Hard Drive Audio

Created
04/22/06

Related Items
none

Materials
Dead hard drive
CD or MP3 player
Copper clad
ICs: LM386, 7805
100K pot
Assorted R & C

A hard drive can be modified to produce sound and play music like an incredibly crappy speaker.

Background

A hard drive consists of one or more rigid disk platters, read/write (R/W) heads, a mechanism to sweep the R/W heads across the platter radius, and support circuitry. The platters are spinning, so any data location can be accessed by positioning the R/W heads over the correct 'track' and waiting for the platter rotation to bring the data under the heads. Early hard drives used a stepper motor to position the R/W heads. Increasing storage requirements and decreasing physical volumes eventually (about when drives reached the 100 MB mark) demanded a precision unattainable with stepper motors, and a new actuator mechanism was brought into service. The industry chose to use a coil of wire, suspended in a strong magnetic field, which would generate lateral motion under precise control of a servo circuit which regulated the current in the coil. Because this device resembles the mechanism used in most speakers, it is called a 'voice coil.'

With a name like that one can't help but to wonder if this coil could be hacked to produce sound. In fact, I didn't have to wonder; I remembered seeing mentions of such things on the Internet. Unfortunately, the day that I made this wasn't a good day for my ISP and I had no Internet access. This led to boredom, and boredom led me downstairs to my Underground Lab of Doom to rip open some hard drives. Now that I have internet, I have found: Afrotech's Hard Drive Speakers. Anyway, on to my own hard drive abomination.

First I needed an audio amplifier. Nothing special here; the most basic use of an LM386 audio amp IC:

Amplifier Schematic

The goofy sickle-shaped thingy on the left is my symbol for a signal input. Expected gain is around 50. Honestly, I have no idea why I put a 5v regulator on there; this should run perfectly fine powered directly by a 9v battery. Potentiometer R1 provides volume control by attenuating the input signal. L1 is the voice coil in the hard drive. I have absolutely no idea what the coil power rating is, but this circuit didn't damage it.

Construction

The amplifier was constructed in dead-bug point-to-point fashion on copper clad. Not as neat as a printed circuit, but only requires 1% of the effort, is reconfigurable, and fast.

I removed and discarded the circuit board from the bottom of the hard drive, and removed the top cover to expose the guts of the beast.

Apparently I didn't get any close-ups of the voice coil assembly. It is supplied by a flexible ribbon cable that carries data traces to the R/W head driver and power traces to the voice coil. I located the voice coil power traces and followed them to a connector that brought all these signals out to the underside of the drive case, where they once attached to the circuit board. I cut away the connector pins I didn't need, leaving only the two voice coil pins. To these I soldered the output wires from the amplifier.

I connected a portable CD player and 9v power, and lo and behold, the R/W heads twitched and there was the sound of music. Not terribly good music, but quality was high enough that I could understand lyrics in most of the songs I played. I noted a definite change in tonal character when the R/W head arm was moved between inner radius and outer radius. Outer areas of the platter reproduced deep sounds more faithfully than the inner areas. Sounds are louder when it is positioned close to the outer edge of the platter.

The system produces sound even in a hard drive that has had the platters removed. In fact, the coil alone will produce sound when brought near a strong magnet. However, music is much louder with the platters in place. I think the lateral vibration of the heads is setting up sympathetic vertical vibrations in the platters, which are then able to 'pump' greater quantities of air due to their large surface area. Not sure how to go about testing that idea though.