Coaxing more sectors & tracks from a bad ST-251

[Forgotten Machines]

David & All,

I did a read tonight of a Seagate ST-251 with internal damage.  

Not surprisingly, Autodetect did not work, so I did a

./mfm_read -t ../NGEN_ST-251-L114682.raw -c 820 -h 6 -d 1

Followed by a

./mfm_util -I -t ../NGEN_ST-251-L114682.raw -e /mnt/NGEN_ST-251-L114682.bin -c 820 -h 6 -s 16,0 -l 512 -f WD_1006 --header_crc 0xffff,0x1021,16,0 --data_crc  0xffff,0x1021,16,0

Again, not surprisingly, about 50% of the tracks errored out.  Attached is the console output status, so you can get a full picture of the scope of failure.  

And yet, when extracting strings from the resulting .bin file, a surprisingly tremendous amount was read successfully and human readable.  So your device has proven highly valuable even with highly compromised drives.  A partial recovery on such a badly damaged drive with so few action steps is absolutely amazing!

BUT, before I relegate this drive to spare parts and other entertainment uses, I thought I could do a little more to "squeeze" more sectors and tracks from this, just for my own education and experience.  After all, this procedure on a Maxtor XT2190 is what got me into the whole Vintage Computing scene in my own way 8 years ago.  

Anyway, can I do some more options where I get this thing to "beg harder" for valid sectors on every track before moving on?

In the past, I've used "retries", such as

-r 250,20

But only when -a works, and I have only a handful of tracks fail, not every other one like this case.

Or is this my best option here?  Or are there other tricks I can try?

Of course, in my above format conversion line I used -I for ignore_seek_errors.  Will simply removing this make it work better, or just take 2 days longer with the same result?

Thanks for David & Everyone's input here.

Best,

AJ

[David Gesswein]

On Tue, Apr 12, 2022 at 02:03:44AM -0700, Forgotten Machines wrote:
>
> ./mfm_util -I -t ../NGEN_ST-251-L114682.raw -e /mnt/NGEN_ST-251-L114682.bin
> -c 820 -h 6 -s 16,0 -l 512 -f WD_1006 --header_crc 0xffff,0x1021,16,0
> --data_crc 0xffff,0x1021,16,0
>

> BUT, before I relegate this drive to spare parts and other entertainment
> uses, I thought I could do a little more to "squeeze" more sectors and
> tracks from this, just for my own education and experience. After all,
> this procedure on a Maxtor XT2190 is what got me into the whole Vintage
> Computing scene in my own way 8 years ago.
>

> In the past, I've used "retries", such as
>
> -r 250,20
>
> But only when -a works, and I have only a handful of tracks fail, not every
> other one like this case.
>

All the mfm_util parameters can be specified on the mfm_read so it can do
retries.

> Or is this my best option here? Or are there other tricks I can try?
>

With the error messages you attached it is seeing data from two cylinders.
This seems to be a common problem where the head stack is out of alignment
so some of the heads are out of alignment. Normally it gets worse from
first to last head or reverse.

> Of course, in my above format conversion line I used -I
> for ignore_seek_errors. Will simply removing this make it work better, or
> just take 2 days longer with the same result?
>

Don't think removing it will help.

If its worth the effort this may be able to recover the contents.
http://www.pdp8online.com/mfm/microstep/index.shtml