Talk:Hard disk drive

From Wikipedia, the free encyclopedia

This is the talk page for discussing improvements to the Hard disk drive article.
Article policies
Archives: 1, 2, 3
Hard disk drive is a former featured article candidate. Please view the links under Article milestones below to see why the nomination failed. For older candidates, please check the archive.
September 9, 2007 Featured article candidate Not promoted
This article is within the scope of Computing WikiProject, an attempt to build a comprehensive and detailed guide to computers and computing. If you would like to participate, you can edit the article attached to this page, or visit the project page, where you can join the project and/or contribute to the discussion.
B This article has been rated as B-Class on the quality scale
Top This article has been rated as top-importance on the importance scale
Archive
Archives


Contents

[edit] perpendicular recording -- when?

"Another technology used to overcome thermal effects to allow greater recording densities is perpendicular recording, which has been used in many hard drives as of 2007 [7][8][9]."

Then a few paragraphs later:

"As of 2006, some disk drives use perpendicular recording technology to increase recording density and throughput.[15]"

So is it 2006 or 2007? This self-contradiction should be corrected. T-bonham (talk) 12:28, 3 April 2008 (UTC)

No year is definitive. Like a lot of high-tech products, drives are developed in stages -- demonstration units, prototypes, pre-production models, and production models. Often the technology itself also goes in stages, where the early versions are feeble or even impractical. Any reference we could use for the exact date of any specific technology usually ends up with just conflicting marketing claims. In this case, the statements "some" drives in 2006 and "many" drives in 2007 is about the best we can do. --A D Monroe III (talk) 15:21, 18 April 2008 (UTC)
There is no inconsistency between the two sentences but the second one appears to have disappeared. Perpendicular recording was first successfully shipped to market in 2005 by Toshiba[1]. Others followed in 2006 and 2007. There were many earlier attempts to commercialize perpendicular recording going as far back as IBM in the late 1950's but nothing really happened until Toshiba circa mid 2005. Tom94022 (talk) 16:06, 18 April 2008 (UTC)

[edit] please unprotect page

why is this page protected? anyway, i recommend the following columns to this table in the middle of the article:

Standard Name Width Largest capacity to date (2007) Platters (Max)
5.25" 5.75"
3.5" 4" 1TB 5
2.5" 2.75" 160GB 2
1.8" (PCMCIA) 54 millimeters
1.8" (ATA-7 LIF) 2.12"
Check the History tab on the article. There was a persistent vandal recently, thus the article was protected. I do think information on platters and capacity belongs in the article, but I'm not sure whether the "Physical Dimensions" table is right for it. Whether this means the table should be renamed, or another table is needed, or some other solution is best, is something I'll need to think about, or defer to wiser minds. Also worth bearing mind is that this information is not "static", but describes something of an ongoing development. Not sure what the implications of that are, to be honest. It's not like changes come every day, but it's definitely info that's subject to change. Paul.w.bennett 16:41, 2 April 2007 (UTC)

[edit] Should this article be titled "Hard disk drive"

A hard disk is a component of a hard disk drive, and such disks are not in today's embodiments user removable.
A floppy disk is the medium of a floppy disk drive and such disks are generally user removable
An optical disk cartridge is the medium of an optical disk drive and such cartridges are generally user removable.
A tape cartridge is the medium of a tape drive and such cartridges are generally user removable.


So why is this article entitled hard disk? More important, who is going to revert me if I go thru the effort to make this terminology consistent. Tom94022 23:49, 6 January 2007 (UTC)

I like the idea of changing the article to "hard disk drive". Unless you want to change it to "hard file" ;-) I would say the convention for optical media is to use the term "disc". Unless you work for a certain HDD vendor, to be unnamed, then it also applies to magnetic media. :-D GMW 07:44, 9 January 2007 (UTC)
if you want to make it consistent, please also rewrite all instances of "hard drive". _All_ drives are hard; it's the medium inside that sometimes isn't. Hint: A "floppy-disk drive" is a drive for floppy disks. A "floppy disk-drive" is something you send in for repair. Geira (talk) 22:11, 24 May 2008 (UTC)

[edit] DoJ

"2005: Seagate and Maxtor announce their intent to merge. US DoJ approval was given for Maxtor to be acquired by Seagate for US$1.9 billion,[7] and the merger closed in mid-2006."

I was just wondering, could anyone explain the relevence of the DoJ being involved in the merge between two hard drive manufacurers?

Binarysaint 12:25, 17 January 2007 (UTC)

DoJ reviews large mergers for possible violations of the antitrust laws. See http://www.usdoj.gov/atr/public/guidelines/hmg.htm --agr 14:00, 17 January 2007 (UTC)

[edit] temperture for harddrive

what is a safe temp for a harddrive? is 103degrees frienheit to hot --Falcon866 15:15, 17 January 2007 (UTC)

Check out the specifications for your specific HDD. The spec always lists the allowable environmental operating conditions of the drive, including temperature. Most HDDs go up to 55 or 60 degrees Celsius, which corresponds to 131 to 140 degrees Fahrenheit. 103 deg F is typical. GMW 16:20, 26 February 2007 (UTC)

[edit] Is This Possible?

