Flaky Firewire connection to external hard drive in Win XP

[ggull]

I am having trouble working with moderate to large files on external hard
drives connected to a Win XP SP2 computer with Firewire. It appears that
accessing files on the external drive leads to sporadic errors. There does
not seem to be a problem writing to the external drive (see below for
evidence). What I'm looking for is some guidance as to
1) whether it is more likely a hardware or software (driver or XP) problem
2) how to further diagnose it

I initially noticed this when downloading large zip archives (about 0.7 to
1.6 GB) directly to an external drive. Trying to unzip them led to an error
in the extraction wizard. I then downloaded a sample archive (0.8GB)
directly to an internal drive, copied the zip to the external drive and had
the same problem, though on the internal drive it unzipped without error.
In an earlier thread ("Problem unzipping files on external HD",
microsoft.public.windowsxp.general), Pegasus suggested using "FC /B file1
file2" from the command prompt. That APPEARED to show that indeed the copy
to the external drive was corrupt -- at least the FC produced a list of
different bytes. .

However, I NOW BELIEVE THE WRITE TO THE EXTERNAL DRIVE IS WORKING.
WHATEVER IS WRONG IS IN THE READ PROCESS. I tried a second external drive
that has both Firewire and USB connections. I copied my test files to the
external on Firewire, and then compared them to the originals with the
Firewire connection, with the usual failures. Then I switched the
connection to USB2.0, ran the FC comparisons, and all files matched --
i.e., the copy on the external was correct, the problem just came when
trying to read it back over Firewire, either to unzip or to make the
byte-by-byte comparison. The 0.8GB zip archive successfully extracted over
the USB connection.

I ran a number of the usual sort of tests to try and isolate the problem:

DIFFERENT CABLES I used at least two different Firewire cables, Same
problems with all. ==> NOT JUST A BAD CABLE.

DIFFERENT EXTERNAL DRIVES I tried three Firewire external drives, all the
same make (Acomdata) but two substantially different models. All had the
same problems. ==> NOT JUST A FAULTY DRIVE/ENCLOSURE.

DIFFERENT PORTS I tried all three Firewire ports (2 front, 1 rear) on the
computer, with the same failure. ==> NOT JUST A BAD CONNECTOR.

DIFFERENT COMPUTER In addition to the Win XP computer, I had available a
Win Me computer. I ran the same tests with the same cable and external
drive (not all combinations, though!), tranferring the test file on a DVD-R.
This time, there were no problems over the Firewire connection. ===> Again
arguing against a problem with cable or external HD.

IS IT SOMETHING ABOUT ZIP FILES? It shouldn't make a difference, but I
tried other large files and had the same problems.

Finally, is it a problem with large files only? Is there some "threshold"
effect, that problems occur only for files larger than X bytes? Or more a
random effect, one failure in X bytes?

First, I tried a bunch of files of several types (.zip, .wav primarily), and
found that there did seem to be a fuzzy break around 100 MB -- over, you
get differences with FC, under, no differences.

Then I tried a set of 7 .WAV (audio) files from 20 to 87 MB, and had the
problem
with several. The pattern was somewhat odd -- the first 4 comparisons were
OK, the next 3 had differences. Even though these were all run separately,
it's as if the connection was getting 'tired'. I tried running the
comparisons in different order and different comparisons failed, with no
obvious pattern.

When a comparison fails, the FC output does show some patterns --
1) the different bytes occur in runs, not randomly. Even a large file with
many different bytes will only have a few runs.
2) although the start of a run seems random, the run always appears to end
on ...FFF, even ...FFFF.
3) there are stretches where the "external HD" byte (i.e. as read over
Firewire) is 00. E.g.,
0F0FFFFE: FF 00
0F0FFFFF: 0A 00
0F101D6C: FC FF

Anybody who wants to play games looking for patterns, I'd be happy to email
a typical output file, but it's large -- 536KB, or about 32K lines.

XP does seem to have some problems working with Firewire, e.g.:
http://support.microsoft.com/kb/885222/en-us
http://support.microsoft.com/kb/330174/en-us
http://support.microsoft.com/kb/885464/en-us
Although these aren't the problem I'm having (nor was anything in a KB
search for "Firewire errors" that was), it makes me wonder if this could be
a
software issue. Any ideas?
Or how can I narrow it down to software, hardware or any other suggested
diagnostics?

