Harddrive runs at half speed

[berend]

Hi,

I have two IDE harddrives, both are about 5 years old. Now one of them
seems to run at a noticeably slower speed than the other, here are
some specs:

=========================
OS: WinXP sp2
=========================
disk0, pri. master
40GB
5400rpm
Cache: 1.8 MB
2 partitions, C and D, both NTFS
Avg Access time: 27 msec
Avg read speed: 12 MB/sec <== this sucks!
Ultra DMA-mode 6
=========================
disk1, pri. slave
20GB
5400rpm
Cache: 2.0 MB
1 partition, E, NTFS
Avg Access time: 22 msec
Avg read speed: 28 MB/sec <== Could be faster but hey, it's old
Ultra DMA-mode 5
=========================

I've been through the usual options such as defragging, but that
doesn't seem to help. Both drives have about 50% free space.
Does anyone have an idea where to go next?

Regards, Berend

 

[Paul]

Hi,

I have two IDE harddrives, both are about 5 years old. Now one of them
seems to run at a noticeably slower speed than the other, here are
some specs:

=========================
OS: WinXP sp2
=========================
disk0, pri. master
40GB
5400rpm
Cache: 1.8 MB
2 partitions, C and D, both NTFS
Avg Access time: 27 msec
Avg read speed: 12 MB/sec <== this sucks!
Ultra DMA-mode 6
=========================
disk1, pri. slave
20GB
5400rpm
Cache: 2.0 MB
1 partition, E, NTFS
Avg Access time: 22 msec
Avg read speed: 28 MB/sec <== Could be faster but hey, it's old
Ultra DMA-mode 5
=========================

I've been through the usual options such as defragging, but that
doesn't seem to help. Both drives have about 50% free space.
Does anyone have an idea where to go next?

Regards, Berend

Start by running this benchmark.

http://www.simplisoftware.com/Public/index.php?request=HdTach

The shape of the curve, and the noise in the curve, can suggest
a few things. For example, I have a couple Seagate drives, and
the one that is a couple years old had a jagged detailed curve,
while the new Seagate drive with only a few hours of test time
on it, gave a very smooth graph. It could be, that reallocated
sectors cause response time issues for the older drive.

Also, the shape of the curve tells you something about whether
the disk is media or cable limited. If the line is curved, that
is a media limitation. My older 7200 RPM drive, is about 60MB/sec
at the beginning of the disk, and 40MB/sec near the end of the
disk. The curve is "tilted", and that tells you that the IDE
cable is not a limitation.

If the transfer curve is a flat horizontal line, then the
interface is running at a reduced rate. If there are a lot
of CRC errors detected, Windows will drop down to PIO transfer
mode, at 4MB/sec. A person suffering from PIO polled transfers,
would see a flat line at 4MB/sec approximately.

So run the benchmark, and see what the results tell you.

Paul

 

[Gerry]

Try running HD Tune(freeware).

Download and run it and see what it turns up.
http://www.hdtune.com/

Select the Info tabs and place the cursor on the drive under Drive
letter and then double click the two page icon ( copy to Clipboard )
and copy into a further message.

Select the Health tab and then double click the two page icon ( copy
to Clipboard ) and copy into a further message. Also do a full surface
scan with HD Tune.

--



Hope this helps.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~

 

[Bob Willard]

Start by running this benchmark.

http://www.simplisoftware.com/Public/index.php?request=HdTach

The shape of the curve, and the noise in the curve, can suggest
a few things. For example, I have a couple Seagate drives, and
the one that is a couple years old had a jagged detailed curve,
while the new Seagate drive with only a few hours of test time
on it, gave a very smooth graph. It could be, that reallocated
sectors cause response time issues for the older drive.

Also, the shape of the curve tells you something about whether
the disk is media or cable limited. If the line is curved, that
is a media limitation. My older 7200 RPM drive, is about 60MB/sec
at the beginning of the disk, and 40MB/sec near the end of the
disk. The curve is "tilted", and that tells you that the IDE
cable is not a limitation.

If the transfer curve is a flat horizontal line, then the
interface is running at a reduced rate. If there are a lot
of CRC errors detected, Windows will drop down to PIO transfer
mode, at 4MB/sec. A person suffering from PIO polled transfers,
would see a flat line at 4MB/sec approximately.

So run the benchmark, and see what the results tell you.

Paul

Right. For a non-antique HD, the curve should be a staircase:

 

[Berend Veldkamp]

Start by running this benchmark.

http://www.simplisoftware.com/Public/index.php?request=HdTach

The shape of the curve, and the noise in the curve, can suggest
a few things. For example, I have a couple Seagate drives, and
the one that is a couple years old had a jagged detailed curve,
while the new Seagate drive with only a few hours of test time
on it, gave a very smooth graph. It could be, that reallocated
sectors cause response time issues for the older drive.

