why no 5.25" hard drives?

[Dungeness]

Why are internal hard drives for desktop computers only manufactured in 3.5 inch format? Assuming all else is the same (same data density, number of platters, etc), a 5.25" hard drive would have 2.25x the surface area of a 3.5" hard drive. Therefore 2.25x the capacity.

Proof:
pi * (radius^2) = area of a circle
3.14 * (3.5 inches ^ 2) = 38.5 square inches
3.14 * (5.25 inches ^ 2) = 86.6 square inches

Other size hard drives are certainly possible. 2.5" and 1" drives are already being made for ipod, memory cards, etc.

Computer cases already have 5.25" bays for optical drives. Why not just use 5.25" bays for all drives? Should simplify designs and lower manufacturing costs.

Legacy formats can always be overcome. Witness the replacement of the 5.25" floppy by the 3.5" floppy, which will soon be dead itself.

Just a pondering another mystery of life.

 

[cy7878]

costs more for the raw material ie metal for the casing, the board will have to be bigger. Electrinics are going towards miniturization to save cost, look at the Mac mini.

 

[WackyDan]

..and pc's are getting ever smaller... HOWEVER.

Quantum (forgot who bought them or who else they became) used to make a 5 1/4 HD called the "Bigfoot" for the very reason you specify- Higher density.

But rotational speeds were slower then. What I may suggest you do is figure out the outside edge rotational speed (NOT THE RPM) of a current 3.5 inch drive at 7200, 10,000, or even 15k rpm.... now do that for a 5 1/4. You will find that the rotational speed on the outside edge of the platter in a 5 1/4 would be insane. When the first 15krpm SCSI drives came out they were thicker as they had to use smaller platters for this very reason. Things tend to break/disassemble at those speeds, and the forces within, air turbulance, and friction would be a huge issue.

 

[SNM]

Besides outside edge speeds, the larger drives need to be stronger simply to avoid "drooping" platters and maintain alignments. Besides which, 3.5" bays work just fine, so why bother changing? It's not like we can't fit all the storage we want on 3.5" drives.

 

[Mark R]

The problem with making drives bigger is speed. Quantum (now Maxtor) launched their 'bigfoot' drive as a budget model for bulk storage.

Despite good transfer rates, it's acutal real world performance was disastrously bad. The rotational speed was lower than 3.5" drives (either 3600 or 4500 rpm I think), and because the seek movements were larger, seek times were longer - something like 15-20 ms.

High end SCSI drives are actually considerably smaller physically than lower-end drives, they just retain the 3.5" form factor for compatability. Smaller platters are better able to withstand the radial forces of high speed rotation, and smaller seek movements give improved random access performance. Seagate is now beginning to migrate their high-end drives into a 2.5" form factor, as the 3.5" form factor just wastes space - and space is becoming a premium in high-density servers (e.g. blades and 1Us).

 

[corkyg]

All of the above - plus 5.25-in HDDs would cost more to manufacture. This is a very tight margin business. Back in the "old days" there were 5.25 drives. (See Tiap's note below!) The market place wants smaller - not bigger.

 

[Zepper]

Besides being "low and slow", the Quantum (now part of Maxtor) "Bigfoot" had a notoriously poor reliability record too.

.bh.

 

[jamesbond007]

Originally posted by: Mark R
...and because the seek movements were larger, seek times were longer - something like 15-20 ms.

The latest Seagate 7200.8 Baracudas fall into this range as well, according to some benchmarks!

http://www20.tomshardware.com/storage/20050317/seagate-05.html

 

[tiap]

Perhaps this was before your time but back in the early to mid 80's all pc hardrives were 5.25, specifically on the original ibm 8086 and 8088 pc. It took a few years before the first 3.5 drives. I had a 10 mb mfm drive that was also known as double height. It literally took the height of 2-5.25 bays stacked. It was state of the art then and very very expensive.
But they are just gone now with the real floppy drives (5.25) and full length slot cards

 

[Zepper]

Actually, real floppy drives are 8". 5.25" were called minifloppy drives when they first came out... I know, I was there.

.bh.

:moon:

 

[Intelman07]

As said above, the rotational speed would be too fast. Intresting fact, when making a 15000 RPM hard drive, engineers ran into problems where they had sonic booms on their disks, so in turn they had to make the platter smaller

 

[mechBgon]

High end SCSI drives are actually considerably smaller physically than lower-end drives, they just retain the 3.5" form factor for compatability. Smaller platters are better able to withstand the radial forces of high speed rotation, and smaller seek movements give improved random access performance. Seagate is now beginning to migrate their high-end drives into a 2.5" form factor, as the 3.5" form factor just wastes space - and space is becoming a premium in high-density servers (e.g. blades and 1Us).