I know someone who'se hard drive got dropped, and he says that the head went through the disk. But I thought that there wasnt much space inside the casing, so how could the head actually have bounced around enough to gain enough speed to actually puncture right through the disk? Was he exxagerating, or should this type of incident be listed under a section titled "Disadvantages of Hard Drives"? Thanks. Ilikefood 16:56, 24 January 2007 (UTC)

I'd like to see a picture; IMHO, it is not possible, for many reasons not just speed. It is possible for shock to cause the head to "ding" a disk leading to a "bad" spot which may or may not be recoverable. Many mobile drives have drop detecting (accelerometers) circuits that move the head to a safe position before the impact preventing such damage. Bottom line is the shock spec on drives exceeds that of many other components, like LCD's, cases, etc., so that i would NOT list this as a disadvantage. Tom94022 17:25, 24 January 2007 (UTC)
Some drives use plated glass platters. I suppose one might find a platter that had the plating completely removed from some part of the platter surface, and one might naïvely describe that as "going through the disk". But other than that, I can't see it either.
Atlant 17:34, 24 January 2007 (UTC)
It is impossible that "dropping" a hard drive would generate enough acceleration for the head to penetrate the platter, unless maybe we are talking about a fall of hundreds of feet. The platters are extremely rigid and very strong (either aluminum or a ceramic). It is possible for a head crash to scratch the media which is sputtered onto the platter, but no way the head "went through" the disc. - O^O 18:27, 24 January 2007 (UTC)

Maybe someone can add to "capacity" section information about 2.5" hard drives (I guess 120Gb is the largest available now, but not sure), please ?

[edit] "Exponential"

It's unusual, and encouraging, to see someone use the word "exponential" correctly: usually it just means something like "fast," but the user has never heard of an exponential function. The graph makes it clear that someone has a clue!

Donreed

[edit] Write precompensation

This section may be unsuitable, but please do not simply remove - possibly move to its own article. - Zephyris Talk 17:29, 17 February 2007 (UTC)

Is this a religious issue with you or are speaking with some knowledge of disk drive technology? The article is too long by most standards and write precompensation is minor one of many technical details of disk drives not covered in the article. IMHO, it doesn't even rise to the level of inclusion in Wikipedia at all, but if u want to move it to a separate page please do so. But you might then consider covering all other hard disk drive technologies. BTW, the PC's CMOS setting is not used by any modern HDD. I'll hear your response, before I delete the section Tom94022 20:15, 17 February 2007 (UTC)
I must say I have some very strong religious views on hard disk drives :) I do understand that it may be an irrelevant detail for the purposes of this article - its both an obscure and superseded technology - but it should merit inclusion as most things do; there probably is a page on hard disk drive religions! Ill move it to its own article.
Wow, my first edit debate! Feels refreshing - I should move out of the obscure world of biochemistry articles... - Zephyris Talk 10:56, 18 February 2007 (UTC)
Precomp is most certainly used on modern HDDs. As Tom94022 said, though, the CMOS setting is no longer used.
Referring to the precomp entry, modern precomp does not involve increased write current. The timing of transitions on short magnets is shifted in order to counteract non-linear transition shift (NLTS). In longitudinal recording, this is largely due to the demagnetization field from previously written data (initial magnetization). GMW 16:11, 26 February 2007 (UTC)

[edit] Google failure study

Google has a paper detailing error rates from a farm of 100,000 drives. Failure Trends in a Large Disk Drive Population

Would it be a suitable link? I found this comment interesting:

Our analysis identifies several parameters from the drive’s self monitoring facility (SMART) that correlate highly with failures. Despite this high correlation, we conclude that models based on SMART parameters alone are unlikely to be useful for predicting individual drive failures. Surprisingly, we found that temperature and activity levels were much less correlated with drive failures than previously reported.

How should I incorporate it into the article? DanBeale 15:08, 18 February 2007 (UTC)

Dan this is a great find and should be incorporated into the article. I would suggest the section Integrity be renamed Reliability and you make a major rewrite to summarize this article Tom94022 20:05, 18 February 2007 (UTC)

[edit] Data Rate

"SCSI originally had just one speed, 5 MHz (for a maximum data rate of 5 megabytes per second)" In the article... I am no expert, But wondering if that would be 5 megabits per second since 5 mhz would be 5 million cycles per second, meaning a maximum of 5 million bits per second, making maximum of 5 megabits per second. Maybe I am wrong, just thought I would put my 2 cents.

My guess is that this talk is unrelated to the Google article above, so I made it a separate talk session. The SCSI-1 bus was 8 bits wide so each transition transferred one byte. I think the transaction rate was more complex than just a maximum of 5 MHz but if that was the maximum then the rate is 5,000,000 bytes/sec. Tom94022 20:05, 18 February 2007 (UTC)

[edit] Sound

What causes the "creaking" sound you often hear when a hard drive is reading/writing? Is it the sound of the heads moving?

Most probably. On a Quantum Bigfoot (2GB model) you can actually watch the head moving (without the drive dying) under the transparent foil. —Preceding unsigned comment added by 145.253.2.232 (talk) 15:27, 23 September 2007 (UTC)
Sort of. On a modern drive, what you actually hear is the position signal changing as the heads approach a track, by way of the voice coil motor. The arm itself makes almost no noise unless the drive is damaged or the bearings holding up the arm are worn out (very rare; usually the spindle goes first). -lee (talk) 19:19, 25 January 2008 (UTC)

[edit] Change to capacity measurements section

I made some changes to the "Capacity measurements" section expanding on the difference between base 10 and base 2 measurement systems. Tom94022 undid my changes two hours later claiming my additions were POV, historically misleading and factually incorrect. Could Tom94022 or somebody else give a clearer reason why my changes deserved to be removed? I think the previous (and current) section is POV, historically misleading and factually incorrect. That is why I changed it. Tom94022 simply undid my changes instead of trying to improve upon them.The Goat 21:02, 20 March 2007 (UTC)

