Sudden permanent slowdown - it's a hardware thing



Steve Campbell
07-10-2005, 12:57 AM
I booted up the morning to find that my PC had slowed to a crawl. It's
normally a very able and stable machine. Win XP/HE SP2, Athlon XP 2100+, 1
GB DDR RAM, etc

I thought it might be file corruption, so I used System Restore and then
Powerquest's Drive Image. The first time I used Drive Image (which has
always previously got me out of hot water) the system was as slow as it had
been before I used it but the second time took ages, 2.5 hours instead of 12
minutes! And bear in mind this was using the bootable CD environment i.e.
totally outside Windows.

Then when it rebooted I got STOP error 0000007B, which turns out to be a
bootable CD error. A trip to Safe Mode got me back in the next time without
doing anything but run AV and Spyware scans. These scans (repeated many
times) were all clean as a whistle. Booting then took a good 10 minutes,
slow by anyone's standards. Applications like Winamp hung and caused even
Task Manager to hang with apparent 100% cpu usage (as far as I could tell).

I checked the IDE properties and found that my main disk (250GB Maxtor
DiamondPlus 10) was registered as PIO !!! The second disk (60GB Maxtor
DiamondPlus 9) was UDMA 2. Both should have been UDMA 5. A 'quick' trawl
around newsgroups brought up the following possibility - when XP detects IO
errors it can ratchet disk performance down the modes (presumably to ensure
data integrity). This can't be reversed / reset. The experience with the
bootable disk disproves any theory that the problem might be software based.

Now here's the dilemma, is it the disks' fault? Both are slower than they
ought to be, the primary cripplingly so. Could it be a mobo fault
(controller?)? In which case would simply replacing the disks only replicate
the crisis? Do I replace both the disks and the mobo and what else besides or
am I barking up entirely the wrong tree?

Does anyone have a view on this? Your ideas are very welcome.

Steve

Ryan
07-10-2005, 12:57 AM
Reinstall drivers...

"Steve Campbell" wrote:

> I booted up the morning to find that my PC had slowed to a crawl. It's
> normally a very able and stable machine. Win XP/HE SP2, Athlon XP 2100+, 1
> GB DDR RAM, etc
>
> I thought it might be file corruption, so I used System Restore and then
> Powerquest's Drive Image. The first time I used Drive Image (which has
> always previously got me out of hot water) the system was as slow as it had
> been before I used it but the second time took ages, 2.5 hours instead of 12
> minutes! And bear in mind this was using the bootable CD environment i.e.
> totally outside Windows.
>
> Then when it rebooted I got STOP error 0000007B, which turns out to be a
> bootable CD error. A trip to Safe Mode got me back in the next time without
> doing anything but run AV and Spyware scans. These scans (repeated many
> times) were all clean as a whistle. Booting then took a good 10 minutes,
> slow by anyone's standards. Applications like Winamp hung and caused even
> Task Manager to hang with apparent 100% cpu usage (as far as I could tell).
>
> I checked the IDE properties and found that my main disk (250GB Maxtor
> DiamondPlus 10) was registered as PIO !!! The second disk (60GB Maxtor
> DiamondPlus 9) was UDMA 2. Both should have been UDMA 5. A 'quick' trawl
> around newsgroups brought up the following possibility - when XP detects IO
> errors it can ratchet disk performance down the modes (presumably to ensure
> data integrity). This can't be reversed / reset. The experience with the
> bootable disk disproves any theory that the problem might be software based.
>
> Now here's the dilemma, is it the disks' fault? Both are slower than they
> ought to be, the primary cripplingly so. Could it be a mobo fault
> (controller?)? In which case would simply replacing the disks only replicate
> the crisis? Do I replace both the disks and the mobo and what else besides or
> am I barking up entirely the wrong tree?
>
> Does anyone have a view on this? Your ideas are very welcome.
>
> Steve
>

JefN
07-10-2005, 12:57 AM
"Steve Campbell" <SteveCampbell@discussions.microsoft.com> wrote in message
news:3A968D97-A4A9-4A85-90DC-F3DB7B317426@microsoft.com...
|I booted up the morning to find that my PC had slowed to a crawl. It's
| normally a very able and stable machine. Win XP/HE SP2, Athlon XP 2100+,
1
| GB DDR RAM, etc
|
| I thought it might be file corruption, so I used System Restore and then
| Powerquest's Drive Image. The first time I used Drive Image (which has
| always previously got me out of hot water) the system was as slow as it
had
| been before I used it but the second time took ages, 2.5 hours instead of
12
| minutes! And bear in mind this was using the bootable CD environment i.e.
| totally outside Windows.
|
| Then when it rebooted I got STOP error 0000007B, which turns out to be a
| bootable CD error. A trip to Safe Mode got me back in the next time
without
| doing anything but run AV and Spyware scans. These scans (repeated many
| times) were all clean as a whistle. Booting then took a good 10 minutes,
| slow by anyone's standards. Applications like Winamp hung and caused even
| Task Manager to hang with apparent 100% cpu usage (as far as I could
tell).
|
| I checked the IDE properties and found that my main disk (250GB Maxtor
| DiamondPlus 10) was registered as PIO !!! The second disk (60GB Maxtor
| DiamondPlus 9) was UDMA 2. Both should have been UDMA 5. A 'quick' trawl
| around newsgroups brought up the following possibility - when XP detects
IO
| errors it can ratchet disk performance down the modes (presumably to
ensure
| data integrity). This can't be reversed / reset. The experience with the
| bootable disk disproves any theory that the problem might be software
based.
|
| Now here's the dilemma, is it the disks' fault? Both are slower than they
| ought to be, the primary cripplingly so. Could it be a mobo fault
| (controller?)? In which case would simply replacing the disks only
replicate
| the crisis? Do I replace both the disks and the mobo and what else besides
or
| am I barking up entirely the wrong tree?
|
| Does anyone have a view on this? Your ideas are very welcome.
|
| Steve
|

Hi Steve -

Before throwing the baby out with the bathwater...

