Spiral Layout for Hard Disk Drives

[oberschelp]

(Out of my field, so I apologize if I'm presenting this poorly, or if it is noobish.)

Why are hard disk drives not made with user data and the servo information in a single spiral a la DVD?
I ask because track-to-track seeking is so common, and reducing it from 1ms to 0 would be a big deal.
From my couple hours of research (before coming to you all with this) it looks like the voice coils could handle it. It seems to me you just need to add a sensor to track the platter(s) angle. Then it's just a software problem, right?

 

[xwb]

If you had a spiral architecture, the calculations would be more complicated. To start with, how would you detect the sync point unless the disk had a special kink in it to tell the head where it was.

The problem is that you are reading across a diagonal track so it is very difficult to determine where you are.

Say once you established where the start is. Would you go for single speed or variable speed (like in the old Commodore/Sirius disk drives).

In a spiral, the track/sector would no longer make sense. Everything would be one long sequential stream, possibly divided into tracks and sectors but why bother. It is no longer go to track t, sector s. More like go to bit n, since the head would have to move in both directions to get from bit n to bit n+1.

Say you have a sync point and your spiral works, would the hardware be easier to manufacture. Look back in history. SCSI was more efficient but IDE won because it was simpler. PATA was more efficient theoretically but SATA was so simple that it could be made faster. 1394 was faster but USB was simpler to implement (how many people even know that there is a 1394 port on their PC?).

Having said all that, in the days of SSDs, spirals are probably meaningless as it is now just a block of memory with zero seek and settling times.

 

[oberschelp]

Point by point:

“If you had a spiral architecture, the calculations would be more complicated.”

Yep.

“To start with, how would you detect the sync point unless the disk had a special kink in it to tell the head where it was.”

Did you see in the original post where I mentioned “a sensor to track the platter(s) angle”?

“The problem is that you are reading across a diagonal track so it is very difficult to determine where you are.”

Its 1st year geometry. I think the engineers can handle it.

“In a spiral, the track/sector would no longer make sense. Everything would be one long sequential stream, possibly divided into tracks and sectors but why bother. It is no longer go to track t, sector s. More like go to bit n, since the head would have to move in both directions to get from bit n to bit n+1.”

Yes but the heads are already being adjusted a fraction of a track through feedback from the sync data in existing drives with some engineered expectation that they will remain within tolerance through the sector r/w. So what I'm talking about is an engineered drift.

“Say you have a sync point and your spiral works, would the hardware be easier to manufacture. Look back in history. SCSI was more efficient but IDE won because it was simpler. PATA was more efficient theoretically but SATA was so simple that it could be made faster. 1394 was faster but USB was simpler to implement (how many people even know that there is a 1394 port on their PC?).”

No it would not be easier to manufacturer. Neither are other (bad?) ideas like multiple platters, variable speed, embedded sync data, ...

“Having said all that, in the days of SSDs, spirals are probably meaningless as it is now just a block of memory with zero seek and settling times.”

Then perhaps you should write to the hard disk manufacturers, because I'm pretty sure they are still wasting their time improving hard drives.

 

[MacGyverS2000]

Did you see in the original post where I mentioned “a sensor to track the platter(s) angle”?

Its 1st year geometry. I think the engineers can handle it.

If it were just the platter angle that mattered, the entire problem would be as simplistic as you seem to believe it is. But more things matter... so it's not.

No it would not be easier to manufacturer. Neither are other (bad?) ideas like multiple platters, variable speed, embedded sync data, ...

Go work in the industry for a few years and get back to us. Just because two ideas are both not simplistic in their manufacture does not equate to both being equal in manufacturing cost. Even if both ideas were equal in cost to manufacture, you go with the proven method, not just choose the other one because you came up with the idea.

Then perhaps you should write to the hard disk manufacturers, because I'm pretty sure they are still wasting their time improving hard drives.

Perhaps you should understand manufacturing trade-offs as a whole, because you obviously do not. You came up with an idea that you feel is better, but despite numerous engineers telling you why your idea comes up short, you insist we're all wrong. So why are you here, just to troll? If you think everyone here is wrong, then you go to the HD manufacturers and pawn your ideas. Tell them they've been doing it wrong all of these years. I'm sure they'll see the light when you tell them how incompetent their engineers are, and they'll surely offer you a position as chief of engineering.

Dan - Owner

http://www.Hi-TecDesigns.com