Post by thread:[OT] USB to IDE

Hey Everyone, Sorry for the wrong subject line on my previous post... I have a quick question I'm hoping someone has experience with. I just purchased two USB to IDE adapters from USB Geek (http://www.usbgeek.com/prod_detail.php?prod_id=0044 , also seen in article: http://www.dansdata.com/usbadapt.htm). Unfortunately I was having trouble getting either one of them to work in windows 2k and XP. Windows (both versions) recognizes device, installs driver, sees the hard drive connected, and that’s it. I was expecting to see a new hard drive in explorer, but nope. Tried installing new hardware, but the adapter and hard drive are already installed and working properly. I have no idea how to access the drive (tried with two different HDD's, made sure they are set as master). I've rebooted multiple times with device connected, and tried connecting the power supply, IDE to HDD, and USB to PC connections in every possible order. I must be missing something real simple here... Any clues? Thanks, --Nick __________________________________ Do you Yahoo!? Yahoo! Mail Address AutoComplete - You start. We finish. http://promotions.yahoo.com/new_mail ___________________________________________

FDisk will detect new HDD. When you are in FDisk, you will be given an opportunity to decide whether the entire disk space allocated to one drive name (eg. C:\) or partition into multiple drives (eg. C:\; D:\; E:\ etc..). First drive always in Primary partition while subsequent drive names on the SAME hardisk will have to go to the secondary partition. Once you have completed the above. Go back to Windows and you should will see the 'drive(s)' you created in FDisk. Format the drive(s) according to your needs (eg. FAT16, FAT32 or NTFS). After formatting, you can use the drive(s) as per norm. William Chops Westfield <.....westfwKILLspam@spam@mac.com> wrote: On Oct 28, 2004, at 9:10 PM, Nick wrote: > I must be missing something real simple here... Do they show up to fdisk? I think you need to build a partition table (and perhaps do a format) before they show up to windows. BillW ____________________________________________

part 1 975 bytes content-type:text/plain; charset=ISO-8859-1; format=flowed (decoded 7bit) I've had the same problem with some older drives. The drives that I have tested with have been formatted as FAT16 and FAT32 via a standard board based IDE port . The drives must conform to the ATAPI spec, the other issue is that with some older drives, ie; not ATA but PIO mode drives may not even work. The only way that I have found this out was to search for details about the driver used and the manufacturer of the USB-IDE interface. The unit that I have used is based on a Cypress chip and such the PIO incompatibility warning is given. I hope this helps. Nino. William Chops Westfield wrote: > On Oct 28, 2004, at 9:10 PM, Nick wrote: > >> I must be missing something real simple here... > > > Do they show up to fdisk? I think you need to build a partition > table (and perhaps do a format) before they show up to windows. > > BillW > > ______________________________________________

Oh yes ! I remember a seller told me something about the USB controller chip used on this type of USB to IDE card or external casing, was that, some of the USB controller used are unable to support large capacity HDD while others can. I did ask the seller, how can I tell the difference ? The answer is, I won't be able to tell. In general, the price difference can be used as a yardstick. Although I hate this method but, when I observe the price difference between external USB 2.0 3.5" casing, it can range from say, USD 30 to 70. Before I bought my USB 2.0 3.5" external casing at USD 53.00, I was eyeing on one which is abt USD 30.00. The seller then asked me what is my HDD capacity and I told her, 80GB. She said I can get by with the USD 30.00 one. BUT, if I plan to upgrade to a higher capacity HDD in the near future (eg. 200GB) then the USD 30 WILL NOT be able to support.I should then go for the USD 50+ ones. The above is just FYI. Dominic Stratten <dominic.strattenKILLspamntlworld.com> wrote: I had the same problem with Seagate 120gb drives and an external USB->IDE controller. It would not under any circumstances recognise the drive and would lock the controller up. I got round the problem by setting the drive as a slave but this only worked on one computer. The other problem I had was that the USB->IDE controller was not backwardly compatible with USB 1.x - I ended up buying a new motherboard with USB 2.0 on and that solved almost all of the problems (still didn't work properly with the 120gb Seagate though). {Original Message removed}

