USB2.0 to IDE Drive - frequently corrupts drive. EVent ID 55, Source: Disk

[trmbr]

Hi,

I have a "USB2.0 to IDE" external case and a 120GB Seagate drive (ST312002
2A drive) - which I connect to my laptop and desktop system - and
periodically gets corrupted. The external case has its own power source,
USB2.0 to IDE cable which is connected to a USB2 PCI adapter.

I have had to FDISK and FORMAT the drive repeatedly and am now frustrated as
there must be a reason for the corruption.
Has anyone experienced this before ort have a resolution?

I format the 120gb drive using a Windows 98SE bootable floppy. The system
the drive is connected to only permits a Primary Partition of 48.9GB.
However I can format the full 120GB (111GB actually due to the drive
overhead) in NTFS format, successfully on my Win2000 laptop.
I have been able to successfully backup my 40GB laptop to the 120GB drive a
few times however it intermittently corrupts, indicated by a repeated entry
in the Application log of
Source NTFS
Event Id 55,
Category isk and a
description of "The file system structure on the disk E: is currupt and
unusable. Please ruin the chkdsk utility on the volume E"

Any assistance will be appreciated
Thanks
trmbr

 

[Odie Ferrous]

Hi,

I have a "USB2.0 to IDE" external case and a 120GB Seagate drive (ST312002
2A drive) - which I connect to my laptop and desktop system - and
periodically gets corrupted. The external case has its own power source,
USB2.0 to IDE cable which is connected to a USB2 PCI adapter.

I have had to FDISK and FORMAT the drive repeatedly and am now frustrated as
there must be a reason for the corruption.
Has anyone experienced this before ort have a resolution?

I format the 120gb drive using a Windows 98SE bootable floppy. The system
the drive is connected to only permits a Primary Partition of 48.9GB.
However I can format the full 120GB (111GB actually due to the drive
overhead) in NTFS format, successfully on my Win2000 laptop.
I have been able to successfully backup my 40GB laptop to the 120GB drive a
few times however it intermittently corrupts, indicated by a repeated entry
in the Application log of
Source NTFS
Event Id 55,
Category isk and a
description of "The file system structure on the disk E: is currupt and
unusable. Please ruin the chkdsk utility on the volume E"

Any assistance will be appreciated
Thanks
trmbr

I'm researching data corruption with USB devices at the moment; it is
occurring fairly extensively. USB should not be considered as a method
of connectivity for (important) backup media. Maxtor is about to learn
this with their OneTouch drives.

Firewire is more reliable, although there are certain chipsets to avoid.


Odie

 

[Neil Maxwell]

I have a "USB2.0 to IDE" external case and a 120GB Seagate drive (ST312002
2A drive) - which I connect to my laptop and desktop system - and
periodically gets corrupted. The external case has its own power source,
USB2.0 to IDE cable which is connected to a USB2 PCI adapter.

There's a long thread here that addresses large file corruption with
USB. The context is True Image support, but I understand it's not a
TI issue.
http://www.wilderssecurity.com/showthread.php?t=65129

You might want to give it a quick browse and see if it's what you're
seeing. I use Firewire as my primary backup and keep the file sizes
to 700 MB for flexibility, and haven't seen any corruption problems.

 

[Guest]

Firewire is more reliable, although there are certain chipsets to avoid.

which ones?

 

[trmbr]

There's a long thread here that addresses large file corruption with
USB. The context is True Image support, but I understand it's not a
TI issue.
http://www.wilderssecurity.com/showthread.php?t=65129

You might want to give it a quick browse and see if it's what you're
seeing. I use Firewire as my primary backup and keep the file sizes
to 700 MB for flexibility, and haven't seen any corruption problems.

Thanks for the responses - much appreciated.
I follow the link provided, and below is my experience.

I have experienced problems using other backup software, i.e. SyncBack,
which has resulted in intermittent USB-IDE Drive corruption.
I use a combination of Windows2003Server, Windows 2000 Pro and Windows XP. I
connect my external 120gb USB11/20 external drive to all 3 systems, and
irrespective of the O/S, corruption intermittently occurs, rendering the
entire drive inaccessible - an FDISK and FORMAT are again required.
The drive isnt faulty as Ive run diagnostics and have used it successfully
before. Only when connected to a USB port, its corrupts.

I "unplug" the device as per the USB Unplug/Eject hardware utility. I dont
power off the external power, until thr drive is reported to be safe to
remove.

The external USB-IDE case was purchased at a computer market, therefore I
dont know the brand. I suspect the corruption occurs when the 120GB drive
exceeds 70GB - has anyone experienced this capacity limit?

Regards
trmbr

 

[CWatters]

I suspect the corruption occurs when the 120GB drive
exceeds 70GB - has anyone experienced this capacity limit?

I don't see any problem at 70GB on my XP only set up. My drive hovers around
75GByte all the time and I change about 20GB every night - seems to work ok.

Perhaps it's a problem with older OS?
Power supply?
Overheating?
Suspect USB cable (USB 2.0 cables are different to 1.1 only cables)

What about running a drive test program? Leave it going over a weekend to
see if you have some duff sectors.

 

[J. Clarke]

I don't see any problem at 70GB on my XP only set up. My drive hovers
around 75GByte all the time and I change about 20GB every night - seems to
work ok.

Perhaps it's a problem with older OS?
Power supply?
Overheating?
Suspect USB cable (USB 2.0 cables are different to 1.1 only cables)

There is no standard that specifically defines a USB 2.0 cable--a cable that
meets the USB 1.1 standard works fine for USB 2.0 and is the specified
cable for that use. Where the problem comes in is that the USB 1.1
standard overengineered the cable so some cable manufacturers started
producing substandard cables that still worked fine for 1.1.