I don't speak for Tom, however, I have to ask . . . What makes you think that the base10 choice had anything to do with manufacturers advertising larger disk drives? Do you have a reference? ~a (user • talk • contribs) 21:36, 20 March 2007 (UTC)
No, there aren't any references, since the "hard drive manufacturer size inflation conspiracy" is bogus. Hard drives have been measured in decimal units since they were invented. See Talk:Binary_prefix#History for a closely related discussion. — Omegatron 21:57, 20 March 2007 (UTC)
Sorry I didn't make my self clearer. The "conspiracy" statement was IMHO POV. I can point to at least 3 early important computers that did NOT have main storage sized in binary related amounts so there is no reason to ever use k or K as other than 1,000 in these early computers. The Stretch documents from the late 1950's do NOT use k or K in a binary sense. The PDP8 did come with main storage measured in 4,096 words of 12 bits, but the 1966 Users manual refers to this in decimal numbers and never uses k or K to refer to main storage. Interestingly the manual does use K as 1,000 when, e.g. referring to a drum as 65K words. The IBM S/360 mainframe product literature did NOT use k K or M in a binary sense for main storage. Amdahl did use K clearly marked as =1024 in his seminal mid-1960's article on the s/360 but else where in the same article he used K as 1,000 without marking - no confusion. The 8" Floppy disk of the 1970's was accurately characterized in decimal units and prefixes by IBM and most OEM's. DEC may have misapplied the SI prefixes but I can't be sure. So to somehow say this occurred in the 50's and 60's is IMHO factually and historically incorrect as demonstrated by these several citations. Personally, I don't think it became a public problem until much later, perhaps as late as the Apple Macintosh. MSDOS is pretty clean, for the most part spelling out storage capacity, main or disk, in decimal units without prefixes. I couldn't come up with anything better than what was already there so I reverted it. If I had to say something, it could only be that the beginnings of the public misuse of the SI units is not clear but it appears to be after the Intel chip, and after floppy disks.Tom94022 22:55, 20 March 2007 (UTC)
Reading these responses and the thread in Talk:Binary_prefix#History I see two groups arguing about two related but slightly separate things. One groups says, "HDD's have always been base 10." The other says, "computer bytes prefixes are historically powers of two." These are not diametrically opposed facts. HDD can be manufactured to base 10 capacities and byte measurment prefixes and be powers of two at the same time. The sticky point is when HDD manufactures stopped advertisings 5 million bytes and started using 5 megabytes instead. Like it or not SI does not have magical power over the electronics/computer industry. They don't get to define Greek and Latin word prefixes for units of measure not included is SI. It is confusing but the computer industry adopted the same prefixes with different definitions. This was not a mistake. This was not a misuse of SI units. I admit without a federal supeana there is no evidence to support a conspiracy argument behind the current choice in HDD advertising lingo.The Goat 12:25, 21 March 2007 (UTC)
Like it or not, every publisher has an obligation to define his terms when he uses standard terms in a non standard way. For example, in the U.S. I can publish a temperature as 15 degrees without having to say Celsius, but if I meant Kelvin then I should so annotate. There is no disagreement that kilo, k and K meant 1,000 long before it was used in a binary sense with regard to computer main storage. The disagreement is only about when it became common to use SI prefixes in a binary sense with regard to computer main storage; that is when were [ 1) products specified using binary SI prefixes, or 2) advertised as such] AND 3) the associated operating systems reported binary SI prefixes AND 4) there was no disclaimer, (e.g., K=1024 as in the Amdahl article). The earliest such system I can clearly identify is Apple Macintosh; MSDOS and Apple DOS are relatively clean and to the best of my recollection so was CPM. IBM MVS and DOS were clean. I don't know about UNIX or DEC OS's but I suspect they were clean. So in the absent any facts, postings establishing an early date for the onset of the mis-use of SI prefixes are speculative and inappropriate.Tom94022 15:21, 21 March 2007 (UTC)
Sorry, SI didn't invent the prefixes they are Greek and Latin. SI can only define the prefixes for SI units. The unit the byte is not governed by SI. So there is no such thing as "mis-use of SI prefixes" when they are used with the byte. The standard use of prefixes when used to modify the unit the byte is the base 2 way. The fact that some early articles included a legend (not disclaimer) highlighting K=1024 doesn't sway the argument one way or the other. People often include extra information so that a unit of measurement is clearly understood. Short/long/metric ton is a good example. When writing an article I would always clarify I am referencing gross tons. That doesn't mean my definition of ton is mis-used or wrong in anyway. If a OS (such as MSDOS) reports sizes as thousands, millions or billions of bytes then it has nothing to do with prefixes. It is just reporting a number of bytes. Not a number of kilobytes, megabytes or gigabytes as those units are never used. i.e. output from the dir command shows I have "14,717,014,016 bytes free". Where is the prefix? Any given OS' implementation isn't really important anyway. What is important is the convention used by the everyday members of the particular industry. I don't think anybody can make a argument against the computer industry's widespread convention being kilobyte=1024 bytes. Of coarse it is possible to find many examples that contradict the industry convention. Like a byte use to not always be eight bits and so on and so forth.The Goat 18:56, 21 March 2007 (UTC)
We agree MSDOS did not contribute to the establishment of the binary SI units because it didn't use prefixes. Again the question is not whether there is a "standard" but when did the computer industry generally adopt such a standard. The only evidence, presented so far is that it didn't happen in the 1950's or 1960's as u originally asserted. If you have any evidence, I'd like to see it.Tom94022 21:37, 21 March 2007 (UTC)
So let me submit that it is very much in question whether there is any kind of "standard" that assigns the meaning 1024 to the prefix "k". I, personally, do not know a single person who has ever thought that a megabyte is "exactly one million byte", but I also don't know anybody who even knows the number of bytes in a MB. You, the reader of this text, cannot name how many bytes are in a MB, so why claim that this number is some kind of "standard"? As far as I know, everybody has always used the prefixes "kilo-" or "mega-" in the computer context as "about a thousand (or a million or whatever)". That is, if I go and ask random people how many bytes are in a gigabyte, the most common answer (after "I don't know" and "a lot" and such) is going to be "about a billion bytes". Or maybe "about a thousand megabytes". With emphasis on the "about". A megabyte is that odd number of bytes that has to do with the way computers count stuff that is close to a million. Good enough for anybody I've ever talked to. Claiming that the number 1024*1024 (which you'd need a calculator to figure out because you can't even name it from the top of your head) has ever been some kind of "standard" would require either a standards body adopting it (nope, SI says megabyte=10^6byte) or some large majority of industry insiders using it in this precise way (nope, they say megabyte="about a million byte") or the majority of folks on the street understanding it that way (nope, if they can name a number to the prefix "mega" at all, they'll say "million"). How is something a "standard" if nobody actually thinks this is exactly the meaning of that prefix?Iron Condor (talk) 23:30, 14 February 2008 (UTC)