On Fri, 29 Oct 2004, techy fellow wrote: > FDisk will detect new HDD. When you are in FDisk, you will be given an > opportunity to decide whether the entire disk space allocated to one > drive name (eg. C:\) or partition into multiple drives (eg. C:\; D:\; > E:\ etc..). First drive always in Primary partition while subsequent > drive names on the SAME hardisk will have to go to the secondary > partition. > > Once you have completed the above. Go back to Windows and you should > will see the 'drive(s)' you created in FDisk. Format the drive(s) > according to your needs (eg. FAT16, FAT32 or NTFS). After formatting, > you can use the drive(s) as per norm. > > William Chops Westfield <.....westfwKILLspam.....mac.com> wrote: > On Oct 28, 2004, at 9:10 PM, Nick wrote: > >> I must be missing something real simple here... > > Do they show up to fdisk? I think you need to build a partition > table (and perhaps do a format) before they show up to windows. Afaik fdisk cannot be used for usb, firewire and other 'modern' drives before the drives are formatted. However, there is often a utility cd that comes with the drive and allows formatting. There should be a cdrom or a download for this somewhere imho. Go to the adapter enclosure maker's page and see what instructions there are. Peter ____________________________________________

"Peter L. Peres" <EraseMEplpspam_OUTTakeThisOuTactcom.co.il> wrote: > On Fri, 29 Oct 2004, techy fellow wrote: > > FDisk will detect new HDD. > > Afaik fdisk cannot be used for usb, firewire and other 'modern' drives > before the drives are formatted. On Win2K, go to Control Panel --> Administrative Tools --> Computer Management --> Storage --> Disk Management This will show all drives, formatted or not, including USB and Firewire, and you can allocate space for partitions on unformatted drives from here. AFAIK, this would be the preferred path on XP as well. -- Dave Tweed ____________________________________________

Nick, One thing I've noticed about usb and mass storage devices is that the drivers generally need to be installed BEFORE the physical drive is first plugged in. If not, the device will get installed as a standard/generic USB mass storage device and will usually not work after that. What needs to happen in this case is that drive should be removed, and ALL the approriate drivers & control-panel devices be uninstalled before re-installing the device. Perhaps try this on another Windows installation in the correct order to see if this really solves the problem, and to see which drivers should be removed. Cheers, -Neil. On Thursday 28 October 2004 11:10 pm, Nick scribbled: {Quote hidden}> Hey Everyone, > > Sorry for the wrong subject line on my previous > post... > > I have a quick question I'm hoping someone has > experience with. > > I just purchased two USB to IDE adapters from USB Geek > (http://www.usbgeek.com/prod_detail.php?prod_id=0044 , > also seen in article: > http://www.dansdata.com/usbadapt.htm). Unfortunately > I was > having trouble getting either one of them to work in > windows 2k > and XP. > > Windows (both versions) recognizes device, installs > driver, sees > the hard drive connected, and that’s it. I was > expecting to see > a new hard drive in explorer, but nope. Tried > installing new > hardware, but the adapter and hard drive are already > installed > and working properly. I have no idea how to access > the drive > (tried with two different HDD's, made sure they are > set as > master). > > I've rebooted multiple times with device connected, > and tried > connecting the power supply, IDE to HDD, and USB to PC > connections in every possible order. > > I must be missing something real simple here... Any > clues? > > Thanks, > > --Nick > > > > __________________________________ > Do you Yahoo!? > Yahoo! Mail Address AutoComplete - You start. We finish. > http://promotions.yahoo.com/new_mail > _____________________________________________

Do I have this right? All adapters do not support large drives (200G+ ) USB 2 is faster than firewire (480mbs vs 400mbs) USB1.1 works but is slow USB2 is more than 4X as fast as a drive attached over a 100mbs network. USB attachment is not affected by bios limits(?) I am looking at ($60.88)Brand IOGEAR Manuf Part GHE135C from http://www.provantage.com/buy-7IOGR03R-3-5in-usb-2-0-firewire-hd-ion-drive-combo-enclosure-w-cables-iogear-ghe135c-shopping.htm Any comments? John Ferrell My Competition is not my enemy! http://DixieNC.US ____________________________________________