 

[Carey Frisch [MVP]]

Make sure your external drive is formatted NTFS
and not FAT32. Also, try downloading large,
zipped files to your primary hard drive, unzip
them and then copy the files over to your external
hard drive.

--
Carey Frisch
Microsoft MVP
Windows Shell/User

----------------------------------------------------------------------------------------------------

:

| I am having trouble working with moderate to large files on external hard
| drives connected to a Win XP SP2 computer with Firewire. It appears that
| accessing files on the external drive leads to sporadic errors. There does
| not seem to be a problem writing to the external drive (see below for
| evidence). What I'm looking for is some guidance as to
| 1) whether it is more likely a hardware or software (driver or XP) problem
| 2) how to further diagnose it
|
| I initially noticed this when downloading large zip archives (about 0.7 to
| 1.6 GB) directly to an external drive. Trying to unzip them led to an error
| in the extraction wizard. I then downloaded a sample archive (0.8GB)
| directly to an internal drive, copied the zip to the external drive and had
| the same problem, though on the internal drive it unzipped without error.
| In an earlier thread ("Problem unzipping files on external HD",
| microsoft.public.windowsxp.general), Pegasus suggested using "FC /B file1
| file2" from the command prompt. That APPEARED to show that indeed the copy
| to the external drive was corrupt -- at least the FC produced a list of
| different bytes. .
|
| However, I NOW BELIEVE THE WRITE TO THE EXTERNAL DRIVE IS WORKING.
| WHATEVER IS WRONG IS IN THE READ PROCESS. I tried a second external drive
| that has both Firewire and USB connections. I copied my test files to the
| external on Firewire, and then compared them to the originals with the
| Firewire connection, with the usual failures. Then I switched the
| connection to USB2.0, ran the FC comparisons, and all files matched --
| i.e., the copy on the external was correct, the problem just came when
| trying to read it back over Firewire, either to unzip or to make the
| byte-by-byte comparison. The 0.8GB zip archive successfully extracted over
| the USB connection.
|
| I ran a number of the usual sort of tests to try and isolate the problem:
|
| DIFFERENT CABLES I used at least two different Firewire cables, Same
| problems with all. ==> NOT JUST A BAD CABLE.
|
| DIFFERENT EXTERNAL DRIVES I tried three Firewire external drives, all the
| same make (Acomdata) but two substantially different models. All had the
| same problems. ==> NOT JUST A FAULTY DRIVE/ENCLOSURE.
|
| DIFFERENT PORTS I tried all three Firewire ports (2 front, 1 rear) on the
| computer, with the same failure. ==> NOT JUST A BAD CONNECTOR.
|
| DIFFERENT COMPUTER In addition to the Win XP computer, I had available a
| Win Me computer. I ran the same tests with the same cable and external
| drive (not all combinations, though!), tranferring the test file on a DVD-R.
| This time, there were no problems over the Firewire connection. ===> Again
| arguing against a problem with cable or external HD.
|
| IS IT SOMETHING ABOUT ZIP FILES? It shouldn't make a difference, but I
| tried other large files and had the same problems.
|
| Finally, is it a problem with large files only? Is there some "threshold"
| effect, that problems occur only for files larger than X bytes? Or more a
| random effect, one failure in X bytes?
|
| First, I tried a bunch of files of several types (.zip, .wav primarily), and
| found that there did seem to be a fuzzy break around 100 MB -- over, you
| get differences with FC, under, no differences.
|
| Then I tried a set of 7 .WAV (audio) files from 20 to 87 MB, and had the
| problem
| with several. The pattern was somewhat odd -- the first 4 comparisons were
| OK, the next 3 had differences. Even though these were all run separately,
| it's as if the connection was getting 'tired'. I tried running the
| comparisons in different order and different comparisons failed, with no
| obvious pattern.
|
| When a comparison fails, the FC output does show some patterns --
| 1) the different bytes occur in runs, not randomly. Even a large file with
| many different bytes will only have a few runs.
| 2) although the start of a run seems random, the run always appears to end
| on ...FFF, even ...FFFF.
| 3) there are stretches where the "external HD" byte (i.e. as read over
| Firewire) is 00. E.g.,
| 0F0FFFFE: FF 00
| 0F0FFFFF: 0A 00
| 0F101D6C: FC FF
|
| Anybody who wants to play games looking for patterns, I'd be happy to email
| a typical output file, but it's large -- 536KB, or about 32K lines.
|
| XP does seem to have some problems working with Firewire, e.g.:
| http://support.microsoft.com/kb/885222/en-us
| http://support.microsoft.com/kb/330174/en-us
| http://support.microsoft.com/kb/885464/en-us
| Although these aren't the problem I'm having (nor was anything in a KB
| search for "Firewire errors" that was), it makes me wonder if this could be
| a
| software issue. Any ideas?
| Or how can I narrow it down to software, hardware or any other suggested
| diagnostics?
|
|
|