Instead of wasting time with futile personal battles, we should be finding all the evidence we can and collecting it in a single location; probably Talk:Binary prefix#History. In other words, start making a timeline of actual sizes and the words used at the time to describe them. First example entry: Moved to Talk:Binary prefix#Timeline. — Omegatron 21:49, 21 March 2007 (UTC)

Actually that's what I've been doing here and in the Binary Prefix page; so far I have identified the RAMAC 305, IBM 1620, IBM Stretch papers, the PDP8 and all IBM/360 as not using Binary Prefixes. I will now add to our list the S/370 announcement:Tom94022 23:16, 21 March 2007 (UTC)

[edit] Manufacturing

Could there be a manufacturing section, or another article? (Like CD manufacturing) Yuletide 16:46, 21 March 2007 (UTC)

I am suggesting that the disk spinning rate of 2007 be changed from an average of 5400 rpm to 7200 rpm. LAPTOP HDDs have a 5400 rpm, but 99% of SATA or IDE HDDs on desktop computers use 7200 RPM drives. Either way, enterprises usually use a SCSI interface on their servers, which go @ ~10,000-15,000 rpm.

Just a thought from an computer consultant. 74.132.138.151 01:47, 12 April 2007 (UTC)Admiralthrawn999 (was looking something up, and didn't sign on)

Yes, if none will oppose, I'll do that change (just something from a computer consultant/technician)--Doktor Who 22:44, 12 April 2007 (UTC)

[edit] {{editprotected}} request for edit and change

request for edit and change the 30 GB size in the seagate hard disk image label into the correct 40 GB size; if you look at the big image carefully you can read in the upper right corner 40 instead of 30 ... and... i have got one on my table right now ;-) i have corrected it in the image, but the main article is protected

Appears to have been done. Harryboyles 12:30, 15 April 2007 (UTC)

[edit] Error in Capacity section?

"The first 3.5" HDD marketed as able to store 1TB is the Hitachi Deskstar 7K1000. The drive contains five platters at approximately 200 GB each, providing 935.5GB of usable space." Shouldn't this say 935.5 GiB? It is in fact a 1 trillion byte drive. I'm making the change. --711groove 13:29, 21 May 2007 (UTC)

you have it the wrong way around

its actually a 1 TiB drive, 935.5GB. its correct the way it is -- BBnet3000 23:19, 22 May 2007 (UTC)

What you just wrote makes no sense. 1×240 bytes != 935.5×109 bytes or 935.5×230 bytes. 711grove is correct. -- mattb 23:48, 22 May 2007 (UTC)

[edit] Bias in measurements

As a tangent to the endless "discussion" about the use of standardized computer units on Wikipedia, User:Centrx is using this section to try to push a POV that hard drive manufacturers should have marketed their devices using binary quantities, and that their use of decimal quantities is deliberately misleading. (See User talk:Centrx#Hard drives.) In other words, they're deliberately mislabeling their 80,000,000,000 byte drive as "80 GB" to inflate the actual capacity. (Since, by golly, everyone knows that the operating system reports it as "74.5 GB". ZOMG I've been ripped off!) I've pointed out quite a bit of evidence that hard drive measurements, when abbreviated, have always been abbreviated in a decimal way, that the decimal meaning was the one used by engineers long before they even invented these devices, and that the binary meaning was never standardized, de facto or otherwise. So there is absolutely no reason to expect them to switch to the non-standard binary units, and there is no reason to assume that they are using the standardized units in a deliberately misleading way. The hard drive manufacturers have explicitly stated their position. Without evidence of malicious intent (like a leaked internal memo or the like), there is no more reason to suspect other motivations behind this than there is to suspect that the Titanic was intentionally sank for the insurance money.

In CMOS, FDISK, Windows File Manager and Windows Explorer, my IDE disk reports smaller capacity than the number of bytes reported by CHKDISK. Why?

Because there are two different definitions of a megabyte and a gigabyte. Hard disk drive capacity is calculated by taking the number of Cylinders x Heads x Sectors x 512 bytes per sector. Hard disk drive manufacturers define a megabyte as equaling 1,000,000 bytes and CHKDSK calculates based on this value as well. However, your system's CMOS, FDISK, Windows File Manager and Windows Explorer calculate disk capacity based on a megabyte equaling 1,048,576 bytes. Therefore different utilities might report different capacities for the same drive.

– FAQs - Fujitsu

My computer shows less drive capacity than on the drive box label

Hard Drive Manufacturer Capacity Definitions

The listed capacity is an unformatted (raw) capacity. After partitioning and formatting, actual storage capacities may vary depending on the operating system and configuration. Maxtor adheres to the National Institute of Standards and Technology (physics.NIST.gov) definition of Megabyte and Gigabyte.

Examples:

  • 1 MB = One Million Bytes
  • 1 GB = One Billion Bytes

Operating System and BIOS Capacity Definitions

Storage devices are marketed and sold in terms of decimal (base 10) capacity. In decimal terms, one Gigabyte (GB) is equal to one billion bytes. Most BIOS’s follow this definition as well. However, many operating systems use the binary (base 2) numbering system. That would be two to the thirtieth power, or 1,073,741,824 bytes equals one-Gigabyte.