John Ferrell writes: > Do I have this right? > All adapters do not support large drives (200G+ ) I believe this is correct, though I have a recollection that the cutoff is typically 137G. > USB 2 is faster than firewire (480mbs vs 400mbs) > USB1.1 works but is slow > USB2 is more than 4X as fast as a drive attached over a 100mbs network. You are talking strictly theoritical numbers here. Reality is quite a bit different. First off, you need to know a few things about USB 2.0. The USB 2.0 spec allows for transmission at 1.5, 12, and 480 Mbps. There are a *lot* of USB 2.0 devices that only operate at 1.5 and 12 Mbps speeds! To make matters worse, the 12 Mbps USB 2.0 is designated by the name "Full Speed USB 2.0" while 480 Mbps USB 2.0 is designated "Hi-Speed USB 2.0" Now, just because a device can operate on on the 480 Mbps bus does not mean that the device can operate at those speeds! You've got bottlenecks in the IDE->USB convertor, the USB bus itself, the USB host controller, the speed of the host CPU, and the quality of the drivers. Needless to say, the device itself needs to be capable of generating data above the bus speeds to fully utilize what speed is available on the bus. Most moden IDE hard disks are capable of 50 - 80 Mbytes/sec when connected to a local IDE bus, so you *should* be able to saturate the Hi-Speed USB bus. This is all simply assuming that the hard disk is the only device on the bus. Don't forget that if you have any other devices on the bus, particularly slower devices, that they will take up quite a chunk of bandwidth as well! > USB attachment is not affected by bios limits(?) The only time BIOS limits come into play is when you want to *boot* off of the device. Once you've got your OS up and running, it goes directly to the hardware. -p. ____________________________________________

On Sat, 2004-10-30 at 09:03, John Ferrell wrote: > Do I have this right? > USB 2 is faster than firewire (480mbs vs 400mbs) No. While firewire has a lower theoretical speed, in the real world the difference isn't as certain, there are cases where USB is faster, there are cases where firewire is faster. Remember, firewire and USB were never meant to be direct competitors. USB generally requires far more power from the host CPU to transfer data, firewire requires far less. > USB1.1 works but is slow For a HD? Yes, I'd consider it to slow. > USB2 is more than 4X as fast as a drive attached over a 100mbs network. Most certainly not. There is alot of overhead in USB2. Again, raw bandwidth is never a good determination for technologies that are so different. In the real world USB2 is generally faster then 100Mbps ethernet, but then only marginally. Also, remember that you will also be hitting limits of hard drive speed, few hard drives can pump data out continuously at 480Mbps, so the interface becomes less important. > USB attachment is not affected by bios limits(?) True, unless you intend to use a USB hard drive for boot, in which case BIOS issues do come into play. TTYL ----------------------------- Herbert's PIC Stuff: http://repatch.dyndns.org:8383/pic_stuff/ ____________________________________________

Herbert Graf writes: > Also, remember that you will also be hitting limits of hard drive speed, > few hard drives can pump data out continuously at 480Mbps, so the > interface becomes less important. That's not quite true. On my Linux box which is hardly state-of-the-art, I get 70+ Mbytes/sec (560+ Mbits/sec) as reported by hdparm. Needless to say, this isn't a real-world application speed, but it does show what the hardware is capable of from the drive through the cable through the host controller, through the drivers, all the way to the application. It is quite likely that raw speed from the drive itself is even greater, though how much so I couldn't begin to guess. -p. ____________________________________________

AHA! I am glad I asked..... >> USB2 is more than 4X as fast as a drive attached over a 100mbs network. > > Most certainly not. There is alot of overhead in USB2. Again, raw > bandwidth is never a good determination for technologies that are so > different. In the real world USB2 is generally faster then 100Mbps > ethernet, but then only marginally. > My current backup has the large drive in an empty IDE slot on the network wherever convienient. Since there is no need for portability, I might as well continue with that and save my $60. John Ferrell My Competition is not my enemy! http://DixieNC.US {Original Message removed}

