Build a Disk Drive for under $20!
|"So", you say, "There's got to be a
catch!". You're right, there is a catch. It has lousy information density and the
worst cost-per-bit in the history of disk drives. And the transfer rate is somewhere in
the range of decibits per minute. So why build it? For teaching, of course.
I teach a course at the Carnegie Museum of Natural History on "Your Computer from Inside". It is designed for the 12-14 year old audience, and I cover topics such as how a transistor works, how a processor chip works, and, of course, how a disk drive works.
Disk drives are mysterious. They sit there and spin and produce only noise, but somehow hold information. I built a disk drive large enough to show what is going on.
The basic disk drive is a "Lazy Susan" ball bearing with a plywood plate on it (this turntable unit was the second most expensive component, at $7.50). It is marked off for 18 positions (20º), of which 16 are drilled. The two unused positions form the "intersector gap". Some very long spikes are inserted in the 16 holes (it has a whopping total of 16 bits of storage! One sector per cylinder, one cylinder per drive.)
Using a magnet, I stroke the spikes to get patterns of N-S or S-N magnetism. The "read head" is a compass, the most expensive part ($10). The feet cost $1.58. Everything else, including the plywood, was lying around somewhere in my basement so was "free".
The only thing to remember is that you must not use any magnetic material near the "read head". The compass came with two holes on one edge to tie a lanyard to; instead, I used two nylon screws (6/32, with nylon nuts, that I had laying around) that exactly fit the holes in the compass, to keep the compass in alignment. Brass or aluminum screws would work equally well, but steel screws will mess up the functioning of the "read head" because their magnetism will affect the compass (even if they are completely degaussed, the magnetic characteristics of iron will cause the magnetized compass needle to swing towards them). With a bit more work I could have built a fitted cradle to hold the compass, but my solution was easier, given the materials I had on hand (if you do this, use glue, not nails, unless you have little brass nails!)
To use it, ideally there are four students involved. The "drive motor" student slowly rotates the turntable in the direction of an arrow drawn on it. After the sector gap, the "read head electronics" student reads off the compass position as each nail passes. This information is relayed to the "disk controller" student, who tells the "memory controller" student what to write on the board (the compass needle is "white" and "red" so the memory controller writes W or R on the board).
Finally, when the transfer is complete, the "disk controller" comes over and taps me on the shoulder (I have been ostensibly reading a newspaper). This is the interrupt that tells me (the CPU) that the transfer is complete. The pattern I magnetize is the ASCII code for my initials, "JN".