Talk:USB flash drive/Archive 1

Name?

Many months and no objections to the change. I suppose I'll be bold and change it. That'll either respark discussion or fix things depending on the reaction. --Sketchee 21:44, Jun 3, 2005 (UTC)

Agreed. The title of this article is definetly misleading and somewhat inaccurate. Kail Ceannai 19:57, 2005 May 9 (UTC)

I agree - the name should be USB Flash Drive. It's succinct enough, yet it describes what the devices are - flash memory drives that plug into usb ports. And this name seems to be starting to be adopted as a standard...--Blackcats 08:00, 24 Mar 2005 (UTC)

[1] is an nytimes.com vote from 2004-09-23 for USB Flash Drive, I quote: "these slender solid-state memory chips -- known by many names, but officially U.S.B. flash drives". (I link you thru Google because otherwise the nytimes.com copy protection kicks in to force you to register directly with nytimes.com). 23 Jan 2005

Would this article be better named "USB flash memory drive". To me "Keydrive" seems to fail the test of precision, particularly with the advent of "keydrives" that are disk-based (eg the Digitalway MPIOHS100) rather than flash-memory-based. Nurg 06:59, 20 Jan 2004 (UTC)

That's my feeling as well. Deh 15:23, 8 Oct 2004 (UTC)

Isn't keydrive a thumb drive double as a keychain? :P wshun 11:27, 10 Apr 2004 (UTC)

I think it would fit better with Wikipedia conventions to be called USB flash drive. --65.185.15.125 23:14, 19 Jun 2005 (UTC)

phew! that was double redirect hell. if someone comes along and says Flash is a proper noun or something, they can do the work to change it back... - Omegatron 02:54, Jun 20, 2005 (UTC)

Max speed

I just added a note about the max speed of USB2.0 devices being around 100Mbit/s - the current fastest (that I can find) is a bit faster than this, but most are a fair amount slower. Here's a reasonable survey of the current state of the art: [2]. -- Finlay McWalter | Talk 23:16, 27 May 2004 (UTC)

There are currently 2 USB standard as defined in www.usb.org. The maximum transfer speed for USB1.1 is 12Mbit/sec (or 1.5MB/s) and for USB2.0 is 480Mbit/sec (or 60MB/s). However, no usb drives can achieve the theoretical speed due to constraints in hardware and software limitations. That is why "depending on system configuration" is the small fine print behind the speed claim.

Many suppliers boost of high speed transfer referring only to reading speed. The catch is "slow" speed write that is important when you wish to transfer data from PC/notebook to usb drive.

The fastest usb disk is Onlydisk offered by www.netac.com. The reading speed is 160Mbit/sec (20MB/s) and writing speed is 128Mbit/sec (or 16MB/s). The company claimed that they have a patented technology (aka Ultra Stable Technology) that prevent disk or data corruption which is a prevailing problem found in most usb drives. When a disk is disk corrupted, user can only throw and buy a new one. But valuable data will be lost forever with the death of the disk.

Resource

This Ars Technica article has some excellent info about what's in the box, and gives concrete numbers for speed. When I have time I'll mention some of the stuff it does (or links to) into wikipedia's article. -- Finlay McWalter | Talk 20:27, 3 Jun 2004 (UTC)

That link is bad. Do you mean this article? - John Fader

Lexar Jump Drive

Please, somebody could tell me, What is the difference between usb flash memory and usb jump drive?, this two are a keydrive, but what is the difference?

As far as I can tell, Jumpdrive is a brand name of Lexar, for their top-of-the-range keydrives. If that's what you mean, then a jumpdrive is just a keydrive. -- John Fader 17:00, 16 Feb 2005 (UTC)

Lexar for whatever reason luvvvvs to make you install a custom driver set for their products rather than the generic drivers. SchmuckyTheCat 21:59, 25 Mar 2005 (UTC)

I'm not sure that's right. If you have an updated version of Win XP, ME or 2000, most Lexar flash devices should work fine without custom drivers. The only ones that explicitly require custom drivers are the devices with special DRM encryption. -- FP 02:04, Apr 10, 2005 (UTC)

This is the truth. I don't remember if my Lexar drive even came with a driver CD. I remember it recommending letting the native OS drivers handle it, but I remember that you had to download or install drivers from CD if you wanted to use it on Windows 9X machines, as with all flash drives.