On Sat, 2004-10-30 at 12:31, Peter Johansson wrote: > Herbert Graf writes: > > > Also, remember that you will also be hitting limits of hard drive speed, > > few hard drives can pump data out continuously at 480Mbps, so the > > interface becomes less important. > > That's not quite true. On my Linux box which is hardly > state-of-the-art, I get 70+ Mbytes/sec (560+ Mbits/sec) as reported by > hdparm. Needless to say, this isn't a real-world application speed, > but it does show what the hardware is capable of from the drive > through the cable through the host controller, through the drivers, > all the way to the application. It is quite likely that raw speed > from the drive itself is even greater, though how much so I couldn't > begin to guess. Sorry Peter, but I think you are misinterpreting something. First off, how did you run hdparm? Did you use -t or -T? For example, on my box: [hgraf@pIII866 /tmp]# hdparm -t /dev/hda /dev/hda: Timing buffered disk reads: 52 MB in 3.05 seconds = 17.03 MB/sec [hgraf@pIII866 /tmp]# hdparm -t /dev/hda /dev/hda: Timing buffered disk reads: 74 MB in 3.06 seconds = 24.83 MB/sec [hgraf@pIII866 /tmp]# hdparm -t /dev/hda /dev/hda: Timing buffered disk reads: 84 MB in 3.03 seconds = 27.68 MB/sec [hgraf@pIII866 /tmp]# hdparm -t /dev/hda Now, my drive is about 6 months old, but still, newer drives aren't going to be 3 times faster (well, newer CONSUMER drives, there are server drives which ARE much faster, but if you had such a fast drive you would tell us right? :) That said, we are talking about the average consumer hear, I don't see ANYONE putting a 12000 rpm server drive into a USB casing...). On top of this, realize that hdparm is a VERY simple way of measuring a hard drives speed. From the manpage: This measurement is an indication of how fast the drive can sustain sequential data reads under Linux, without any filesystem over- head. Notice the word "sequential". This is the BEST case for a hard drive, a fetch of sequential data. Unfortunately rarely in real life does data end up being sequential. A TRUE gauge would be a random fetch of data, which would come up far slower. There is no doubt that there are cases where a hard drive will reach 480Mbps, but in normal use that happens very rarely. TTYL ----------------------------- Herbert's PIC Stuff: http://repatch.dyndns.org:8383/pic_stuff/ ____________________________________________

Herbert Graf writes: > Sorry Peter, but I think you are misinterpreting something. First off, > how did you run hdparm? Did you use -t or -T? For example, on my box: I didn't just fall off the turnup truck. ;-) Those results were done with both -T and -t. # hdparm -T -t /dev/hde /dev/hde: Timing cached reads: 1484 MB in 2.00 seconds = 742.00 MB/sec Timing buffered disk reads: 156 MB in 3.03 seconds = 51.49 MB/sec # hdparm -i /dev/hde /dev/hde: Model=Maxtor 6Y160M0, FwRev=YAR51BW0, SerialNo=XXX Config={ Fixed } RawCHS=16383/16/63, TrkSize=0, SectSize=0, ECCbytes=4 BuffType=DualPortCache, BuffSize=7936kB, MaxMultSect=16, MultSect=16 CurCHS=16383/16/63, CurSects=16514064, LBA=yes, LBAsects=268435455 IORDY=on/off, tPIO={min:120,w/IORDY:120}, tDMA={min:120,rec:120} PIO modes: pio0 pio1 pio2 pio3 pio4 DMA modes: mdma0 mdma1 mdma2 UDMA modes: udma0 udma1 udma2 AdvancedPM=yes: disabled (255) WriteCache=enabled Drive conforms to: (null): * signifies the current active mode This happens to be an older SATA drive that was never optimised for my system. (older SATA drives tend to be slower than their PATA counterparts, BTW) The IDE-133 drive that gave 70 MB/sec was imaged for a client. (Unfortunately, I cannot remember the specific model.) > On top of this, realize that hdparm is a VERY simple way of measuring a > hard drives speed. From the manpage: I know this as well. That is why I specifically mentioned that these were the results from hdparm and *not* from any real-word usage or even simulated real-world usage. > Notice the word "sequential". This is the BEST case for a hard drive, a > fetch of sequential data. Unfortunately rarely in real life does data > end up being sequential. A TRUE gauge would be a random fetch of data, > which would come up far slower. Even more than this, hdparm only tests the beginning of a partition (or disk if the raw device is given) where even sequential access is fastest. If I remember correctly, the drive above drops down to something list 38 MB/sec at the innermost tracks. Now as to real world usage, this depends entirely on your application. If you happen to be imaging disks on two different IDE busses, then you really will see this throughput. If you happen to be processing large files such as video and your disk has low fragmentation, you'll still see enough end-to-end throughput such that even the Hi-Speed USB 2.0 bus will be a bottleneck. In many cases, this is exactly what people use external USB/Firewire drives for and is one of the reasons Apple came out with 800 Mbps Firewire. Point is, you need to optimise for your application. USB bus speed may be a factor, or it may not. But you need to know enough background as to how things work so you can develop some proper tests and benchmarks and properly interpret the results. -p. ____________________________________________