Yeah, check out the internal photo of a Cheetah 15k.4 versus a Barracuda 7200.8.

Interesting fact (at least to me): if you get a detailed .PDF from Seagate, the actual capacity of a 36.7GB Cheetah is over 48GB. I'm pretty sure all that extra is spare capacity that the drive keeps in reserve for when it runs its self-diagnostics in its idle time and finds bad areas in use.

edit: oh, and my first PC was a refurbished CTX with a Cyrix PR166 and it did have a Quantum Bigfoot in it. We're talkin' slow everything!

 

[piasabird]

The wider the disks are the more the platter tends to wobble. This makes it harder to read the disk so it would have to spin slower to read it. It is just not mechanically feasable.

 

[Pariah]

The history of the computer industry is to continually make everything smaller rather than larger. Smaller means less space (duh), less heat and less power, all desirable traits.

Why would we want to move back to giant 5 1/4" drives? The only positive a larger platter drive would yield would be higher capacity. Hitachi just released a 3.5" 500GB drive. Is that really limiting to the majority of the public? Everything else about a 5.25" hard drive would be a step backwards both environmentally and performance wise. Not sure how you see a larger drive being cheaper to produce than a smaller one. 3.5" hard drives are a very mature field, and are certainly cheaper to produce at this point than a 5.25".

The reason SCSI 15k platters are smaller than 10k platters which are smaller than 7200RPM is purely because of power issues, not because they couldn't make them the same size. OEM customers deliver desired specs to HD makers and they try to meet them to the best of their abilities. When a company is trying to replace 100's if not 1000's of HD's, they don't want to have to replace the supporting infrastructure as well which could be very time consuming and expensive. For this reason, they always want new generations of drives to adhere to the same power usages as the previous generation at worst.

These power limitations also limit how fast a read/write head can move. HD makers could make drives seek a lot faster if they wanted to. The problem is the only way to make the head move faster is to supply it with more power which they can't do to remain within desired specs.

 

[Pariah]

You're in luck, I found your answer from an engineer that works for Maxtor over on SR's board:

"Larger platters were abandoned for two reasons. First, the power required to spin them increases with the square of diameter (I think, though it might be the cube). You can't achieve high RPM and stay within the allowable power budgets with large diameter media. Second, larger platters tend to "flap" or "potato chip" under the stresses induced during high RPM rotation. This deformation causes problems with data integrity ("stress erasure") and makes it harder for the heads to fly and the servos to track follow (forcing lower TPI and so lower recording efficiencies).

Larger diameter hubs do nothing to "improve" the RPM performance of an HDD. Bigger bearings are *less* precise than smaller bearings. And a larger inner diameter means less of the overall radius of the media is available for data tracks, meaning lower capacity for a given platter diameter.

The number of platters is a function of the height of the HDD, the thickness of the media substrate, and the manufacturing tolerances of the design. You can fit more platters into an HDD by spacing them more closely together. But to do that, you need to make them thinner, which makes them less resistant to the flapping issue, which requires them to be smaller diameter.

Manufacturers have tried multi-actuator drives (one spindle but two VCMs). The added cost doesn't justify the performance benefit, and this type of drive is less reliable than a single-actuator type.

The issue with higher RPMs has nothing to do with mechanical strength. 22K RPM (the next "proposed" operating point after 15K, though don't ask me why) designs have already been tried (years ago). Unfortunately, power consumption goes up radically with RPM, 22K RPM drives won't fit the power budgets required by the OEMs (who want to be sure new generation HDDs will work in the systems they shipped last year). There are other complications associated with getting a servo to function at high RPMs that degrade recording efficiency. And while a 22K RPM drive has 2/3rds the rotational latency of a 15K drive, the seek times wouldn't be much different and the higher RPM wouldn't give you a substantial improvement in performance."

So, there you have it.

 

[Promethply]

Wonder why Quantum came up with the idea of producing a 5.25 inch "bigfoot" in the mid 90s.

 

[hopejr]

my 286 had a 5.25" hdd (capacity was 30MB)

 

[Promethply]

Interesting, almost sounds like as if Quantum was trying to sell older design/technology :Q but obviously it didn't work since the bigfoot performance was inferior compared to its peers.