Also, the shape of the curve tells you something about whether
the disk is media or cable limited. If the line is curved, that
is a media limitation. My older 7200 RPM drive, is about 60MB/sec
at the beginning of the disk, and 40MB/sec near the end of the
disk. The curve is "tilted", and that tells you that the IDE
cable is not a limitation.

If the transfer curve is a flat horizontal line, then the
interface is running at a reduced rate. If there are a lot
of CRC errors detected, Windows will drop down to PIO transfer
mode, at 4MB/sec. A person suffering from PIO polled transfers,
would see a flat line at 4MB/sec approximately.

So run the benchmark, and see what the results tell you.

Paul

Paul, thanks for your detailed answer. It gave me a better understanding
of what's going on. I put the results here, I you have time to take a
look: http://members.tele2.nl/b.veldkamp/hd/hdtach.png

The fast drive (blue) has a curved graph with just a few spikes, and
slowing down towards the end of the drive. The slow one (red) stays more
or less on the same level, but has a very jagged graph. If I understand
correctly, this could suggest a cable limitation. However they are on
the same cable. Hm, maybe I'll try to replace it anyway and see what
will happen.

What exactly is a reallocated sector, is that where a sector goes bad,
and Windows marks it as such? I just ran chkdsk, but it gave me exactly
0 errors. Is there a way to fix reallocated sectors?

Thanks in advance, Berend

 

[Berend Veldkamp]

Gerry,

Thanks, this more or less confirms what I found using the hdtach tool. I
compared my results to the ones found on the hdtune website, and it
seems that my 40Gig drive is running at normal speed, the other one
should be about 4x faster.

Here are the results from the info and health tabs for the slow drive.
It fails on "Seek error rate", but I'm not sure how to interpret those
values.

======================
HD Tune: MAXTOR 6L040J2 Information

Firmware version : A93.0500
Serial number : 662202221736
Capacity : 37.3 GB (~40.0 GB)
Buffer size : 1863168 bytes
Standard : ATA/ATAPI-5
Supported mode : UDMA Mode 6 (Ultra ATA/133)
Current mode : UDMA Mode 2 (Ultra ATA/33)

S.M.A.R.T : yes
48-bit Address : no
Read Look-Ahead : yes
Write Cache : yes
Host Protected Area : yes
Device Configuration Overlay : yes
Automatic Acoustic Management: yes
Power Management : yes
Advanced Power Management : no
Power-up in Standby : no
Security Mode : yes
Firmware Upgradable : yes

Partition : 1
Drive letter : C:\
Label : windowsxp
Capacity : 20481 MB
Usage : 59.86%
Type : NTFS
Bootable : Yes

Partition : 2
Drive letter : D:\
Label : mirror
Capacity : 14841 MB
Usage : 40.31%
Type : NTFS
Bootable : No

Partition : 3
Drive letter :
Label :
Capacity : 2847 MB
Usage : 0.00%
Type : unknown (F2h)
Bootable : No
======================
HD Tune: MAXTOR 6L040J2 Health

ID Current Worst ThresholdData
Status
(01) Raw Read Error Rate 100 253 20 0
(03) Spin Up Time 80 80 20 2506
(04) Start/Stop Count 97 97 8 2131
(05) Reallocated Sector Count 100 100 20 0
(07) Seek Error Rate 62 1 23 5
Failed
(09) Power On Hours Count 98 98 1 1871
(0A) Spin Retry Count 100 100 0 0
(0B) Calibration Retry Count 90 60 20 1
(0C) Power Cycle Count 97 97 8 2120
(0D) Soft Read Error Rate 100 100 23 0
(C2) Temperature 83 76 42 44

(C3) Hardware ECC Recovered 16 1 0 -128810360
(C4) Reallocated Event Count 100 100 20 0
(C5) Current Pending Sector 100 100 20 0
(C6) Offline Uncorrectable 100 253 0 0
(C7) Ultra DMA CRC Error Count 200 200 0 0
Power On Time : 1871
Health Status : Failed

 

[Berend Veldkamp]

Please ignore the values for "Current mode", I switched the cable back
to an 80pin, and now they're equal to "Supported mode". My original
issue remains, unfortunately.

Berend

 

[Gerry]

http://www.ariolic.com/activesmart/smart-attributes/seek-error-rate.html

--



Hope this helps.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~

 

[Berend Veldkamp]

http://www.ariolic.com/activesmart/smart-attributes/seek-error-rate.html

Gerry,
I found that page by myself too. But what's the error rate compared to?
Is it the percentage of total seek actions, in that case 100 would mean
"all seeks are erroneous". Or is it errors/sec, per hour, ever? And why
is the "worst" value 1?

Is it time for a new HD?

 

[Gerry]

Berend

You need to get readings above the threshold not to fail. The threshold
for your drive is 23. My two Seagate drives have a threshold of 30 and
when I just checked both showed worst figures of 60 and are not
failing. Check the readings over 12 hours to confirm that your figures
are not a blip.