According to the NIST standard, an 80 GB hard drive would contain eighty billion bytes. 80,000,000,000 bytes divided by 1,000,000,000 bytes equals eighty decimal Gigabytes. In binary terms, 80,000,000,000 bytes would be divided by 1,073,741,824 for a total of 74.5 binary GB. However, there are still 80 billion bytes on the drive in either case.

– Personal Storage 3000DV: Installation Guide - Maxtor

Surely Western Digital cannot be blamed for how software companies use the term “gigabyte”—a binary usage which, according to Plaintiff’s complaint, ignores both the historical meaning of the term and the teachings of the industry standards bodies. In describing its HDD’s, Western Digital uses the term properly. Western Digital cannot be expected to reform the software industry. (Apparently, Plaintiff believes that he could sue an egg company for fraud for labeling a carton of 12 eggs a “dozen,” because some bakers would view a “dozen” as including 13 items.)

– Defendant Western Digital Corporation’S Brief In Support Of Plaintiff’s Motion For Preliminary Approval - Western Digital

Since this inflation hypothesis has led to lawsuits against the drive manufacturers, it should actually be mentioned here in more detail than it is. — Omegatron 16:40, 3 June 2007 (UTC)

[edit] platter sizes and form factors

"The names refer to the width of the disk inserted into the drive rather than the actual width of the entire drive."

This is not entirely correct. For example, so called 3.5" drives typically have 95mm platters—which is closer to 3.7". (3.5" is about 90mm).

I would like to see a small table of the sizes, with specific measurements, in both measurement systems. —überRegenbogen 02:03, 4 July 2007 (UTC)

See Form Factor Comparison Note there are several different diameter disks used in 3.5 inch form factor drives. I believe this is unique to 3.5" Tom94022 02:32, 5 July 2007 (UTC)

[edit] Operations per second

I'm not sure there is any agreement on "operations per second" since can be a strong function of block size but, be that as it may, there is at least one publication that has a modern drive at 561 ops, well above the numbers cited in the article. Accordingly I deleted the statement BTW there are standard tests of IOPS in the unix world but they mainly apply to storage subsystems and not individual drives. If anyone knows of a valid reference and wants to cite such, it would be a useful addition. Tom94022 17:11, 10 July 2007 (UTC)

[edit] FYI: answers.com copies this page..

http://www.answers.com/topic/hard-disk-drive-1

The table on disc interfaces looks "quite similar" .. :-) Oh btw, I made that table asfair. Electron9 07:07, 16 July 2007 (UTC)

Read Answers.com and read Wikipedia:Mirrors_and_forks/Abc#Answers.com. It is normal for Wikipedia to be copied if the license is closely adhered. ~a (user • talk • contribs) 13:57, 16 July 2007 (UTC)
I know about the license. Guess I've seen some bad behaviour from about.com in the past. And this kind of makes it obvious that they don't really have any content of their own. Also it has implications for public presence. Electron9 12:44, 18 July 2007 (UTC)

[edit] Disk Rotation Speeds

Why do manufacturers all use the same rotation speeds? 7200,5400, 10200 etc I can't imagine they all use the same motor manufacturer or do they?

Why isn't one of them busy Marketing there drive as being 7800 RPM instead of just 7400 for those really on the move etc!

Years ago, there were drives with speeds we'd see as odd now. 3200, 3800, 5800, 6300, 14400...eventually the market settled on 4500, 5400, 7200, 10k and 15k for the sake of clarity. The first three are fractional multiples of 3600 RPM, an artifact of now-ancient drives that ran their spindles directly off the AC mains (they used synchronous motors, and mains frequency in the US is 60 Hz, so they spun at 60 revolutions per second -- 3600 RPM). Since 3600 was something of a standard when the first small drives came out in the early 1980s, it stuck, even though the electronic servos they used could theoretically run at any speed. 10k and 15k are nice round numbers (even multiples of 1000). -lee 18:11, 10 September 2007 (UTC)

[edit] Physical interfaces on 1.8" disks

The section "Hard disk drive characteristics" states: "A previous 1.8" HDD standard exists, for 2–5GB sized disks that fit directly into a PC card expansion slot." Shouldn't this ideally describe that two previous standards exist. Apart from PCMCIA and ZIF there is also one commonly used with it's connector on the long side of the drive. Search for HTC426060G9AT00 to see an example of it. 213.114.65.148 13:24, 22 July 2007 (UTC)

[edit] SMD and IPI-2 harddisks missing completely

This wikipedia article is missing several (still used) harddisk types like SMD and IPI-2 (there might be several others but can't remember their names at the moment). Adding these would improve quality of this page quite a lot. Simakuutio 07:15, 9 August 2007 (UTC)

May I suggest the place for such additions, if at all, is the History Of Hard Disk Drives page not this page. SMD already has a page and probabaly should be cited and linked in the 1970's section - u might want to make such an addition. I'm not sure IPI-2 rises to a level sufficient for inclusion in any article; there were many such interfaces that IMO never achieved sufficient technical or market success to justify inclusion in any Wikipedia article. Tom94022 17:35, 9 August 2007 (UTC)

[edit] Number of platters

Updated from 2 to 3. IBM DTCA-23240 2.5 inch HDD (3.2GB (binary)), produced september 1998. had 3 platters (just disassembled). I assume, more recent HDDs can have more than 2 platters as well. -Yyy 16:09, 8 August 2007 (UTC)

[edit] SCSI = Small Computer "Serial" Interface

It's become popular to rewrite history by saying that SCSI stands for Small Computer System Interface, but the second S actually stands for Serial.

