Hard disk clunk on one machine but not another

[Dale Walker]

Could there be any other reason for that special hard disk clunk (with
and occasional system reset that normally suggests a disk is on it's
way out?

A friend of mine got worried when his drive started making these
noises one a day. It increased over a period of a couple of weekd to
once an hour and then every ten minutes. Fearing the worst he gave the
disk to me to get all the data off before it completely died.

However, it's behaving perfectly on my machine. I've had the thing
running for days non-stop without a glitch. I've also done massive
series of reboots to see if I could reproduce the clunk but nothing.
On my machine the drive works fine.

So, if the drive is actually OK, what could be causing the clunking?

 

[J. Clarke]

Could there be any other reason for that special hard disk clunk (with
and occasional system reset that normally suggests a disk is on it's
way out?

A friend of mine got worried when his drive started making these
noises one a day. It increased over a period of a couple of weekd to
once an hour and then every ten minutes. Fearing the worst he gave the
disk to me to get all the data off before it completely died.

However, it's behaving perfectly on my machine. I've had the thing
running for days non-stop without a glitch. I've also done massive
series of reboots to see if I could reproduce the clunk but nothing.
On my machine the drive works fine.

So, if the drive is actually OK, what could be causing the clunking?

Most likely power problems. The fact that he has a large name brand power
supply doesn't mean that the power supply isn't defective or broken.

Heat is another, less likely possibility--are the air intakes on his machine
blocked with dust by any chance? If there's a fan for the disks, is it
working properly?

 

[Dale Walker]

Most likely power problems. The fact that he has a large name brand power
supply doesn't mean that the power supply isn't defective or broken.

Heat is another, less likely possibility--are the air intakes on his machine
blocked with dust by any chance? If there's a fan for the disks, is it
working properly?

I'd already considered both of those.

I don't think it's a power problem unless the PSU is defective as
there's nothing in the machine. Even the graphics are on the
motherboard. This is a business machine just being used for email and
the odd Word/Excel document so it's not as though it's being thrashed
by playing processor intensive games either.

As for heat, well it's not anything to do with anything being blocked
or clogged. All looks almost brand new in there. Sure, we've had a few
days of pretty warm weather but nothing that should make a PC fall
over. Processor and m'board temperatures were normal too.

So, after all that, I suppose I'd better check the PSU a bit more
thoroughly.

 

[Arno Wagner]

On Sat, 14 Aug 2004 08:55:07 -0400, "J. Clarke"

A HDD cannot trigger a system reset. There simply is no signalling
connection for it to do that. The only way I can think of is it
shortening its power connections (with associated pyrotechnics
display).

The system reset is likely caused by the real problem, e.g. a PSU with
a massively too low output voltage or drop-outs. A PSU can trigger a
system reset directly (by way of the power-good signal) and indirectly
(by delivering too low output voltage, also in short negative spikes).

I would say power. The clicking slunds like seeks that fail because
the 12V line does not deliver enough power and the voltage dops too
much during some seeks, since a seek causes a (relatively small)
current spike. After that the HDD does the 'clunk' to recover the now
mispositioned head.

I'd already considered both of those.

I don't think it's a power problem unless the PSU is defective as
there's nothing in the machine.

That is what John suggested. And I agree. Could e.g. be defective
buffer capacitors on the mains side.

Arno

 

[J. Clarke]

I'd already considered both of those.

I don't think it's a power problem unless the PSU is defective as
there's nothing in the machine. Even the graphics are on the
motherboard. This is a business machine just being used for email and
the odd Word/Excel document so it's not as though it's being thrashed
by playing processor intensive games either.

As for heat, well it's not anything to do with anything being blocked
or clogged. All looks almost brand new in there. Sure, we've had a few
days of pretty warm weather but nothing that should make a PC fall
over. Processor and m'board temperatures were normal too.

So, after all that, I suppose I'd better check the PSU a bit more
thoroughly.

I'll give you an example. Had a machine that was pretty well
stripped--motherboard with onboard video, hard disk, CD, diskette, power
supply, and case and nothing else that I can recall that periodically had
problems accessing the disk. Replaced the disk, same problems. Monitored
the voltages and found that for no apparent reason they were going low
periodically, and when that happened had problems with disk access.
Replaced the power supply and no problems since. Once I learned to look
for it I started finding it with some regularity in machines that were
having disk access problems.

 

[Arno Wagner]

Dale Walker wrote:
[...]

I'll give you an example. Had a machine that was pretty well
stripped--motherboard with onboard video, hard disk, CD, diskette, power
supply, and case and nothing else that I can recall that periodically had
problems accessing the disk. Replaced the disk, same problems. Monitored
the voltages and found that for no apparent reason they were going low
periodically, and when that happened had problems with disk access.
Replaced the power supply and no problems since. Once I learned to look
for it I started finding it with some regularity in machines that were
having disk access problems.

It seems that compared to other PC components, PSUs (even quality
ones) are pretty unreliable. Maybe some HDDs come close in
unreliability, but they have more obvious failure modes.

Arno

 

[J. Clarke]

Dale Walker wrote:
[...]

I'll give you an example. Had a machine that was pretty well
stripped--motherboard with onboard video, hard disk, CD, diskette, power
supply, and case and nothing else that I can recall that periodically had
problems accessing the disk. Replaced the disk, same problems.
Monitored the voltages and found that for no apparent reason they were
going low periodically, and when that happened had problems with disk
access.
Replaced the power supply and no problems since. Once I learned to look
for it I started finding it with some regularity in machines that were
having disk access problems.