 

[ggull]

Thanks for the reply.

(1) You're right, I had been using FAT32. This was at least partly to
maintain compatibility with the Windows Me machine, for file sharing. The
archives are not large enough to run into the limits of FAT32, if that's
what you had in mind (max seems to be less than the 2GB of FAT16, much less
4GB, probably by design). Other than that, should there be an issue?
Remember that on one drive, I can use either Firewire or USB, and the only
problems are when reading bytes from the external drive via Firewire, either
to run FC or unzip or simply to copy the file back to an internal drive (see
(2)). I suppose I can convert one of the external drives to NTFS as a test
if you'd suggest ... it will have to wait to later tonight or tomorrow.

(2) As to unzipping on an internal drive, then copying to the external for
storage --
I had considered this at one point, but now I believe I'd still have trouble
using the files once stored on the external. Most of the contents of the
zipped archives are compressed audio (.SHN and .FLAC) files in the range of
10 to 80 MB or so which are what I want to keep (for several reasons). I'd
need to run an experiment, but since a simple file compare failed
sporadically for files in this size range copied to the external (see about
2/3 of the way through the original post), I imagine that trying to
decompress them on the external drive would also be flaky. Indeed, I'd
forgotten to report this experiment, but simply copying back from the
external to internal corrupts the files, even "small" files (that set of 7
..WAV files but I imagine pretty much any files in this size range). Thus I
could not even copy back to the internal in order to decompress or copy to
CD/DVD for sharing, etc.

I'll run that experiment converting a drive to NTFS late tonight or early
tomorrow -- I'll be away from the computer for the evening. Be interesting
to see what happens, even though I'm pretty sure the results will be the
same.

Thanks,
gg

 

[Jonny]

I am having trouble working with moderate to large files on external hard
drives connected to a Win XP SP2 computer with Firewire. It appears that
accessing files on the external drive leads to sporadic errors. There
does
not seem to be a problem writing to the external drive (see below for
evidence). What I'm looking for is some guidance as to
1) whether it is more likely a hardware or software (driver or XP) problem
2) how to further diagnose it

I initially noticed this when downloading large zip archives (about 0.7 to
1.6 GB) directly to an external drive. Trying to unzip them led to an
error in the extraction wizard. I then downloaded a sample archive
(0.8GB) directly to an internal drive, copied the zip to the external
drive and had the same problem, though on the internal drive it unzipped
without error. In an earlier thread ("Problem unzipping files on external
HD", microsoft.public.windowsxp.general), Pegasus suggested using "FC /B
file1 file2" from the command prompt. That APPEARED to show that indeed
the copy to the external drive was corrupt -- at least the FC produced a
list of different bytes. .

However, I NOW BELIEVE THE WRITE TO THE EXTERNAL DRIVE IS WORKING.
WHATEVER IS WRONG IS IN THE READ PROCESS. I tried a second external drive
that has both Firewire and USB connections. I copied my test files to the
external on Firewire, and then compared them to the originals with the
Firewire connection, with the usual failures. Then I switched the
connection to USB2.0, ran the FC comparisons, and all files matched --
i.e., the copy on the external was correct, the problem just came when
trying to read it back over Firewire, either to unzip or to make the
byte-by-byte comparison. The 0.8GB zip archive successfully extracted
over the USB connection.