One can understand the confusion - after all, SCSI uses parallel data transfer, not serial data transfer. The name comes from the fact that devices are connected in a series along the cable (though not in series electronically - the devices are independent, and won't all stop working if one burns out, like some strings of tree lights). --Thanny 11:11, 16 August 2007 (UTC)

I believe that Alan Shugart disagreed with you; if I remember, I'll check my reference material tonight.
Atlant 12:54, 16 August 2007 (UTC)
I am looking at the cover of the ANSI X3.131-1986 "small computer system interface" specification; in its forward it acknowledges its origin in the "Shugart Associates System Interface" circa 1982. I personally attended many of the early SASI/SCSI meeting and never heard or saw the usage of "serial" in the context of either SASI or SCSI. BTW, Alan Shugart had nothing to do with SASI or SCSI, he left Shugart Associates long before SASI was conceived.Tom94022 15:46, 16 August 2007 (UTC)

[edit] Can we add a quote or some info on how precise hard drives are?

I don't think this article's "technology" section quite captures just how amazing hard drives are. To shorten a quote from Scott Mueler's Upgrading and Repairing PCs (17th Edition):

“ ...to such a scale, the heads in this typical hard disk would be about [...] equal to the Sears Tower if it were tipped over sideways. These skyscaper sized heads would float on a cushion of air that to scale would only be 5mm thick (about 0.2") while [...] circling the earth once every 5 seconds! ”

I don't think this quote specifically is needed, although I enjoy it's accuracy (I can give someone more detail about it in IM where I don't have to worry about copyvios) and proper sense of scale. I think an explanation of how precise the mechanisms are is important --Lucid 14:49, 16 August 2007 (UTC)

Anyone? --lucid 08:34, 29 August 2007 (UTC)
it's a great quote. it's short enough that, properly cited, it wouldn't be a copyright violation to add it. go for it, i say. Anastrophe 15:38, 29 August 2007 (UTC)

I'm not sure it is a reasonable analogy and I'm pretty sure Scott Mueler is not qualified to originate it. I once asked an aeronautical engineer (phD) working on flying heads what was a reasonable analogy so he came up with one based upon scaling the lift pads of a slider to the wings of a 747, something like a 747 flying 1/8" off the ground at 2k mph. He scaled holding the Mach number constant. That slider flew at about 10 times today's flying heights. The other analogy I've heard is kicking a football in the SF Giants ball park and having the ball come down between the goal posts in NYC's Polo Grounds (every 30 msec or so :-). Great quotes, but not of the quality that belongs here. Suggest u look for a quote from a disk drive technologist and not a journalist. Tom94022 01:18, 30 August 2007 (UTC)

There is one in Magnetic Storage Systems Beyond 2000 By George C. Hadjipanayis but it is copyrighted so unless u want to fight the fair use wars u will have to get permission to use it. His numbers are neither Mueler's or mine :-)Tom94022 01:28, 30 August 2007 (UTC)
It's not a matter of who wrote it, just the size comparisons. any analogy is going to vary based on what scale is used, and what hard drive it is based off of. --lucid 11:34, 2 September 2007 (UTC)

[edit] drive limits

Shouldn't this article mention something along the lines of : http://www.tldp.org/HOWTO/Large-Disk-HOWTO-4.html#ss4.2

?  —Preceding unsigned comment added by 70.55.86.204 (talk) 22:01, 3 September 2007 (UTC)

[edit] Should the history section mention Winchester?

Because I was referred here after searching for 'Winchester drive' (as many 'old guys' know this drive), should one not mention 'Winchester' in the history section? Topics on the floppy disk and mini-floppy disk acknowledge the role of Burroughs in Scotland, and no one ever called these 'Glenrothes disks'. Geologist 22:36, 1 November 2007 (UTC)

I don't think so; the history section is quite truncated and given the size of this article, IMHO, rightly so. Winchester is mentioned in footnote 1, covered in more detail in History of hard disk drives article and still more detail in Early IBM disk storage article. That's probably enuf Tom94022 02:56, 2 November 2007 (UTC)
I was also disappointed to see a lack of discussion of any disk technology prior to approx 1998. Mini & Mainframe systems (eg: DEC, IBM) from the 1950s onward had all sorts of interesting disk hardware. 208.62.177.45 (talk) 14:47, 29 May 2008 (UTC)
Name one :-) Actually, you make a good point regarding the History of hard disk drives article, but when I think about it there really were no significant non-IBM disk drives until the late 1960s. Since the early IBM drives are covered in their own articles, the only candidates that come to mind are the Memorex 630 (aka DEC RP01) - first significant non-IBM HDD and the CDC 9760 SMD -first non-IBM media. When I have a few more moments I'll add them. I suppose we could add Bryant as a curiosity. There were also several fixed head disk drives and drums, but that is another article. Tom94022 (talk) 15:09, 29 May 2008 (UTC)

[edit] Performance of competing technologies

I'd like to see a proper summary of the competing technologies - these are actually very badly covered in the individual technology sections of Wikipedia as presumably they are written by people who favour each technology. I've worked for 3 storage vendors all of whom displayed little honesty or knowledge in this area generally preferring the "what we have and more expensive is best" line to performance and reliability discussions. Comparisons SCSI, SAS, Fibre Channel, SATA. SSD

Here is my view of the performance and comparison criteria requirements to truly asses disk technology