 

[jackschmittusa]

The Bigfoot drives weren't too bad as a cheap file storage device. They got most of their bad performance rap when OEMs started using them for primary storage during a time when ram was expensive, and there was a lot of file swapping going on since there wasn't much ram in the systems. Shortage of ram + slow primary drive = awful performance. Consider as well that most of these were budget machines with slow processors too...well, you can see it went downhill fast.

 

[Jeff7]

Originally posted by: mechBgon

High end SCSI drives are actually considerably smaller physically than lower-end drives, they just retain the 3.5" form factor for compatability. Smaller platters are better able to withstand the radial forces of high speed rotation, and smaller seek movements give improved random access performance. Seagate is now beginning to migrate their high-end drives into a 2.5" form factor, as the 3.5" form factor just wastes space - and space is becoming a premium in high-density servers (e.g. blades and 1Us).

Yeah, check out the internal photo of a Cheetah 15k.4 versus a Barracuda 7200.8.

Interesting fact (at least to me): if you get a detailed .PDF from Seagate, the actual capacity of a 36.7GB Cheetah is over 48GB. I'm pretty sure all that extra is spare capacity that the drive keeps in reserve for when it runs its self-diagnostics in its idle time and finds bad areas in use.

edit: oh, and my first PC was a refurbished CTX with a Cyrix PR166 and it did have a Quantum Bigfoot in it. We're talkin' slow everything!

So your amazing, expensive SCSI drive could actually be losing data to bad sectors like crazy - gigabytes of bad space, and you wouldn't even see it. (Yes, IDE drives do this, but with much smaller reserve pools.)
Wow, no wonder SCSI drives can be more reliable. It's got almost another half a drive in there as reserve space.

 

[Pariah]

No, that's not really a function of reliability. If the drive actually had to remap that many sectors, the performance would be dreadful due to all the extra seeking, especially during activities that required streaming data.

Higher areal data density generally means better performance. Since SCSI uses industry standard capacities (18GB,36,72,144,300), the manufacturers squeeze as much capacity onto the platters as then can then just use what is actually needed to meet the necessary capacity. If they can short stroke a portion of the drive off, that's bonus performance for the drive. Though this effect isn't evident with the Seagate 15k.4 which has pretty awful performance compared to the competition.

 

[Oyeve]

Originally posted by: tiap
Perhaps this was before your time but back in the early to mid 80's all pc hardrives were 5.25, specifically on the original ibm 8086 and 8088 pc. It took a few years before the first 3.5 drives. I had a 10 mb mfm drive that was also known as double height. It literally took the height of 2-5.25 bays stacked. It was state of the art then and very very expensive.
But they are just gone now with the real floppy drives (5.25) and full length slot cards

Yeah, the old ESDI drives. Bout the size of a carborator.

 

[mechBgon]

Originally posted by: Jeff7

Originally posted by: mechBgon

High end SCSI drives are actually considerably smaller physically than lower-end drives, they just retain the 3.5" form factor for compatability. Smaller platters are better able to withstand the radial forces of high speed rotation, and smaller seek movements give improved random access performance. Seagate is now beginning to migrate their high-end drives into a 2.5" form factor, as the 3.5" form factor just wastes space - and space is becoming a premium in high-density servers (e.g. blades and 1Us).

Yeah, check out the internal photo of a Cheetah 15k.4 versus a Barracuda 7200.8.

Interesting fact (at least to me): if you get a detailed .PDF from Seagate, the actual capacity of a 36.7GB Cheetah is over 48GB. I'm pretty sure all that extra is spare capacity that the drive keeps in reserve for when it runs its self-diagnostics in its idle time and finds bad areas in use.

edit: oh, and my first PC was a refurbished CTX with a Cyrix PR166 and it did have a Quantum Bigfoot in it. We're talkin' slow everything!

So your amazing, expensive SCSI drive could actually be losing data to bad sectors like crazy - gigabytes of bad space, and you wouldn't even see it. (Yes, IDE drives do this, but with much smaller reserve pools.)
Wow, no wonder SCSI drives can be more reliable. It's got almost another half a drive in there as reserve space.

I'd have to look at the list of grown defects, yeah.

 

[grant2]

Originally posted by: Promethply
Wonder why Quantum came up with the idea of producing a 5.25 inch "bigfoot" in the mid 90s.

they were cheap, it's that simple.