R: Winchester SA1004 file recovering

[Enrico email.it]

I wish apologize me. the work was done by a dear friend and I had mistakenly
understood that he had used the Fluxengine card instead he used this card
http://www.pdp8online.com/mfm/ with add-on daughter card interfacing with
the SASI controller SA-1403. Since, despite repeated readings, we are not
able to correctly extract the data from the second Winchester Hard Disk, we
will probably operate by trying to rely on someone able to intervene on the
platters of the SA-1004 Hard Disk, hoping to be able to recover as much as
possible.

Enrico

-----Messaggio originale-----
Da: cctech [mailto:cctech-bounces at classiccmp.org] Per conto di David
Gesswein via cctech
Inviato: giovedì 4 febbraio 2021 01:45
A: cctech at classiccmp.org
Oggetto: Re: Winchester SA1004 file recovering

On Wed, Feb 03, 2021 at 10:43:48AM +0100, Enrico email.it wrote:
> 
> 
> A very dear friend first read in raw mode a 10MB Winchester hard disk
model
> Shugard SA1004 from which the files were correctly extracted and then set
up
> a test bench from which you can see in this video the complete system
> operating with the HD ed il controller Shugart SASI SA1403:
>
https://www.dropbox.com/s/d9iacfgrnexso3o/GP%20T20%20con%20HD%20banco%20prov
> a.mp4?dl=0
> 
> Despite we having attempted different readings (even using cards from the
> first hard disk), using the same FLUXengine card
> (http://cowlark.com/fluxengine/doc/building.html) the raw file still does
> not present the Hard Disk directory, instead  we read the information of
the
> files present but there may be bad bytes in these files and in the rest of
> the disk as well. 
> 

I have only heard of FLUXengine being used for floppies. Can it handle
hard drives? I hadn't heard of other boards that can read MFM hard drives
other than mine http://www.pdp8online.com/mfm/.

Other than the drive being faulty the biggest reason reading fails is
the heads are no longer on track. You can tell that by looking at a
histogram
of the flux transition timing for a track. If they show peaks with
good nulls between them the data is likely recoverable. If they don't
you have to do something to shift the head position.
Example histogram from by unit for ST506 type drive.
https://groups.google.com/g/mfm-discuss/c/kjKez8vfapU/m/4XamyQJgCAAJ
The SA1004 is a stepper drive. Some people have had success on other stepper
drives with putting some rotation force or drag on the stepper to shift the 
heads. Creating a microstepper driver would give better control.

May be able to say more if you provide more information on what you have 
done and checked to try to recover the contents.