1: Reliability - The single most important thing unless you have server level redundancy.
2: Performance under load when 80% full (Read and Write) - everyone runs their disks nearly full - and the newest and most accessed stuff is nearly always near the middle of the disk. Main cause for “my laptop used to be great and now it sucks.” syndrome.
3: Raw transfer speed - Like top speed in your car – oft quoted rarely used.
4: Hot swapability - important for servers and storage systems - can I unplug the disk without interrupting my business and how often does it work.
5: Performance when empty - a few people run lots of large disks empty to maximise performance for the rich and ungreen
6: Length of cable possible - Not normally important
7: Power consumption
8: Heat - normally related to power consumption but a big problem for people who buy 10K disks for their home systems.
9: disk rotation speed - together with how full the disk is probably a key performance indicator - but 10K and above disks don't survive well in uncontrolled environments.
10: Relationship of components - or why having a 2.4 Ghz CPU and a 5400 RPM disk do nothing for either performance or power consumption in your laptop.
11: Interface cheating. - More common than you think - a vendor needs to offer a new technology but doesn't have the technology today so puts on an interface converter - or converter chip on the disk. Voila your ATA disk is a SATA disk but with a negative not positive performance impact.
12: Operating system support. Almost always when new disk technologies are introduced .Example NCQ - the operating system dids not support it. And the first release drivers are always flaky. And then the drivers reveal a problem in the hardware so release 2 hardware comes out.
13: Chipset/Bios support - see above but for the PC or Server hardware
14: Who owns the technology -Is it a licensed technology - this generally affects the price and initial quality. Licensed technologies generally have better initial quality because the specification is controlled not interpreted - the "open/free" technologies generally better later in the cycle. Licensed technologies are almost always more expensive to implement. Firewire and USB are great examples of this cycle.
15: Price
--CDXP 11:19, 10 November 2007 (UTC)

Add to this list an important but always understated element of perceived performance: access time. Compared to drives of the early 90s, today's readily available hard drives are on the order of 10000x bigger and transfer data 50x faster, yet access time has only about doubled. A marginal increase in access speed will pay a much greater dividend than even a sizable increase in capacity or bandwidth to the average consumer who cannot fill a 750GB drive even if he sets out to do so. Whelkman (talk) 17:58, 23 November 2007 (UTC)
It's not for us to decide how HDDs should be judged. Any list of features we'd compare drives by would have to come from some cited, commonly referenced source. Anything else would be possibly tainted by our personal views, and fall under WP:OR. --A D Monroe III (talk) 15:30, 27 November 2007 (UTC)
It is a valid point though, access times are incredibly important, and that fact can easily be made out not to be subjective, through linking to comparisons between gaming load times etc. It just needs to be mentioned anyway, and the surprising development that well, there is no development, access times still suck. This is why SSD's are interesting, or rather one of the reasons.Talrinys (talk) 22:16, 18 December 2007 (UTC)

[edit] Capacity and access speed

I made an edit in this section that was a citation of a source proving the statement "throughput increases, which themselves have not kept up with growth in storage capacity." It was reverted immediately as link spam. The cited page is http://www.mattscomputertrends.com/Time.html. Its my page and research. The editor looked at the page for all of 12 seconds, according to the logs. I think the page is legitimate proof of the statement which otherwise has no cite. Due to my COE here (its my research), I want other opinions please. Matt21811 (talk) 09:23, 21 November 2007 (UTC)

Here's my two cents. Like you say, you're trying to include research that you wrote and it's your page, so there is an obvious conflict of interest in including this information. What's more, the research hasn't been published by any reliable sources so it's considered to be original thought. There's a little more info in the "Advertising and conflicts of interest" section that says that "You should avoid linking to a website that you own, maintain or represent, even if the guidelines otherwise imply that it should be linked." ~a (user • talk • contribs) 12:56, 21 November 2007 (UTC)
i'm the editor who reverted the addition. your page says that all data came from Storage Review. SR is a WP:RS. your site is not. furthermore, by your own post above - you seem to understand part of the issue - it's original research. so i'm not sure why you would have made the edit in the first place, knowing that it fails at least one criterion for inclusion. Anastrophe (talk) 15:54, 21 November 2007 (UTC)
He does absolutely have a point so, what we're looking at is a way to document it? A rather simple comparison would be the time it takes to fill up a harddrive now compared to maybe 10 years ago? That would be straightforward and undeniably accurate Talrinys (talk) 22:42, 7 January 2008 (UTC)

[edit] The caption for a picture is way too long

The picture of an HDD without the platter has a caption that is way too long. I was thinking if I should make it into a new section with the picture and a short caption? 21:46, 8 February 2008 (UTC)

agree, I think we can strike the caption in its entirety since much if not all of it seems to be covered in the paragarph or is too much informationTom94022 (talk) 22:36, 8 February 2008 (UTC)
This is an encyclopedic article, so let's be careful when contemplating 'striking' information. Are we going to lose anything? Is everything the caption states really covered elsewhere? Daniel B. Sedory (talk) 17:49, 10 February 2008 (UTC)
Changed! As near as I can tell the only thing added was the copper motor coils & bearing so I left it in. I changed the term disks to platters to agree with the sketch above, most of what was said in the caption is in the sketch above and in the text of the article. If u find anything in the caption that is not in the article, I suggest it goes in the article. Tom94022 (talk) 19:02, 10 February 2008 (UTC)

[edit] instaling widows

can instaling and uninstaling windos all the time damage your hard drive?eg 4 in two month.from KHATHU —Preceding unsigned comment added by 164.151.129.37 (talk) 11:13, 13 March 2008 (UTC)

Yes, that's an interesting question... My 6 years old Quantum 30GB got its sound very little changed after one reinstallation of Windows 2000 3 years ago. Scandisk found one bad sector and it seems that the heads are seeking slower than before (I have defragmented it before and after that reinstallation and I am absolutely sure that this is not caused by fragmentation). I don't know if anyone else have noticed but my drive is making a little louder noise when the computer is running Windows 2000/XP than when running Windows 98 (and more louder when running Linux) although I use the same software and open the same files. Although after 3 years my drive is still working normal without new bad sectors, I think that the question about wearing the actuators of a drive should be taken seriously, not only when reinstalling Windows, but when defragmenting and choosing which operating system to run. Unfortunately it is difficult to find any information on the Web about drive wearing... Are there any experts who can help? --Lefter 13:01, 8 April 2008 (UTC)

[edit] Screwed Up Captions

The captions in the "Integrity" section clearly do not refer to the pictures they're attached to. They should be changed by someone who is familiar with the significance of the two images. —Preceding unsigned comment added by 128.36.126.120 (talk) 21:53, 23 March 2008 (UTC)

I’m also not familiar with the significance of the two, but if the pictures are in no way in reference to the integrity section then it should be changed.--DavidD4scnrt (talk) 10:04, 25 April 2008 (UTC)

[edit] Turning a drive too soon after shutdown can damage its heads?

