Comment by Lagged2Death | Hacker News Reader

Comment by Lagged2Death | Hacker News Reader

Parent Full thread

... to reduce density it might have been easier to simply reduce pressure and then create a hermetically sealed chamber...

Disk heads "fly" above the surface of the disk on a cushion of gas that's pulled under the head by the spinning disk. If there isn't enough gas pressure inside the disk enclosure, this method of locating the heads won't work. Wikipedia actually says:

If the air density is too low, then there is not enough lift for the flying head, so the head gets too close to the disk, and there is a risk of head crashes and data loss. Specially manufactured sealed and pressurized disks are needed for reliable high-altitude operation, above about 3,000 m (9,800 ft).

http://en.wikipedia.org/wiki/Hard_disk_drive#Integrity_and_f...

Atmospheric pressure on the top and bottom faces of a 3.5" drive case (~22 square inches) is in the neighborhood of 300 pounds each, and there's still more on the front, back, and sides. A square box of thin aluminum is probably not a great starting point for building a pressure vessel. If the box is filled with a gas at close to atmospheric pressure, you don't have to worry about it.

Helium's specific heat is five times that of air, so it probably offers better heat conduction than air at 1/7th atmospheric pressure would.

I'm sold, but I have an additional question though. Can you please explain why hard-drives follow a flying head approach? Why don't they just perpendicularly fix the reader above the disk, rotate the disk and then use a linear actuator to move the reader back and forth - exactly like optical drives?

I'm guessing that the status quo exists and that my suggestion is inherently flawed because of the precision required to maintain the head at 10 nm above the surface and as such letting fluid dynamics wrestle with gravity might do the trick for you (given that you can make the spindles in a precise manner that the forces cancel out in a way that ensures that the head is exactly at the desired distance above the drive)

That leads to another thought, would the following conjecture be in the realm of possibility? Imagine that you have coated the disk with mystery thing X - which is an etched semiconductor that does something very magical. When an incident beam of light strikes a point on the coating, it releases electrons, which are then controlled through magical mechanism Y etched on X to create a magnetic field that flips the bit. If you do this then you can remove all of this stuff and it might lead to some gains.

"Can you please explain why hard-drives follow a flying head approach? Why don't they just perpendicularly fix the reader above the disk, rotate the disk and then use a linear actuator to move the reader back and forth - exactly like optical drives?"

The head is only "flying" relative to the platter. A hard drive works pretty much like you describe -- a short stack of fast-spinning platters and a read/write head on a swing arm that pivots back and forth over the platters.

A pic: http://msuweb.montclair.edu/~maciakl/blog/img/hd.jpg

They are child's play to take apart (http://turmon.org/Images/celeste-disk.jpg)

I think you should be careful about taking the notion of 'flying' the head too literally. Think of the air as a 'spring' which is holding the head at a precise distance above the platter, the head is pushing "down" and the spring is pushing "up", so if the head is too far away its down force is stronger and it moves closer, if the head is too close the air's "up" force is stronger so it moves away. This is simply a very precise way of placing the head over the platter, which is necessary because the shape of the magnetic field is a sphere and the 'circle' of its intersection with the platter is determined by the distance from the platter by the head.

In terms of the mystery thing 'x', most of what you might think of as candidate materials have been tried. The challenge is to reliably (well at least reasonably so) flip the state of a bit on the substrate in about a microsecond. And do that as cheaply as possible.

It is exceptionally challenging engineering.

Your description is a much more precise way of looking at it. I was trying to reach for the same concept when I talked about fluid dynamics and the forces acting on the head cancelling out so that it rests at exactly the right distance, but I lacked the clarity to state it so neatly. Thanks!

>>> It is exceptionally challenging engineering. <<<

It's also very beautiful... When you look at something like that it's a bit like looking at a work of art - I don't know how to explain it but it is awe inspiring. I wonder how people end up working on such things... How do you get to the point where you can raise your sleeves and create a system as beautiful as this?

I don't think any one person can create a piece of engineering as complex as a modern day hard drive. There are just too many disciplines involved: Physical, electrical and embedded software engineering, applied physics, information theory for encoding the data and recovering from errors. But certainly a single person or group in those fields can be responsible for specific breakthroughs.

Can you please explain why hard-drives follow a flying head approach? Why don't they just perpendicularly fix the reader above the disk, rotate the disk and then use a linear actuator to move the reader back and forth - exactly like optical drives?

The read/write arm needs speed and extremely fine-tuned movement that can't be had from a mechanical actuator. See: https://www.youtube.com/watch?v=Wiy_eHdj8kg

Re: the last part about etched semiconductors:

Keep in mind that the size of a magnetic grain (which stores 1 bit, roughly) is about 8nm. Compare that to the current cutting edge lithography scale of 14nm, which is the size of the semiconductor 'wires'.

Thus, a bit is smaller than the size even of the wires in any semiconductor circuit we can build, which means a system as you describe would have much lower storage density.

Shot in the dark, but might there be advantages simply because there are more molecules flying around than with air? 10nm ground effect flying probably isn't trivial.

Reply on news.ycombinator.com