MFM Hard disk gurus please help

[nunoalex]

Hi

My Microscience HH-725 has suddenly stopped working ((
This is a "Type 2" 20mb hard disk half height 5 1/4 MFM hard disk

6 month ago I formatted it as an ST 225 with a seagate ST11 controller (formatted it as a ST-225 because it is one of the only compatible options on the controller ROM)
The hard disk worked fine, it had a few bad sectors but just a few (20k or so)
I installed MS-DOS 5.0 and it booted fine
I passed spinrite a few times to try to iron out the bad sectors but some always remained... but the drive was working fine

yesterday I decided to install another language MS-DOS and I connected it, (it had its heads parked), and started normally and booted the previously installed DOS fine
Then I swapped to the floppy disks and started installing DOS.. everything went fine until 97%

At 97% installation suddenly an error message "Cannot write to disk - command.com) (can you bloody believe it had to be command.com ? why not fail to write "compare.com" or "setver.exe" or another non critical file... ??
After that it was always downhill.
Could not boot from it anymore obviously since command.com was not saved properly
I run spinrite 5.0 and it started identifying a few more bad sectors... but not many, just a few "new" ones
At this point i could still run spinrite that was previously saved on the hard disk and I could CD DOS and make "DIR" commands

So I then next tried to low level format it again with the ST11 ROM program
It formatted (I could see the stepper motor slowly rotating) - The stepper motor is not stuck it moves freely
But when the verifying phase started it immediately started retrying to read, right on the 1st cylinder and first head ... read retries
I could do nothing more on the controller ROM program

I then swapped to another controller (a 16 bit western digital controller) and run Speedstor from a floppy disk and the controller test passed, the seek test passed, it could move the heads around from 0 to 612 cylinder
But the read or write tests failed immediately with all "bad sectors"
I tried to initialize it with Speedstor but as soon as it started formatting it it returned with the error "Bad command to drive BIOS"

now the drive is completely inaccessible
The ST11 controller reports "bad drive geometry" when I try to format it with ST-225 parameters and speedstor can only do the seek test, all other tests, initializations fail)
I have tryed the controllers on multiple ISA slots, spraied deoxid and swaped the controller cables around...

Did the drive died in front of my eyes?
I find it strange that a drive that was working only with a few bad sectors has suddenly just stoped working like that
The drive is not making any strange noises, mechanically it behaves normally, I can even issue the park command from speedstor

Can some one with MFM experience help ?

thenk you

Nuno

 

[SomeGuy]

When you tried it on the other controller, did you also use different cables?

I've had issues with worn out cables (the smaller data cable) that can suddenly make a drive look like it is dead or developing random bad sectors.

Also, I would sanity check after LLFing using Norton Utilities to read a few random sectors (select "physical drive" to read, no partition or file system needed). If there were a head crash, then usually some sectors on certain heads will read also try toward the end of the disk - if all sectors fail then that is suspicious and points to controller, cable, or logic board failure.

 

[modem7]

The ST11 controller reports "bad drive geometry"

Possibly due to:
- ST11 is acting as a 'dynamic' controller - see [here].
- At power-on time, ST11 is unable to read the first sector on the drive (where it stores the geometry info), or that sector is read but no geometry info found.

Did the drive died in front of my eyes?
I find it strange that a drive that was working only with a few bad sectors has suddenly just stopped working like that

It does sound odd that your ST-225 would partially fail before your eyes. I wrote "partially" because at the least, your ST-225 is still stepping its heads.

It is sounding like a 'data' related problem.

Regarding seeking/stepping. There are certain models of MFM hard drive controllers that when they receive a seek command to seek the heads to a particular cylinder, the controller only steps the drive's heads - the controller not reading any sectors as part of the seek operation. An example of such a MFM controller is the IBM supplied controller in the IBM AT (IBM 5170). If I disconnect the data cable from the drive, SpeedStor's seek test still results in the heads stepping.

That behaviour can lead people astray:
- For example, you wrote, "It formatted (I could see the stepper motor slowly rotating)". Technically, the heads stepping is only part of a LLF operation.
- For example, people incorrectly thinking that because the heads are stepping, all of the cables must be hooked up properly.

How about, for both of your controllers, you only have the control and power cables hooked up, and see if SpeedStor's seek test still works. Report the results back here. If both still seek, the 'but the drive is seeking (and parking)' behaviour is explained, and the diagnostic direction heads to a data related problem.

Another thing I'll mention: Some people think, "I will try another controller, then go through the LLF process, etc. If that does not work, it must be the drive at fault." That logic is often compromised because more than one variable (the controller) gets changed. Sometimes misconfiguration catches people out (e.g. data cable going to wrong connector on the second controller). Etc.

At this point, if you cannot get a LLF to run to successful completion (even with some bad tracks found during the LLF), there is no point in trying higher level functionality, such as reading and writing tests (they rely on a good LLF).