I ran a number of the usual sort of tests to try and isolate the problem:

DIFFERENT CABLES I used at least two different Firewire cables, Same
problems with all. ==> NOT JUST A BAD CABLE.

DIFFERENT EXTERNAL DRIVES I tried three Firewire external drives, all the
same make (Acomdata) but two substantially different models. All had the
same problems. ==> NOT JUST A FAULTY DRIVE/ENCLOSURE.

Open the enclosures. I'll bet you'll find a drive with a Jupiter or other
off-brand on them. These are junk. Put a more common WD or Seagate in one
and try it.

 

[ggull]

Open the enclosures. I'll bet you'll find a drive with a Jupiter or other
off-brand on them. These are junk. Put a more common WD or Seagate in
one and try it.

Thanks for the suggestion, but .... why then would the same drives work
without error on a Windows Me computer using Firewire, and on the same
WIndows XP computer when using USB (on the one enclosure that has both
connections)? Had you read my OP up to this point, or my reply to Carey
Fisher? {BTW, Carey, if you're reading this ... I'm running into some snags
converting to NTFS, which I guess we could expect if there's a bad
connection. I'll post on that later.)

It would be great to try this with another brand of drive / enclosure, but I
don't have access to one and have little desire to go out and buy another
Firewire drive just on a possibility when I strongly suspect it is something
with the Firewire implementation on the computer (hardware or software or
system) rather than the drives.

 

[Jonny]

Thanks for the suggestion, but .... why then would the same drives work
without error on a Windows Me computer using Firewire, and on the same
WIndows XP computer when using USB (on the one enclosure that has both
connections)? Had you read my OP up to this point, or my reply to Carey
Fisher? {BTW, Carey, if you're reading this ... I'm running into some
snags converting to NTFS, which I guess we could expect if there's a bad
connection. I'll post on that later.)

It would be great to try this with another brand of drive / enclosure, but
I don't have access to one and have little desire to go out and buy
another Firewire drive just on a possibility when I strongly suspect it is
something with the Firewire implementation on the computer (hardware or
software or system) rather than the drives.

I said to pull the hard drive out of one enclosure, and try a well known
brand name of hard drive like WD or Seagate. Not buy anything.

I had an Acom pure firewire enclosure that came with a hard drive. The
firewire enclosure's hard drive lasted two (2) weeks, was flaky and made odd
noises 2 days prior. Put a WD in it, and has worked well since. I use it
primarily for image file storage.
Another idea during the copies you mentioned, your source hard drive and
cable, internal RF interference, not your target is the problem.

 

[ggull]

I said to pull the hard drive out of one enclosure, and try a well known
brand name of hard drive like WD or Seagate. Not buy anything.

OK, I misspoke -- I meant not "another Firewire [external] drive" but just
"another drive". Unfortunately, I don't have any naked drives, high-grade
or plonk, sitting around, nor know of any I could borrow. So I would have
to buy whatever, or start cannibalizing parts, in order to try this.

I had an Acom pure firewire enclosure that came with a hard drive. The
firewire enclosure's hard drive lasted two (2) weeks, was flaky and made
odd noises 2 days prior. Put a WD in it, and has worked well since. I
use it primarily for image file storage.

Yikes, disappointing to hear. One of my Acomdatas, the dual-connection, is
I think about 6 months old, seen a fair amount of use with no problems --
except when connected to one particular computer with one of its two
connections (Firewire). No funny noises either. I also have an Acomdata
USB-only external [hey, they've been having some good sales] which isn't as
old but has given no problems. Unfortunately, just throwing out the
hard-drive part turns these into very expensive bare enclosures.
Incidentally, I just googled Acomdata, and Anandtech's review of the E5 (one
of the models I'm using) says that it uses a Western Digital drive, at least
in the 320GB flavor. Of course, opening one to just look would break the
seals and void the warranty.

Another idea during the copies you mentioned, your source hard drive and
cable, internal RF interference, not your target is the problem.

Again ...

-- If it is the platters / heads / circuitry of the physical drive in the
external enclosure [i.e what would be fixed by putting in a Seagate drive],
then why can I successfully copy, compare, unzip over the USB connection the
same files that fail over the Firewire?

