just to test that.

If the drive spins and indexes the disk, then it's functional, at least in the motor and DC supply. Given that the motor control is completely independent (it's a latch at I/O address 03F2h), it's not the 8237 or any other LSI chip. The next thing to do is to see if MOTOR ON/ is being asserted at the floppy interface. As this is completely independent of other signals (you don't need drive select, for example). If it's not, then it's either the cable or what's usually a 7432 driver IC, though the exact configuration varies with the specific controller.

Once you can verify that the disk spins on demand, you can then chase down the other issues.

But first things first.

 

[Deksor]

Thank for all these answers !

I used proper 3"1/2 DD floppy disks for that computer so no need to cover a hole ^^

About the motor not running it could be an incorrect memory, last time I turned that computer on, it was ~3 month ago. I'll take it out of storage in few days ... (I'm sick right now ) But one thing is sure, the PSU isn't faulty because it behave the same with another PSU.

- Exactly what model of floppy controller(s) are you using?

I don't know, I tried several ones. Some were 16 bit IO cards, the one that's currently inside is an 8 bit IO card with RTC. All of them used to work with floppy disks.

- Are you using a proper "twisted" floppy cable?

Yes, all of my cables have the twist

- What are you using to connect the 3.5" drive? An adapter, or a cable with both types of plugs?

A cable with both types

- Does your secondary 360k drive have a terminator resistor pack installed?

That's highly possible

- Have you tried with just the 1.44mb drive attached?

Yes, and I also tried the 360K alone, but it didn't change much ...

- Are the XT motherboard switch DIP settings correct?

I can't check right now like I said, but I don't think they're incorrectly set.

- Are you covering the density holes on your 3.5" disks?

Like I said, no need for that, I have DD disks

- What kind of system, drive, software, and source images are you using to make the 3.5" boot disks?

I don't remember, I think that was winimage. But that's not a big deal now as I have 360K floppy disks, so I could use these instead.

I think IBM BIOS checks its own checksum. If the ROM is corrupted it won't boot at all.

It's not an IBM BIOS, it's a Phoenix Bios, but I presume it does the same ...


One thing to note is that there used to be another RTC io card with a varta battery. The RTC card was completely busted, and the ISA slot it was in was also quite corroded, but the mobo stayed mostly fine. Traces didn't get damaged. One chip was right underneath it and got a tiny bit corroded so I desoldered it to see if anything could be wrong underneath it but nope, so I installed a socket and I put the chip back in place. The only spot I can think of it could have damaged anything would be underneath the ISA slot, but the ISA slot still works properly and even changing the floppy disk controller to another ISA slot doesn't work so I don't think corrosion got through the ISA slot. Basically the corrosion only made the solder joints to not shine anymore, but it doesn't seem to have eaten any trace.


Anyways, once I recover, I'll take some photos, make several tests or even shoot a video so you'll see exactly what's happening.