1. Home
  2. Electronics Design
  3. HD speed up times
  4. Page 2

HD speed up times

I do not believe this is true. HRD's have millions of heads in an array configuration using standard lithography techniques.

Yes, but sequential access could be immensely improved. Say you had 32 heads on an arm and you were writing and reading sequentially. This would improve the speed by a factor of 32 if not more(since no motion of the head would be needed as you would write in 32-bit blocks).

possibly but again, the distance needed to travel is divided by the number of heads.

Maybe but again, I doubt it. It might be a little work but well understood in other industries. Obviously it's not going to be a walk in the park but all progress takes work.

Probably.

see

formatting link

This is exactly what they are doing.

I'm not talking about random access but sequential. Obviously no matter what, with random access you'll have to wait on the platters to spin around. But with sequential access. i.e., reading large defragmented files, one can get a huge increase. Most files can be defragmented to get this increase so the average increase will be quite high for most things.

Yeah, I thought about that too. Having so many heads and in the configuration one couldn't deal with it in a safe way easily.

Again, see

formatting link

I'm basically thinking of taking a strip of their heads wafer and using it as an arm.

I doubt it. I think it's more a matter of determination. There is simply no desire to do it or it would be done.

Then why isn't this out?

formatting link

The technology is obvious there. It makes sense. The company has done the majority of the work. Why isn't seagate or some other HD company out there buying it up? (The big step here is the formulation of the heads array)

HD companies may be competitive but they also are not going to waste all the investment they have. Also SSD's are really the wave of the future from what I hear(although their long term storage capacity is crap).

The proof that seagate doesn't care is when they brought

formatting link

and dropped it because it wasn't feasible. What it really means is that it wasn't worth it. Too much money to develop and not enough buyers.

In any case the real problem with the fixed head array on platters(and possible the HDR) seems to be the temperature issues. As the platters heat up and expand it throws of the alignment of the fixed heads which at the track densities is quite significant. With one head on a swinging arm one can simply adjust the position algorithm to compensate or use an alignment track.

Anyways, A+ for trying to have an intelligent discussion and not being so egotistical 2-bit engineer.

Reply to
Stretto
Loading thread data ...

To the point where most of the players are copying the main design features of the main two makers.

Reply to
Chieftain of the Carpet Crawle

Why don't YOU grow the f*ck up BEFORE you make another post?

Don't worry, dumbfuck. We know better than to think that you would leave and not return until you get a clue.

Reply to
My Name Is Tzu How Do You Do

message

up

This

very

leak?

Since the poster was both obnoxious and stupid, you were there in spirit.

John

Reply to
John Larkin

Adding info:

We have multiple heads per arm now, since a platter has two sides to be used, and they can push up to four platters into a 3.5" drive, it's the advance of areal density that's getting the data density improvements on the current physical limit in number of platters. The slim drives like current 500GB use one platter, get that down to 500GB per side and you can make 4TB drives, which not quite out yet? Unless I missed it, current max size is a 3TB drive.

My first HDD was a 100MB unit, with a separate optical encoder arm inside for the track servo, these days that servo info is embedded with the data, and, a drive self calibrates on first usage.

Recently I built a 5 x 1TB Seagate drive array, the five drives are from the same factory, serial number sequence, yet they have different data rates over the surface, indicating self calibration at runtime to get best out of the drive, without lengthy factory test. Quite amazing, and observable by anybody (on Linux anyway).

Grant.

Reply to
Grant

Please, seek time and latency are two separate issues, seek to get the heads to the right track, latency of average half revolution period to find the right data on that track.

I don't think you or the OP has opened a hard drive, there one head per surface (cheap Maxtor have one head, period), where does the extra head go? Look at the form factor and try fitting another arm in there. No room. Cannot be done unless you shrink the disk to make room. If you duplicated the arm and electronics you can reduce the drive latency, but as another poster pointed out, it's far easier to add an array of drives to both reduce latency and increase drive data capacity.

Not gonna happen that way, SS drives will be more practical and cheaper, lots of precision electronics to run each head, they now use thermal effect to control the head's flying height, you imagine several thousand of the supports for that?

Yup, if it was easy, they'd do it now.

Grant.

Reply to
Grant

You are even more retarded than Roy is.

Reply to
StickThatInYourPipeAndSmokeIt

I'd consider a 2-sided platter to have two arms... each arm, currently, with one head. One actuator. I think the OP had in mind multiple heads per side of each platter.

and they can push up to four platters into a 3.5" drive, it's the

My first HDD stored 128 kbytes.

John

Reply to
John Larkin

Well, I'd personally call that N separate arms and N heads, ganged together into a single actuator mechanism... there's only one head per surface.

The head assemblies themselves have become quite impressively sophisticated. I'm told that these days, there's actually a small thermal-flex "heater" in the head assembly, which heats up and cools down and creates a bending force which flexes the head itself towards or away from the platter surface. The "flying height" of the head is thus being adjusted dynamically, in real time, in order to keep it at the correct height (and I think the optimal height during writing is different than the height during reading).