 

[Peter]

There is no standard that specifically defines a USB 2.0 cable--a cable
that

meets the USB 1.1 standard works fine for USB 2.0 and is the specified
cable for that use. Where the problem comes in is that the USB 1.1
standard overengineered the cable so some cable manufacturers started
producing substandard cables that still worked fine for 1.1.

What about section 6 and 7 in "Universal Serial Bus Specification Revision
2.0"? There is a lot of information about USB cable specs. They may have
used "high-/full-speed USB cable" term instead.

By the way, there are three USB cable assemblies: standard detachable cable,
high-/full-speed
captive cable, and low-speed captive cable.

 

[Mike S.]

Just found this thread, and found it interesting as I am also experiencing
data corruption in a slightly different setting.

I have one of those pocket-sized micro USB hard drives - the 2.2 GB drive
from GS-MagicStor that CompUSA sells under its house brand.

I primarily use it for short-term storage and transfer of files between
several computers, much as I would a big flash drive. Every now and then,
it gives me a "surprise".

On one occasion, a folder I had created became inaccessible when trying to
read the drive on another computer. The directory entry for the folder
itself had become trashed, rendering the contents to be lost as file
fragments. After that, I started running chkdsk or scandisk immediately
after plugging the drive into a new machine.

Occasionally, I will get a message that the boot sector is damaged and I
allow it to be repaired. Nothing else found.

At one point, however, I had an instructive experience. I had an .RAR file
on the pocket drive, and it needed to be updated. Rather than update in on
the PC's internal drive and then copy the new version back, I allowed
WinRAR to update the archive on the USB drive. When I let WinRAR test the
newly-updated archive, one file failed. Extracting it, there was a random
1-byte error. It occurred to me that, after copying lots of megabytes of
files without error, a critical factor might be CPU activity at the time
of the data transfer.

I don't have access to sophisticated testing tools, and the drive itself
is a closed system so I can't, say, remove it and try it on an IDE
controller to see if there are problems there. But the experience has
really soured me on the prospect of portable high-capacity storage.




tes

 

[J. Clarke]

What about section 6 and 7 in "Universal Serial Bus Specification Revision
2.0"? There is a lot of information about USB cable specs. They may have
used "high-/full-speed USB cable" term instead.

By the way, there are three USB cable assemblies: standard detachable
cable, high-/full-speed
captive cable, and low-speed captive cable.

"captive cable" is permanently attached to a device--your choice is to use
it, toss the device, or second-guess the designer, open it up, and attach
your own cable to it.

As for section 6 and 7 of the USB 2.0 spec, yes, there are cable
specifications. It may have escaped your notice that there are also cable
specifications in 1.1. Now, would you care to list the particulars in
which a detachable cable compliant with the USB 2.0 spec will be different
from one that is in compliance with the final revision of the 1.1 spec?

 

[Peter]

As for section 6 and 7 of the USB 2.0 spec, yes, there are cable

specifications. It may have escaped your notice that there are also cable
specifications in 1.1. Now, would you care to list the particulars in
which a detachable cable compliant with the USB 2.0 spec will be different
from one that is in compliance with the final revision of the 1.1 spec?

I initially did not look at USB 1.1 specs. But, yes, you are right that
cable fully compliant with FINAL revision (emphasize word "final") of the
1.1 spec should be also fully compliant with USB 2.0 spec.

The difference between original and FINAL 1.1 spec are:
1) Add to the frequency range of the attenuations table (Table 7-4) to
include 200MHz <3.2db and
400MHz < 5.8db
2) Specify that the common mode impedance (Zcm) must be 30 ohms +/- 30%
3) Clarify section 7.1.16 so that one way cable delay is 5.2 ns/m (not 26ns
regardless of length)
4) Tighten cable skew (Sec 7.1.3) to 100ps from 400ps
5) DC resistance from plug shell to plug shell < .6 ohms

 

[CWatters]

I initially did not look at USB 1.1 specs. But, yes, you are right that
cable fully compliant with FINAL revision (emphasize word "final") of the
1.1 spec should be also fully compliant with USB 2.0 spec.

Yes I should have been clearer. I can't imagine why a cable manufacturer
would make a USB 1.1 only cable but they do seem to exist. I can only
assume they are sub-spec perhaps.

 

[trmbr]

I don't see any problem at 70GB on my XP only set up. My drive hovers around
75GByte all the time and I change about 20GB every night - seems to work ok.

Perhaps it's a problem with older OS?
Power supply?
Overheating?
Suspect USB cable (USB 2.0 cables are different to 1.1 only cables)

What about running a drive test program? Leave it going over a weekend to
see if you have some duff sectors.

Thanks for all the responses and related feedback - much appreciated.

The drive is ok as I've run diagnostic tests on it - all tests are ok, no
bad sectors or tracks.
The USB-to-IDE external case & cable are both USB1.1 and USB 2.0 compatible.

I've resorted to testing whether the corruption always occurs when connected
to a USB2.0 or USB1.1 port, as the PCI USB2.0 adapter (in my laptop) might
also be suspect. I assume that connecting to either a 1.1 and 2.0 USB port
should not make any difference as the cable is 1.1/2.0 compatible, the only
noticable difference will be the transfer rate. Is this correct?

Will post the results after backing up my laptop drive, 30GB, using USB1.1
port which will take "forever" in comparispon to USB2.0