[dunbar]

Originally posted by DanU
That doesn't sound anything like your typical low-pitched head-positioning sounds. I doubt the voice coil positioner is capable of vibrating that quickly to produce those high pitched tones. Given a seek time a few milliseconds, the most they could hope to produce would be a few hundred Hz. Those tones sounded like a couple KHz.

I worked for Unitrode Integrated Circuits back in the late 1990s. Unitrode was working on a chip for a very large hard disk outfit at the time I was there. The head positioning system is not electronically limited to milliseconds. The electronics are capable of delivering electrical impulses which are shorter than a few tens of microseconds in duration. The electrical pulses generate magnetic fields of the same duration in order to interact with the head positioning systems static magnetic structure. Even if the heads do not move, the mechanical structure is still being whacked with those electrically generated, magnetic pulses which are tens of microseconds in duration. The result is like striking a tubular bell. With a head stuck to a platter, you have a beam which can flex and resonate (I think the resonances are higher in frequency than the audible range), that will transmit those magnetic pulses to the platters. Those head positioners are not slowed because of the electronics, they are slowed because of the mass of the arm and because of air frictions; because of those problems, the settling time to reach the target position becomes a problem of how quickly can you mechanically dampen a swinging pendulum.

Imagine: when you swing a baseball bat as hard as possible, you cannot stop the tip of the bat when it is pointing exactly at the pitcher (on the first attempt).... you over shoot the mark; then you pull back towards the pitcher, and maybe overshoot that realignment attempt as well. That is what causes milliseconds of seek times, not the drive circuitry. That and the air friction of trying to move the heads laterally against the revolving air immediately above the platters.

Also, there's a definite pure, harmonious quality to it. To my poorly trained ear, it sounds like a pure sine wave. Musically, it sounds like a perfect fifth followed by a major third interval. Head movements and their associated vibrations sound more like noise due to all the random interactions. I still think it sounds like a small speaker inside, probably a little piezo tweeter.

As for the sounds of a drive in normal operation, the randomness of the head positioning sounds is more based on the randomness of the distances involved in seeking to the next location of data than it is based on 'random resonances', because the mechanical structure is not changing from head seek to head seek. Hint: we defrag a drive because data gets placed randomly around the disk.

I can't fathom why someone would put a sound transducer into a sealed hard disk box, especially when the device needs electrical connection to the PC board which is outside the sealed cahmber. In fact, why not put the transducer onto the PC board, which is where the transducers are already designed to be placed, and would also thus be better heard?

BTW, I forgot to add that it's a good idea to put the HDD in a bag, suck out as much air as you can, then stick it in the freezer/fridge. This will keep too much moisture from condensing on the HDD.

Will remember that idea!

By the way, the platter motors are certainly driven by high frequency, so a stuck platter structure would also have motor pulses transmitted to the platters where the platters can vibrate.

Anyway, the point is now moot, the drive is spinning for him.