The fact is USB drive cannot work in Win95, NT and Win 98 (1st edition). The reason being Microsoft has not catered for USB software driver for such devices before 1999. With the introduction of OnlyDisk (just another name of USB drive) patented in 1999, Microsoft added the driver in Win98 (2nd edition) and other later versions (Win2000, WinME, Win XP). The USB drive can also work in Linux 2.4.x and Mac 9.x above.

Hmm i run a number of 98 second edition systems and i've always had to use manufacturers drivers to use usb flash sticks with them. There is an unofficial service pack that claims to include a generic driver for theese sticks but i haven't had any luck with that driver either. Plugwash 19:27, 1 October 2005 (UTC)

Read/Write Cycles?

"flash drives wear out after 10,000 read/write cycles"

and

"In normal use, mid-range flash drives currently on the market will support several million cycles, although write operations will gradually slow as the device ages"

Are these referring to different things or ... what is the difference between these two statements? --Sketchee 21:58, Jun 3, 2005 (UTC)

I think this is simply incorrect.

Googling for the second phrase (in quotes) turns up several wikipedia quotes, and one article that appears to be original? (at least they don't mention wikipedia, and the article is significantly different than the current article).

Single Layer Cell flash memory has 100,000 read/write cycles and Multi Level Cell flash memory has 10,000. I don't know where this "millions" number is coming from. Could it be a different measurement of the same limit? - 71.103.83.28 18:14, 14 July 2005 (UTC)

About eight or nine years ago (near the beginning of the mass use of flash) I worked on an embedded product which used flash for data (not just application) storage. I had occasion to discuss flash wear characteristics with field engineers from both AMD and Intel. The issue of how many erase cycles a given part could sustain wasn't nearly so cut and dried. At the time they guaranteed only 1000 cycles (clearly technologies have improved greatly since) but very informally the said you could expect an order of magnitude better than that in reasonable use. Factors which affected this included operating and storage temperatures (and microdamage caused by thermal changes, particularly during shipping), variation in erase voltage (the parts used a 5v erase voltage, against the 3.3v used for reading and writing) and variation due to manufacturing issues both at the fab and at our assembly (SMT) facility. Usage patterns also affect the real-world mtbf of a multisectored flash part - block size, the pattern with which you wrote stuff, and the characteristics of the block erase and block wear leveling algorithms you use. So I'd be willing to accept that the 100,000 is the chip manufacturer's guaranteed number (they're a conservative bunch) and that in most cases, with sensible use and handling, you'd get a million or two cycles from them. -- Finlay McWalter | Talk 18:51, July 14, 2005 (UTC)

The term "read/write cycles" is inherently confusing because it doesn't clarify whether reading or writing wears out the memory, or both -- and if both, which wears it faster. Reading only senses the state of floating-gate charges, but writing requires erasing (discharging) and re-writing (recharging). Some applications might read the memory many times between writes, so the difference between read cycle wear and write cycle wear will be significant in this case.--69.141.120.204 02:11, 31 July 2005 (UTC)

Foreign names

In the list of differing names for these devices, why are other languages listed? There is not confusion just because different languages use different names for things. 4.242.147.147 02:57, 14 Jun 2005 (UTC)

Kikinou ??

Can anyone tell where they heard the word Kikinou ? I never heard that in France !

mini usb

"Most flash drives feature the larger type-A USB connection, although some feature the smaller "miniUSB" connection."

Unless a source is given for the use of miniUSB I'm removing it. It seems very unlikely to be true, given that the whole point of these devices is to plug straight into a USB port.

Strengths and weaknesses

This section lists several strengths but no weaknesses (apart from saying that like all means of data storage it will eventually fail). Maybe something could be mentioned about the problems of non-compatability (I realise this is discussed elsewhere in the article) and memory corruption which, although I'm not an expert on the subject, I think can on occasion occur.