What is the condition of your IDE cables? Are they old?

You might want to consider changing them out.

You can right-click on the controllers in Device Manager and select
Uninstall. Then reboot and Windows will redetect them, refreshing the error
counts and re-enabling full UDMA-5 support.

If you have rounded cables, be aware that many cause signal "cross talk".
High speed IDE cables (with 80 wires) have a ground wire between each signal
wire to help reduce cross talk between the signal lines. When you bunch all
the wires in the cable together, what happens to the ground wire between
each signal wire? Poorly shielded wires in rounded cables are prone to
signal cross talk.

Jef

Sharon F
07-10-2005, 12:57 AM
On Sat, 25 Jun 2005 23:15:45 GMT, JefN wrote:

> "Steve Campbell" <SteveCampbell@discussions.microsoft.com> wrote in message
> news:3A968D97-A4A9-4A85-90DC-F3DB7B317426@microsoft.com...
>> [2 quoted lines suppressed]
> 1
>> [5 quoted lines suppressed]
> had
>> [1 quoted line suppressed]
> 12
>> [5 quoted lines suppressed]
> without
>> [4 quoted lines suppressed]
> tell).
>> [5 quoted lines suppressed]
> IO
>> [1 quoted line suppressed]
> ensure
>> [2 quoted lines suppressed]
> based.
>> [4 quoted lines suppressed]
> replicate
>> [1 quoted line suppressed]
> or
>> [6 quoted lines suppressed]
>
> Hi Steve -
>
> Before throwing the baby out with the bathwater...
>
> What is the condition of your IDE cables? Are they old?
>
> You might want to consider changing them out.
>
> You can right-click on the controllers in Device Manager and select
> Uninstall. Then reboot and Windows will redetect them, refreshing the error
> counts and re-enabling full UDMA-5 support.
>
> If you have rounded cables, be aware that many cause signal "cross talk".
> High speed IDE cables (with 80 wires) have a ground wire between each signal
> wire to help reduce cross talk between the signal lines. When you bunch all
> the wires in the cable together, what happens to the ground wire between
> each signal wire? Poorly shielded wires in rounded cables are prone to
> signal cross talk.
>
> Jef

Don't tell me that. I really like my round cables. Dust carried by the
intake misses the cables and blows right out the back. Well, most of it
anyhow. ;^)

--
Sharon F
MS-MVP ~ Windows Shell/User

JefN
07-10-2005, 12:57 AM
Sharon F wrote:
|
| Don't tell me that. I really like my round cables. Dust carried by the
| intake misses the cables and blows right out the back. Well, most of it
| anyhow. ;^)
|

Hi Sharon -

Re-read my last sentence:

| Poorly shielded wires in rounded cables are prone to signal cross talk.

Properly shielded wires don't have as many problems with cross-talk (or, at
least, not causing Windows XP to throttle down the access speed due to error
thresholds). As with anything, cheaper is generally not better.

Jef


| --
| Sharon F
| MS-MVP ~ Windows Shell/User

Steve Campbell
07-10-2005, 12:57 AM
Thanks, will try this out before I do anything more drastic. The IDE cables
(flat) are a few years old, so will replace it/them. Many thanks again.

What else could cause this?

Steve

Sharon F
07-10-2005, 12:57 AM
On Sun, 26 Jun 2005 04:15:25 GMT, JefN wrote:

> Hi Sharon -
>
> Re-read my last sentence:

I did read that. Your point was one that I had never considered about round
cables. Don't know why as it makes perfect sense. Thank you for mentioning
it. I guess my surprise at this revelation is what sparked the comment.

Anyhow, when I bought my cables, I didn't even think about shielding but
did not buy the cheapest either. I must have made a good choice as so far
they have served me well.

--
Sharon F
MS-MVP ~ Windows Shell/User

JefN
07-10-2005, 12:57 AM
Steve Campbell wrote:
|
| Thanks, will try this out before I do anything more drastic. The IDE
cables
| (flat) are a few years old, so will replace it/them. Many thanks again.
|
| What else could cause this?
|
| Steve
|
|

Hi Steve -

Your best bet at this point would be to replace the IDE cables with good
quality parts, boot into Windows, go into Device Manager and right-click and
Uninstall your IDE controllers. Reboot your system and Windows will
reinstall them, resetting the error thresholds.

Monitor your drive performance for a while (by periodically checking your
controllers access mode). If, after a week or so, you stay at UDMA5, you're
good to go.

If you start throttling back again, one or both of your drives may be on
their way south or your IDE controller (either on your motherboard or a
separate add-on card) may be defective. A disk manufacturer drive fitness
utility may be a good idea in this case to thoroughly test out the drive and
it's communication with the controller. Seagate has their Seatools, Western
Digital has their Data Lifeguard, Maxtor has PowerMax, Hitachi has their
Drive Fitness Utility... (to name a few).

You've already put in place a good practice: an image backup for recovery.
Continue this practice, because no drive lasts forever.

Jef


Sudden permanent slowdown - it's a hardware thing