-- If it is the above or the Firewire circuitry of the enclosure or the
Firewire cable, then why can I successfully work with these same files using
Firewire on another, Windows Me, computer?

Although I could come up with scenarios where the fault lies in the external
drive/enclosure/cable (** below), the obvious focus is the computer --
either hardware, software such as drivers, or another aspect of the problems
XP SP2 is reported to have with Firewire. I was hoping for some advice on
how to sort out those more likely possibilities.

**[for instance: the computer giving the problem has cable internet (the
other is dialup) ... maybe leading to RF interference of poorly shielded
cable/circuit? OK, I'll try with that disconnected and the cable shifted
away from the area]

 

[Guest]

Carey,

I'm having issues with my external firewire drive also. The behavior is
described in Knowledge Base Article 887170 at
http://support.microsoft.com/kb/887170/en-us. How can I get this patch so I
can try to update 1394bus.sys? Is this a valid support option versus the pay
telephone support? I can't imagine that I need to pay to fix my buggy
firewire drivers....

Regards,
Kevin Fritz

 

[jmichaelak]

Did you troubleshoot the firewire hardware? I would try a new firewire
pci card since they are cheap. I have no clue if this would cause your
problem, but you would have another base covered.

 

[ggull]

Did you troubleshoot the firewire hardware? I would try a new firewire
pci card since they are cheap. I have no clue if this would cause your
problem, but you would have another base covered.

Thanks. That's a good suggestion. I'll try it if I don't get any joy in a
couple of other directions over the next few days (too busy to run to the
computer store anyway). .

But a question -- the computer has its Firewire hardware on the motherboard
(Intel D915GAV or ...AG board w/ IEE-1394a controller and three 1394a
connectors. The manual doesn't specify any chip manufacturer). Would this
conflict with adding a pci card?

 

[jmichaelak]

I don't think so, but I'm no expert. I've added pci cards to get extra
connections and never had any problems.

 

[ggull]

I don't think so, but I'm no expert. I've added pci cards to get extra
connections and never had any problems.

Thanks. At least no intrinsic conflict between on-board and extra cards. I
want to try a couple of other things first (I finally thought of a guy who
probably has a machine fairly close to mine that I can plug an external
drive into, for instance).

 

[ggull]

Sorry for the delay, but it took a lot of fidgeting to finally get an
external drive into NTFS.
(The Convert utility had problems, even after I gave it a clean format in
FAT32 on the Win Me machine and ran a complete surface scan. Eventually,
after a few tries, I was able to Format as NTFS on the Win XP SP2 machine in
question.) I'm not surprised at the difficulties, if there is a problem
with the connection.

Anyway, clean format, NTFS, copy big.zip over from an internal drive, run
FC -- same problem.
If anything, even flakier ---
Just running FC in the Command Prompt window seemed to result in a lot of
difference bytes, so I tried capturing the output and somehow the external
drive disappeared. After cycling its power, FC now shows only a handful of
differences, here's the complete output (F: is the external drive)
----------------------
Comparing files big.zip and F:\BIG.ZIP
034BFFF4: 48 4D
034BFFF5: 0D B1
034BFFF6: 00 6B
034BFFF7: 05 AE
034BFFF8: 10 96
034BFFF9: 1A C6
034BFFFA: 00 01
034BFFFB: 05 80
034BFFFC: 0E C8
034BFFFD: 1A E9
034BFFFE: 00 95
034BFFFF: 8C 16
FC: big.zip longer than F:\BIG.ZIP
-----------------------------------
Just an oddity, in spite of the last line, in Properties both files are the
same Size, 855,496,984 bytes.

 

[Rock]

Carey,

I'm having issues with my external firewire drive also. The behavior is
described in Knowledge Base Article 887170 at
http://support.microsoft.com/kb/887170/en-us. How can I get this patch so I
can try to update 1394bus.sys? Is this a valid support option versus the pay
telephone support? I can't imagine that I need to pay to fix my buggy
firewire drivers....

Regards,
Kevin Fritz

[Courtesy of Kelly Theriot, MVP]
Call Microsoft Professional Customer Service @ 1-800-936-4900 (this line
is 24/7), there is an option for hotfix. A rep will email you a link to
the fix.