cylinder hard drives

[TRC]

If only they would make hard drives on cylinders instead of platters. A cylinder 15 cm long and a centimeter in diameter would have the same surface area as a standard platter. ( I did not do the math, please check.) In the shape of a cylinder every part of the surface would be spinning at the same speed without any drop in speed depending on the location of the disk surface. I would think at such a small diameter you could greatly increase the speed of the disk surface without it flying apart that would increase read speeds.

 

[CleanOldMan]

The problem with using a cylinder 1 cm in dia is that it would require so high an rpm that the magnetic material would fly off.
For a cylinder to work, it would have to be at least 2 inches in diameter.
Also, a disc is cheaper to manufacture than a cylinder.

 

[Corasik]

Problem with a cylinder would be the seek times, the heads having to travel a 15 cm length?

 

[Jak_Sparra]

well why not put more heads in? or is a dumb question?

 

[nobly]

15cm = 5.9inches.
So if there's 4 platters, we'd need 4 cylinders that are 6 inches long. We'd probably have to redesign cases as well. Plus did you take into account that platters have data on both sides?

If there's no benefit, there's no point...

 

[threep]

Technically, on a rotating disk, all points on the surface ARE rotating at the same speed. If the inside of the disk is rotating at 5400 rpm, so is the outside.

The reason we use disks instead of cylinders is the increase in surface area. A modern hard drive with multiple disks in it has a far greater surface area than a single cylinder.

As an interesting note, the cylinder was actually used decades ago in some of the original storage devices. I don't remember if it was magnetic, or carved permanently... someone with more time than I should look it up.

- 8)

Correction: Decades? Make that more than a century ago:
http://en.wikipedia.org/wiki/Phonograph_cylinder
And for laughs:
http://video.google.com/videoplay?docid=-1171246024685658304&q=techtv/

 

[Jak_Sparra]

you could like have wax in the middle of the cylinder, when the cylinder spins it coats the inside, then a needle rights info into the wax once its set, to format you just melt the wax. it might be slow and a bit unreliable though and also a bit crap and the wax could crack, or melt by accident.

 

[Ycon]

Nope, the motor rotates at 5400rpm. The farther you go away from the platters center the faster the disk is moving.

 

[theaxemaster]

velocity always increases the further from the rotation point you are. The rotation rate (rpm) is the same, but the velocity of a point on the edge of the disk is way way higher than at the center. Inertia is dependant on velocity, not rotation rate. That's why CDs are limited to 54x speeds, because the rotation rate (and hence, edge velocity) required to go higher than that overcomes the structural integrity of the disc due to the high inertia and the discs explode.

Platters work pretty well except for inconsistant seek/read times due to varying locations of data on the disc. The outer edge of the platter has a larger circumference than the inner section, which means you can cover more data per rotation on the edge of the disc than you can in the center. A cylinder would be very consistant with seek/read times since it is uniform but the surface per volume is much lower and the mechanical complexity is much higher to get the same storage area.

Which is why we use platters/discs. The trade off of consistant seek times for capacity per unit volume is more beneficial.

 

[maury73]

Cylinders where effectively used in the far past: before vinyl discs music was recorded on aluminim cylinders!
But it is simply impossible to do that with magnetic media at the present densities.

First of all a cylinder 15cm long rotating at 5000+rpm must be absolutely perfect and so the motor shaft and all of the mechanics: HDD heads "fly" some microns above the disc surface, any small vibration may cause them to hit the surface and be immediately destroyed.
It is possible of course to manufacture such a system (the big electric turbo generators must have this perfection!) but it would cost 100:1 compared to a perfect platter.

Then there is the seek time: you would have to put many heads with indipendent positioning systems along the cylinder, raising costs very much.

Last but not least with platters you can simply stack them one above the other (and the same for heads), but with cylinders you can't. You could make cylinders bigger in diameter, but this raises weight and it's not a good idea having a 1kg cylinder rotating at 15000rpm in you hard drive!