It seems that compared to other PC components, PSUs (even quality
ones) are pretty unreliable. Maybe some HDDs come close in
unreliability, but they have more obvious failure modes.

I have a couple of Intel Cabrillo-C cases that I got on ebay for about 20
bucks once. Those are Intel's large server cases from the 440BX/GX era,
designed for a dual-Xeon (not the little P4 Xeons, the great honking P2 and
P3 Xeons) motherboard. They have three 400-watt hot-swappable redundant
power supplies. Maybe Intel knows something that we don't?

 

[Arno Wagner]

Arno Wagner wrote: [...]

It seems that compared to other PC components, PSUs (even quality
ones) are pretty unreliable. Maybe some HDDs come close in
unreliability, but they have more obvious failure modes.

I have a couple of Intel Cabrillo-C cases that I got on ebay for about 20
bucks once. Those are Intel's large server cases from the 440BX/GX era,
designed for a dual-Xeon (not the little P4 Xeons, the great honking P2 and
P3 Xeons) motherboard. They have three 400-watt hot-swappable redundant
power supplies. Maybe Intel knows something that we don't?

Maybe Intel has some customers that are willing to pay for
reliability? ;-)

I believe unreliable PSUs are not a massive problem for most
users, if defective ones are diagnosed correctly. If you buy
a quality PSU, it is more likely to be completely dead than
give you sub-standard output. At least that is my experience.

Cheap PSUs should not be used at all, IMO.

Arno

 

[J. Clarke]

Arno Wagner wrote:

Dale Walker wrote:

[...]

It seems that compared to other PC components, PSUs (even quality
ones) are pretty unreliable. Maybe some HDDs come close in
unreliability, but they have more obvious failure modes.

I have a couple of Intel Cabrillo-C cases that I got on ebay for about 20
bucks once. Those are Intel's large server cases from the 440BX/GX era,
designed for a dual-Xeon (not the little P4 Xeons, the great honking P2
and
P3 Xeons) motherboard. They have three 400-watt hot-swappable redundant
power supplies. Maybe Intel knows something that we don't?

Maybe Intel has some customers that are willing to pay for
reliability? ;-)

I believe unreliable PSUs are not a massive problem for most
users, if defective ones are diagnosed correctly. If you buy
a quality PSU, it is more likely to be completely dead than
give you sub-standard output. At least that is my experience.

My experience is that all brands of power supply can experience failures in
service that affect the regulation without completely shutting down the
power supply. Diagnosing intermittent failures of regulation can be
difficult.

 

[Roger Gelder]

Hi,
I have had something similar - a high-pitched 'squeak', like someone
putting a brake on, and then a 2 second system lock-up. I thought this was a
bad drive so replaced but the 'squeak' persists. Have since seen a comment
regarding the cpu fan, which if not true can lock. Not sure whether to
believe this but it sounded plausible, particularly after replacing a drive.
Maybe the power supply could be it but the lock-up is so short in duration
that a fan 'stop & start' feels more likely.
Regards,
RoJ
----------------------------------------------------------------------------
-------------------------------
: Arno Wagner wrote:
:
: >> Arno Wagner wrote:
: >
: >>>> Dale Walker wrote:
: >
: > [...]
: >
: >>> It seems that compared to other PC components, PSUs (even quality
: >>> ones) are pretty unreliable. Maybe some HDDs come close in
: >>> unreliability, but they have more obvious failure modes.
: >
: >> I have a couple of Intel Cabrillo-C cases that I got on ebay for about
20
: >> bucks once. Those are Intel's large server cases from the 440BX/GX
era,
: >> designed for a dual-Xeon (not the little P4 Xeons, the great honking P2
: >> and
: >> P3 Xeons) motherboard. They have three 400-watt hot-swappable
redundant
: >> power supplies. Maybe Intel knows something that we don't?
: >
: > Maybe Intel has some customers that are willing to pay for
: > reliability? ;-)
: >
: > I believe unreliable PSUs are not a massive problem for most
: > users, if defective ones are diagnosed correctly. If you buy
: > a quality PSU, it is more likely to be completely dead than
: > give you sub-standard output. At least that is my experience.
:
: My experience is that all brands of power supply can experience failures
in
: service that affect the regulation without completely shutting down the
: power supply. Diagnosing intermittent failures of regulation can be
: difficult.
:
: > Cheap PSUs should not be used at all, IMO.
: >
: > Arno
:
: --
: --John
: Reply to jclarke at ae tee tee global dot net
: (was jclarke at eye bee em dot net)

 

[Folkert Rienstra]

Could there be any other reason for that special hard disk clunk (with
and occasional system reset that normally suggests a disk is on it's
way out?

A friend of mine got worried when his drive started making these
noises one a day. It increased over a period of a couple of weekd to
once an hour and then every ten minutes. Fearing the worst he gave the
disk to me to get all the data off before it completely died.

However, it's behaving perfectly on my machine. I've had the thing
running for days non-stop without a glitch. I've also done massive
series of reboots to see if I could reproduce the clunk but nothing.
On my machine the drive works fine.

So, if the drive is actually OK, what could be causing the clunking?

CRC erors on the ide bus.
Bad power supply/supply of power.