On Oct 30, 2004, at 9:15 AM, Herbert Graf wrote: >> USB2 is more than 4X as fast as a drive attached over a 100mbs >> network. > > Most certainly not. There is alot of overhead in USB2. ... In the real > world USB2 is generally faster then 100Mbps ethernet, but then > only marginally. Really? There's a lot of overhead for the network as well; I would have guessed considerably MORE overhead than for USB2, although I guess it depends rather a lot of what you're using a a remote file access protocol. In my own experience, firewire and (high speed) USB2 are "fast enough" for disks, and USB1.1 isn't (except perhaps as a backup media.) Strange to think that 10Mbps was "fast enough" for all that early NFS stuff. I guess files and disks were just a lot smaller in those days, so slower transfers weren't as noticeable. BillW ____________________________________________

Herbert Graf <piclistspam_OUTmit.edu> wrote: > USB generally requires far more power from the host CPU to transfer > data, firewire requires far less. Well, yeah... Mostly because USB was deliberately designed that way. The idea was to simplify the design of the devices by making the PC host do most of the work. > There is alot of overhead in USB2. .... In the real world USB2 is > generally faster then 100Mbps ethernet, but then only marginally. Hmm... If the "overhead" you're talking about is outside the USB interface (slow CPU, a chip that can't do high-speed ATA, badly-written drivers, USB applications written in interpreted BASIC, etc.) that may be true, but our USB-to-ATAPI bridge chip can talk to a drive at 96 megabytes/second and stream data over USB at 53 megabytes/second. > Also, remember that you will also be hitting limits of hard drive > speed, few hard drives can pump data out continuously at 480Mbps, so > the interface becomes less important. Yeah, that's definitely true. -Andy === Andrew Warren - @spam@aiwKILLspamcypress.com === === Principal Design Engineer === Cypress Semiconductor Corporation ____________________________________________

On Oct 30, 2004, at 9:10 PM, Andrew Warren wrote: > Mostly because USB was deliberately designed that way. > The idea was to simplify the design of the devices by making the PC > host do most of the work. I was always under the impression that a large part of the host-side effort was done by the USB controller chip hardware, and not in software, giving you (once everything was set up) a rather disk-like interface where chunks of data got DMAed into the address space every tick for the software to hand to the user with little additional effort. Isn't that true? BillW ____________________________________________

Chops <KILLspampiclistKILLspammit.edu> wrote: > > The idea was to simplify the design of the devices by making the > > PC host do most of the work. > > I was always under the impression that a large part of the host-side > effort was done by the USB controller chip hardware, and not in > software Tell that to the guys who had to write USB drivers for Linux... Although a lot of work is done by the hardware -- less by USB 1.1 UHCI controller chips, more by OHCI controllers and USB 2.0 EHCI controllers and hubs -- there's still plenty for the host to do in software. That's one reason Linux machines circa 2002 (kernel 2.4 or so) weren't able to achieve more than about 20% of the theoretical top speed on USB mass-storage transfers. -Andy === Andrew Warren - RemoveMEaiwTakeThisOuTcypress.com === === Principal Design Engineer === Cypress Semiconductor Corporation ____________________________________________