I had a misconfiguration problem with my BIOS and my computer used to turn itself back on as soon as it has shut down. I heard that turning on a hard drive too early (<5 sec) can damage the parking mechanism and the heads. Is that true? During the same couple of months, my drive (Quantum 30GB, bought 2001 y.) started seeking irregularly slow (it was defragmented) and I replaced it with a new to avoid losing data. But I couldn't repair my BIOS settings soon after I bought the new drive. Has the turning on caused a little wear/damage to the new drive? It's model is Seagate Barracuda 160GB (ATA 100 variant), SMART currently reports that the Spin Up time is 97 (all other values are 100, except Seek error rate, which is 69, but I hear this is normal for new Seagates)... --Lefter 13:30, 8 April 2008 (UTC)

Highly unlikely, but possible that frequent power cycling would damage heads/disks. Drives are tested thru 10's of thousands of such cycles. In PC's, typically nothing happens until the power supply decays, then the drive dynamically brakes (using stored energy in the motor if necessary) to stop quickly. If the power supply doesn't decay to zero before it starts up again, a badly designed drive might have problems with the transient voltage of the power supply but it would typically lock up rather than damage the heads. It is not just the PC's drive that will have problems in such a situation. I always let my PC set for a few seconds after a "Bill Gates" moment.
Yr defragmented drive with long seek times was possibly experiencing "heroic" error recovery - the drive had errors and never gave up until it corrected them :-). Sometimes this is spot related and u can cure it by surface scan and deallocating the spot, but if the drive is more than a few years old, it's probably time to upgrade
Tom94022 (talk) 16:44, 8 April 2008 (UTC)

[edit] Wrong image caption

The article contains the following image:

Close-up of a hard disk head resting on a disk platter, and its suspension. A reflection of the head and suspension are visible beneath on the mirror-like disk.
Close-up of a hard disk head resting on a disk platter, and its suspension. A reflection of the head and suspension are visible beneath on the mirror-like disk.

The caption is mistaken. No disk platter or mirror-like disk are visible. AxelBoldt (talk) 19:53, 14 April 2008 (UTC)

The caption is correct, the gray, mirror like surface is a portion of a disk. The reflection of the underside of the supension is clearly visible - e.g. note the two holes. Tom94022 (talk) 20:37, 14 April 2008 (UTC)
I believe that's not right. What you see there is two read-write heads and no mirror-like surface at all; the platter would ordinarily be situated between the two levers (one head writes on one side, the other writes on the other side), but the platter was apparently removed for the picture. It might appear like a reflection because the upper lever looks just like the lower lever; both have two holes and cables attached to the actual read-write heads. If it were a true reflection of a single lever however, you wouldn't be able to see the full length of the lower cables. AxelBoldt (talk) 15:53, 15 April 2008 (UTC)
at first i was going to agree with you, then i looked a while longer and became convinced it was indeed a reflection, then after doing some visual processing on the image (primarily increasing brightness) i believe you're correct. if you turn up the brightness, you can see two areas on the lower arm, offset to the left and down, where light cast from above is hitting the lower arm through the two holes in it. if it were a reflection, it wouldn't look like that. i think the neutral gray background that smoothly darkens across the image gives the impression of a reflective surface, but it's a false impression. Anastrophe (talk) 03:04, 16 April 2008 (UTC)
Well, I just emailed the photographer and he said "Yes, the caption is correct. The picture shows one head resting on the disk surface, and its reflection. You may notice that the wires to the head are arranged in exactly the same way in the reflection." I guess that settles it. About the two bright spots you found: they can also result from the reflection. Light passes through the two upper holes, gets reflected off the disk, forms to bright spots on the underside of the upper arm, and we see the reflection of this as the upper side of the lower arm. AxelBoldt (talk) 14:16, 18 April 2008 (UTC)
yep, i kind of thought that about the reflections but dismissed it. the thing with the wires is, since they're all done by machine, i figured there was a very good chance that they always look just like that, with virtually no variations. but a reflection it is, so, its settled! Anastrophe (talk) 01:26, 24 April 2008 (UTC)

[edit] Lead Image

I changed the image (to this one

that I took specifically for WP use with the aim of providing a good illistrative image whilst I had one in pieces and it's been reverted with the explanation that it doesn't show the reflection. I am interested in more information as to why. The image does show a reflection, it clearly shows the reflection of the arm, which really is much clearer that the distorted hand reflected in the other image. Also, I would say that the new image is a much clearer representation of the disk in use with the arm on the head rather than being disassembled with a screwdriver. Furthermore it shows more components of the hard drive - it shows multiple platters, arm, head, static magnet, PCBs, head park and thus has higher enc value for an article on the complete hard disk? Your thoughts? Mfield (talk) 15:47, 25 May 2008 (UTC)

I agree that your picture is far superior. Furthermore, the reflection in the original illustration is at best irrelevant. Tom94022 (talk) 21:58, 26 May 2008 (UTC)
Good, I am not going mad. I am changing it back. If someone wants to revert, they can provide a better justification here. Mfield (talk) 22:03, 26 May 2008 (UTC)