floppy drives

[chiplonkar]

I have read some where that the 3.5 inch floopies have tracks are seperated by 0.2 mm and each track has a width of 0.115 mm.

I want to know if it is possible to control the movement of the head in steps which are smaller than the gap between any two track.

When INT 13 is used to read a particular track, the head must be moving to some calculated distance from the reference ( track 0 ).If, due to mis alignment , the head falls in the area in between two tracks, then error is reported. Can I , using some other available interrupts or using Floppy Drive Registers (Ports 3F5 and 3F4 ) , control the movement of the head in finer steps ?

On getting the error that "Sector is not found" or " Address mark not found" ,I want to increment the head position in fine steps till a track is read. Purpose is only to have a deeper understanding of the process.

Can any one suggest a way ?

 

[automatic2]

Bios only allows you to send head, track, cyl requests. The floppy disk controller passes this info along to the drive itself. This is where the info is translated into head steps. Head alignment is possible with the right tools, with the adjustments made on the drive itself.

 

[chiplonkar]

Yes, I do agree to what you are saying. I know that BIOS is accepting only integers for head ( track wise ) movement.Therefore, I am asking the way to get finer steps of head movement.I have tried to write a program for head alignment. If the track number read is not the desired one, it will show the actual track number read.I have recorded known data on each track using a good drive. So looking at the data read, the program was supposed to tell the user the direction in which he should move the head ( align ).However, I have found that if the head falls in the blank areas, then nothing is read, indicating only the error. This does not give clue in which direction the head has to be moved. Therefore, I wanted smaller steps, which will take the head to nearest track and then tell the user where his head lies.
I do not know how to attach the program written to this thread. Otherwise you would have seen what I wanted to achieve. If I have your email, I can send it for your comments.

 

[IRstuff]

I think that automatic2's response is correct. There is no there, there. The following links shows what you are asking for does not exist:

http://cvsweb.netbsd.org/bsdweb.cgi....c?rev=HEAD&content-type=text/x-cvsweb-markup

http://members.tripod.com/~oldboard/assembly/765.html

Since head misalignment is a recurrent problem, if it could have been done in the last thirty years, it would have been. head misalignment has been problematic since day one.

The following link likewise confirms that head alignment can only be accomplished on the drive and that it's not cost effective:

http://cma.zdnet.com/book/upgraderepair/ch13/ch13.htm#Heading37

TTFN