That's one of the limitations on trying to run more than one head per surface. It's just the head itself - it's the cost and the power requirement for the electronics which drive the head (vertical and horizontal tracking, read conditioning, write-current drive, etc.).

Real Soon Now, I expect.

Yup. The sectors-per-track will vary, between "identical drives", and between heads/surfaces and across individual surfaces in any specific drive, depending on the quality of the individual head and the quality of the surface.

The days are long gone when the host controller and software had any really good understanding of the drive geometry. These days, you can either use the "logical block address" (as SCSI drives did from the get-go), or you can try to use some CHS (cylinder/head/sector) legacy addressing notation which actually bears no meaningful relationship to the drive's internal storage geometry. The drive controller is managing it all "out of sight" of the host I/O system.

--
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!
Reply to
Dave Platt

Yup, they've been set to revolutionize the industry since 2009.

Nice white papers.

formatting link

John

Reply to
John Larkin

That was actually my point. Adding multiple heads per arm (if that were mechanically possible) might shorten the time to seek the arm/head to the necessary track, but would do nothing to reduce the average rotational latency. Since the rotational latency is on the same order as the average seek time, even cutting the seek time to zero wouldn't result in the extreme access-time reduction that the OP was thinking.

Dunno about the OP, but in my case you'd be mistaken... looking at the innards of modern hard drives is actually part of my day-job.

Agreed. The OP's idea would require major redesign of how arms and head assemblies are constructed (if it were possible and practical, and I don't think it is).

I agree.

I'm going to be very interested to see if memristors turn out to be practical for that sort of usage.

Yup... completely impractical. I believe that the OP did/does not understand the complexity of the "support electronics" required for each head assembly.

No argument there.

--
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!
Reply to
Dave Platt

"The drive controller" on an IDE drive is ON THE DRIVE.

The "host I/O system" is ONLY an I/O system, NOT the drive controller. On IDE.

Reply to
Chieftain of the Carpet Crawle

message=20

them up=20

the=20

20

to=20

This

and very

the=20

leak?

The headers show that it is somebody else. Besides the style is more = like GX.

Reply to
josephkk

message=20

=20

them=20

minimal(just the

20fold(say 20

has to

one.=20

and=20

of the

more).

60+
1970s.

that

more=20

=20

on=20

=20

current=20

inside=20

data,=20

The smallest HD i have heard of were always in MB and larger. Are you thinking of a floppy?

Reply to
josephkk

What do you mean by small, dumbass? Even the "smallest" capacity drive was a LARGE full height form factor device. The half heights didn't come out until the second generation of MFMs.

Even the first 1GB drive was a 5.25", full height SCSI Seagate.

I'll bet your skull cavity is the small form factor. You know... THICK! Outward appearances can be deceiving. "Objects in skull cavity are smaller than you might think from looking." Bwuahahahahahahaa!

Reply to
TheQuickBrownFox

I saw the stretto post, _before_ I also kill-filed him. NymNoNuts is _universally_ kill-filed :-) ...Jim Thompson

--
| 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
Reply to
Jim Thompson

According To Wiki 3.75MB.

AlwayWrong again. What a wealth of disinformation you are.

As usual, nothing at all useful in QuickBrownLoaf's post.

Reply to
JW

Some of the more common DEC 14" platter drives were the RK05 with 2.5 MB capacity and the RL01 with 5 MB capacity.

Reply to
upsidedown

Around 1983 when cp/m machines were still useful and as fast as IBM-PC, a local company that made cp/m semi-portable was in liquidation, one of the items on sale was a hard drive, all of 5MB capacity, for several thousand dollars. At the time I could fit a bit over 1MB onto a quad density 8" floppy disk, couldn't see any point to such a small hard drive capacity.

In October'84 got a shiny new PC-AT to work with, 21MB hard drive as standard, on a '286 processor. They were the ones included dirt from the factory next door. Failed fairly often, but was years before we got to know the reason. Hard drive was a little more convenient then big floppies, but the cp/m box was kept in use for another year, connected via serial cable to the IBM box for backups and files transfers.

There were a few products that suffered lack of clean room build environments back then, anyone remember the Beckman displays? They developed spots of burnt, corroded areas inside, also due to the factory next door's particulates. Took them a while to work it out, Beckman displays were large neon seven segment digits, before LEDs took over, used in industrial instrumentation displays, PID controllers.

Grant.

Reply to
Grant

Oh boy. JW asswipe knows how to google search! WOW!

You're an idiot. We are not talking about mainframe horseshit here, boy. We are talking about commercial products. The Seagate was the first.

As usual, you have posted NOTHING contributory.

If you dispute something I wrote, asswipe, then CITE the contention, don't just spew horseshit, like the little bitch we all know you are.

Reply to
TheQuickBrownFox

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.