That’s a HDA (Head/Disk Assembly) from an IBM 3380 disk drive. I used to work with those in the late 1980s/early 1990s.

If you look at the picture, you will see what appears to be two linear voice coil head assemblies, one on each side of the platter. That’s exactly what is there! Each disk had two completely separate head assemblies, one on each side of the disk, reading and writing off the same platters.

What this meant was that the operating system could read or write on two different places on the disk at once. This could massively speed up disk operations in certain circumstances.

Suppose you want to copy a file somewhere else onto the same disk. If you were doing that on your PC, it would take a long time, because the computer would have to seek a certain track, buffer a bunch of data, then move the head to a different place on the disk to write it. The head would be oscillating all over the disk doing the copy and chewing up time.

With two heads on the 3380, one head could be reading sequential tracks while the second head could write sequential tracks in a different place on the disk, thus massively speeding up file copy and other similar disk operations.

This was all done transparently by the extremely sophisticated and expensive hardware/firmware in the disk controller. No special programming required. This is just one example of how IBM’s mainframe hardware was so far ahead of even modern PC technology that people have no idea why mainframe programmers have such high regard for old IBM iron.