I added some comments on form factor which has been the biggest gotcha for me with various manufacturers products. Garglebutt / (talk) 01:51, 31 July 2005 (UTC)

Confusing text

"Most flash drives feature the standard type-A USB connection allowing them to be connected directly to the port. Some small drives have been made with a thin plug designed to mate with a standard usb port[1] but these are very rare."

These two sentences seem to describe the same thing. A drive that connects to a standard USB port, and a drive that connects to a standard USB port. The only difference seems to be in the wording - "connect" vs "mate", etc. This is very confusing as it's clearly intended to describe two different things.

one describes a standard USB type A plug, the other describes a plug that still mates with a standard socket but looks nothing like a standard plug. Plugwash 09:39, 1 August 2005 (UTC)

Names

I removed one of the references to the memory stick name under the Naming Confusion section, as it was listed twice.Ynos 17:32, 2 August 2005 (UTC)

I got rid of "that flash-card USB thingy" because seriously, no one calls it that, and it's very un-encyclopidic, if thats even a word.

Largest capacity pen drive available?

What's the USB flash drive (a pen shaped one) with the most capacity currently available? On Amazon UK the highest I found was 1gb. I'm considering a purchase. --bodnotbod 15:35, 18 September 2005 (UTC)

Dabs sells several 2Gb ones - http://www.dabs.com/uk/channels/hardware/storage/productView.htm?quicklinx=3KQG http://www.dabs.com/uk/channels/hardware/storage/products.htm?catid=216 -- Finlay McWalter | Talk 15:46, 18 September 2005 (UTC)

Thanks for such a quick reply. Still no match in storage terms for a DVD then, though I appreciate the differing merits (well, a bit). -- (bodnotbod, not logged in).

Jetflash has a 4GB model, as do several other manufacturers. Bloodshedder 03:59, 19 September 2005 (UTC)

dabs is now listing several 4GB drives http://www.dabs.com/productlist.aspx?&NavigationKey=11152&NavigationKey=40096&CategorySelectedId=11152&PageMode=1 nothing above that yet though (well unless you count that 5 gig sony device but thats not flash. the 4 gig ones are bloody expensive though. Plugwash 23:27, 7 October 2005 (UTC)

ThinkGeek exlink

I'm not quite sure which product the ThinkGeek link is supposed to reference (I see a list). If it's the Pocketec one, this is a more direct and informative link, I think. -- Finlay McWalter | Talk 22:57, 24 September 2005 (UTC)

When i added the link it went straight to a product very similar to the one i have now linked (though i belive a different brand). It seems thinkgeek have now dropped the product and made the link for it take you to the section menu. Plugwash 23:15, 24 September 2005 (UTC)

Clarifiaction needed

The flash drive was first invented in 1998 at IBM as a floppy drive replacement for the ThinkPad line of products. Although there is an IBM disclosure, they did not patent it. IBM later contracted M-Systems to develop and manufacture it non-exclusively. M-Systems holds the patent to this device, as well as a few other related patents.

If IBM had publically disclosed this idea and contracted someone to make it non-exclusively how exactly can that other company have a patent on it? Plugwash 19:20, 1 October 2005 (UTC)

A "disclosure" is a legal submission which says "I/we have or may have some intellectual property in this or that field"; it doesn't actually mean to disclose anything substantive about that invention (doing that, publically, would invalidate any unfiled patent applications). Disclosures are most commonly filed with standards bodies (although they're found in all kinds of inter-company agreements). Say, for example, you're on the ANSI Exploding Weasel subcommitee. ANSI will require you to disclose that you have some pending intellectual property in the field of mustela detonation, but you don't have to disclose what it is. In this case it's likely that IBM tranferred (sold/gave/swapped) the revelant IP to M-Systems, having previously disclosed that they were working on something sorta kinda link that, in some public agreement or filing. -- Finlay McWalter | Talk 23:40, 7 October 2005 (UTC)

ok so if there is a patent what number is is it? what exactly does it cover? is it likely to stand up in court? and are other manufacturers generally licenseing it or ignoring it? Plugwash 00:20, 8 October 2005 (UTC)

Looks like patent # 6,148,354. Bloodshedder 05:37, 8 October 2005 (UTC)