Changing a hard drive is much more manageable if it is before rather
than after failure. I appreciate it means a new installation of Windows
and programmes unless you are using cloning software. I would see it as
an opportunity to increase hard disk capacity for a modest investment.
You can get 40 gb drives for £20 but it is likely to be more economic to
go for something bigger.

--



Hope this helps.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~

 

[Berend Veldkamp]

Berend

You need to get readings above the threshold not to fail. The threshold
for your drive is 23. My two Seagate drives have a threshold of 30 and
when I just checked both showed worst figures of 60 and are not
failing. Check the readings over 12 hours to confirm that your figures
are not a blip.

Changing a hard drive is much more manageable if it is before rather
than after failure. I appreciate it means a new installation of Windows
and programmes unless you are using cloning software. I would see it as
an opportunity to increase hard disk capacity for a modest investment.
You can get 40 gb drives for £20 but it is likely to be more economic to
go for something bigger.

Ok, so higher values are better, that makes a bit more sense ;-) Just
one more thing though: I ran chkdsk, and a hdtune Error-scan, but both
didn't report any errors. Still the worst Seek Error Rate is 1. Which
time period do the smart params aply to? The entire lifespan of the
disk, or from boot?

I've just been through a major clean-up because the drive(s) kept
filling up, so I was already considering buying a new one. Reinstalling
Windows is a pain, but you can also get rid of the 'dust' that was
gathered over the years. Nice thing to do over Christmas I guess.

Thanks again for your help,
Berend

 

[Paul]

Paul, thanks for your detailed answer. It gave me a better understanding
of what's going on. I put the results here, I you have time to take a
look: http://members.tele2.nl/b.veldkamp/hd/hdtach.png

The fast drive (blue) has a curved graph with just a few spikes, and
slowing down towards the end of the drive. The slow one (red) stays more
or less on the same level, but has a very jagged graph. If I understand
correctly, this could suggest a cable limitation. However they are on
the same cable. Hm, maybe I'll try to replace it anyway and see what
will happen.

What exactly is a reallocated sector, is that where a sector goes bad,
and Windows marks it as such? I just ran chkdsk, but it gave me exactly
0 errors. Is there a way to fix reallocated sectors?

Thanks in advance, Berend

I just tried Gerry's suggestion of HDTune.

On the Info tab, there is this on my hard drive

Supported UDMA Mode 5 (Ultra ATA/100)
Active UDMA Mode 5 (Ultra ATA/100)

If the two fields match, the drive is doing the best
that it can. If the "Active" value is slower than
the "Supported" value, then that can be improved.

An 80 wire IDE cable helps. It has better signal
quality, and helps with the UDMA modes.

The following article addresses resetting the transfer mode
being used in Windows. (See "Workaround" near bottom of page)

"IDE ATA and ATAPI disks use PIO mode after multiple time-out or CRC errors occur"
http://support.microsoft.com/default.aspx?scid=kb;en-us;817472

Hope that helps,
Paul

 

[Gerry]

Berernd

http://en.wikipedia.org/wiki/Self-Monitoring,_Analysis,_and_Reporting_Technology

http://file-monitor.com/extensions/hdd-smart.php

HD. Tune almost certainly relies on S.M.A.R.T. This can be enabled /
disabled through the BIOS. I cannot ascertain whether the S.M.A.R.T.
data is collected for the lifetime of the disk, for when S.M.A.R.T. is
enabled or for each boot. Turning S.M.A.R.T. off, rebooting and then
renabling might suggest an answer.

Chkdsk does not have any predictive capability. Running the Disk
Manufacturers Test Utility might give an answer. You could also contact
Seagate (who now own Maxtor ) for advice.

--



Hope this helps.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~

 

[Berend Veldkamp]

Berernd

http://en.wikipedia.org/wiki/Self-Monitoring,_Analysis,_and_Reporting_Technology

http://file-monitor.com/extensions/hdd-smart.php

HD. Tune almost certainly relies on S.M.A.R.T. This can be enabled /
disabled through the BIOS. I cannot ascertain whether the S.M.A.R.T.
data is collected for the lifetime of the disk, for when S.M.A.R.T. is
enabled or for each boot. Turning S.M.A.R.T. off, rebooting and then
renabling might suggest an answer.

Chkdsk does not have any predictive capability. Running the Disk
Manufacturers Test Utility might give an answer. You could also contact
Seagate (who now own Maxtor ) for advice.

I'm no further, It seemed smart was already disabled in the bios. I
enabled it, then disabled it, and finally enabled it again, but hdtune
gives me the same information. Apparently the system can still access
smart data, although it's disabled in the bios.

Oh well. I guess I'll buy a new drive anyway, and maybe keep the old one
to see how long it'll last.

Berend

 

[Gerry]

Berend

I have enquiries out to try to find out more about how S.M.A.R.T.
updates but not too many people seem to know too much about that aspect.

Please let me know when you get out your cornet and blow the last post!


--
Regards.

Gerry
~~~~
FCA
Stourport, England
Enquire, plan and execute
~~~~~~~~~~~~~~~~~~~