This happened on two diskettes, thus it is not caused by the media defect. I had formatted two 360K diskettes using a Windows XP machine recently, and noticed a peculiar behavior: heads start at track 0, to somewhere at the ½ of the surface towards the center in double-track step, then heads go back about 20 tracks, and perform formatting in single track step up to the position where they retracted, and then continue as usual in double track step. I do not have a 360K drive, thus I cannot perform trial-and-error to empirically find out the best way of dealing with formatting. The goal is having a diskette formatted using a 1.2M drive to be portable to a 360K drive as much as possible. But none lays out the algorithm in where heads must be positioned and in what order. Lots talk about the differences in track size, magnetic material, how to degauss the media. Having reviewed a number of sources, I cannot find the algorithm to format a 360K diskette using a newer 1.2M drive. This is a low-level tech question, as I am developing a driver for an FDC at its command-set level.

3 1 2 floppy disk not formatted