Why not use multiple read heads per arm instead of 1? One can set them up in a multitude of ways and the additional cost should be minimal(just the extra metal and heads).
One has the potential to speed up read/write access times 20fold(say 20 heads stacked in an array) and reduce mechanical failure(the arm has to move much less of a distance).
Alternatively instead of a radially arm one could have a linear one. This should make it much easier to move and possible higher resolution and very little wear and tear.
In fact there maybe a better way to completely remove the motion of the arm and increase read and write speeds drastically(1000 fold or more). Probably cost prohibited though.
As far as I can tell it would require minimal cost, increase drive rates drastically, and increase life span.
Why hasn't this been done? I've read seagate did a drive with multiple actuator arms but I'm talking about one arm with multiple heads in which case the only added cost is in software + the heads. The design can easily be amortized and I'd imagine one could easily get 20x increase without much work or additional product cost.
I know about SS HD's. This isn't a question about which is better.
I'm not at all certain, but I believe that radial expansion of the platters due to temperature changes would result in loss of head/track alignment which couldn't be servoed out.
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
Remember: Once you go over the hill, you pick up speed
Well, you are proposing something that the computer industry used 60+ years ago. I used fixed-head swapping disks on PDP-11s in the 1970s. Multiple heads per moving arm have been used, too. Did you think that your insight was somehow original?
This was the original problem by seagate(or whatever company that originally did it before seagate got them). If the platters expand that much and it cannot be correspondingly compensated for by the expanding arm(as it too will expand in a similar manner so that everything relatively should be the same) then you might be right.
That is, suppose the radial expansion is non-zero. The distance between the RW heads would then not be the same before and after which would result in mis-alligned tracks. If the inter-head distance is fixed it could not be compensated for.
The only way to deal with it would be to reduce the effects of temperature or to allow variable distance between heads(might not be that difficult).
In any case I saw site about a rectangular disk that had an arrya of RW heads that moved slightly relative to the media. I guess this would be much more effective as it eliminates the any real motion(small vibrational motion is all that is needed). Ofcourse unlikely it will come out on the market but it's nice to know someone tried to improve upon the idea.
Now that I think about it, HD companies have an incentive to make sure such technologies don't make it as they have billions invested in the manufacturing of their crap.
Your ideas may seem original to you, but are actually decades old.
A few minutes googling, followed by some simple math, would tell you why head-per-track disks aren't practical.
I'm not sure about the math of putting multiple heads per arm. Likely the added mass would slow down seek time, and rotational latency may be the real limit anyhow. Since nobody seems to do it any more, it must not be a good idea. It's not as though the idea has never occurred to the folks at Seagate or WD.
Seagate was originally called Shugart Associates. They were sued by Shugart and forced to change their name, so they are still the same company. Any other idiotic comments you'd like to make?
--
You can't fix stupid. You can't even put a Band-Aid? on it, because it's
Teflon coated.
For one thing, the cost of the heads (and the per-head backing electronics) is not minimal... adding even a couple of heads per arm would add significantly to the cost of the drive.
With modern drives, the average head-seek time is actually no more than the average "rotational latency" (one-half revolution of the disk). If you were to load up the arms with so many heads that you could drop the average seek time by 10:1, you would still have the rotation time to deal with... and thus the average time-to-access a sector would decrease only slightly. 10 extra heads, for a net decrease in access time of only 10-20% (i.e. average access still
80-90% of what it was) would be a very poor investment.
You'd also make the arms quite a bit heavier, and thus mechanically more difficult to drive.
I'm not an ME, but I suspect that the mechanical issues for making a linear actuator with that sort of resolution are probably a harder problem than the already-pretty-well-solved issues for a pivoted-arm system.
The pivoted arms and their servo electronics already have a resolution as good as is needed... track spacing is extremely tight, and I believe that the lower limit is currently set by the size of the heads.
To gain 1000:1 speedup you would need to not only eliminate arm motion (one-head-per-track) but also rotational latency - and this would require multiple heads per track. This would require many millions of heads per drive.
Years ago, there were "one head per track" fixed disks in use... Xerox used one called a "RAD" (random-access device) on its Sigma 7 main-frames. Using the RAD for memory swapping/paging was a good deal faster than using one of the Winchester hard drives.
The RAD was somewhat trouble-prone... it would occasionally suffer from head crashes (perhaps as a result of some dust contamination getting inside the bulkhead). When this would happen, you could *see* a circular line scratched onto the disk media!
The cure for this was to open up the electronics compartment, unsolder the crashed head from its read/write electronics, and solder the wires to a spare head, and then re-format the device.
The RAD had a capacity of only a few megabytes. I really don't think that the head-per-track approach would work with today's track densities.
Because the numbers really don't work out the way you think they would, with today's extremely high track densities.
If it were that easy, all of the drive companies would have done it already. They're bright people, in a *very* competitive business (to the point of being cut-throat).
--
Dave Platt AE6EO
Friends of Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.