Help hard drives keep clicking and dying

[Chris Milne]

I have an older PC (2.26 ghz 1gb ram)...here is what happened.

I bought a new Seagate 400gb HD...unplugged my 120 and 80 GB Maxtors,
plugged in the seagate, installed windows Vista (from msdn), was trying
to get my Radeon card to work with windows aero. Shut down pulled out
all the pci cards to see if that would help...no help, shut down
plugged them back in booted and the drive started clicking, then a
screech, i shut down, reboot and the drive was dead. I packaged the
new drive up to send back, plugged in the old drives and one of them
started clicking...i noticed that drive did not appear in windows, i
immediately shut down and im afraid to turn the computer back on. What
could be frying my hard drives? I had an external Maxtor USB 300 gb
drive die 4 months ago which i presume is unrelated...in my whole life
before this ive never had a hard drive go bad.

Thanks.

Other cards:
SB Audigy
UAD 1 DSP Processor
Firewire card
Radeon XT800 (agp)
USB 2 (which recently only was working at USB 1.1)

 

[Chris Milne]

Oh i forgot to mention. Most people say this is a heat related issue,
but i dont even have the side of my pc case on, so its as open as you
can get.

 

[Arno Wagner]

Oh i forgot to mention. Most people say this is a heat related issue,
but i dont even have the side of my pc case on, so its as open as you
can get.

Ok, since these drives seem to die fast, it can be one of the three
main HDD killers: Heat, mechanical shock and bad power. You have ruled
out heat. I guess you would have told us, if you dropped the drives
(even a 5cm drop on a hard surface can exceed to 200G shock
rating...). That leaves bad power.

Now, bad power does not mean that your deive(s) are actually dead.
Usually it just means that power is not stable enough or too
noisy for reliable drive operation. In these cases the drives
may behave as if defect, but are actually fine. The other
potential porblem is overvoltage or power-spikes. These can kill a
HDD. Overvoltage is easily diagnosed with the mainboards PC health
monitor, accessable via the boot-up PC bios screens. Have a look
at the levels of +5V and +12V and post them here.

Power spikes or instable power (e.g. negative spikes) require advanced
measurement equipment. If the voltage levels are fine, the next test
would be with a different, known-to-be-good PSU. If that works,
replace the PSU.

Personally, my money is on a bad PSU. These are often insanely
cheap for what they have to deliver. An industrial PSU comparable
to a typical PC one easily costs in excess of 500 EUR. No surprise
that a cheap PC PSU is usually aproblem waiting to happen...

Arno

 

[Rod Speed]

I have an older PC (2.26 ghz 1gb ram)...here is what happened.

I bought a new Seagate 400gb HD...unplugged my 120 and 80 GB
Maxtors, plugged in the seagate, installed windows Vista (from msdn),
was trying to get my Radeon card to work with windows aero. Shut
down pulled out all the pci cards to see if that would help...no help,
shut down plugged them back in booted and the drive started
clicking, then a screech, i shut down, reboot and the drive was dead.
I packaged the new drive up to send back, plugged in the old drives
and one of them started clicking...

Then the problem is much more likely to be with the system than the drive.

The best test is to see if the Seagate works in another system
and even if it doesnt, it may well be the system killing the drives.

i noticed that drive did not appear in windows, i immediately
shut down and im afraid to turn the computer back on.
What could be frying my hard drives?

You dont know that anything is yet.

I had an external Maxtor USB 300 gb drive
die 4 months ago which i presume is unrelated...

Very likely, particularly given that those are notorious for
their obscene failure rates. They basically run the drives
with inadequate cooling and Maxtor drives hate that.

in my whole life before this ive never had a hard drive go bad.

Yeah, I've only ever had one go bad myself.

Most people say this is a heat related issue, but i dont even
have the side of my pc case on, so its as open as you can get.

That wont necessarily cool the drives adequately, particularly if you
have the drives all stacked adjacent with no free slot between them.
See what Everest says about the drive temps when you have fixed
what is killing the drives if something is actually doing that.

A system can kill hard drives with a bad power supply.
Best check the drives that click in a completely different
system to see if they are actually dead. If they are, I'd
bin the power supply even if it measures correctly.

Its also possible to kill drives by handling the
system roughly with the drives spinning too.

 

[Chris Milne]

Thanks ill try your suggestions. I didnt handle the drives roughly and
they're not stacked close to each other normally..however while testing
the new seagate it was sitting on another drive (that wasnt plugged in)
for a short while.

Another thing to note is that if my computer crashes sometimes the BIOS
can't reset the time, i read in a few threads that a bad battery can
cause problems...what are your thoughts on that?

Thanks.

Chris

 

[Rod Speed]

Thanks ill try your suggestions. I didnt handle the drives
roughly and they're not stacked close to each other normally..
however while testing the new seagate it was sitting on
another drive (that wasnt plugged in) for a short while.

That shouldnt have been a problem temperature wise,
particularly as you are presumably in winter there.

Another thing to note is that if my computer crashes sometimes
the BIOS can't reset the time, i read in a few threads that a bad
battery can cause problems...what are your thoughts on that?

It wont cause drive clicking. Thats more likely to be a bad power supply.

 

[Arno Wagner]

Thanks ill try your suggestions. I didnt handle the drives roughly and
they're not stacked close to each other normally..however while testing
the new seagate it was sitting on another drive (that wasnt plugged in)
for a short while.

Sjould not be a problem...

Another thing to note is that if my computer crashes sometimes the BIOS
can't reset the time, i read in a few threads that a bad battery can
cause problems...what are your thoughts on that?

What do you mean by "reset the time"?

Arno

 

[Chris Milne]

If i get a hard reset i get the msg

CMOS/GPNV Checksum bad. If i go in to the BIOS setup the system date
is reset to 01/01/2002.

....ok i replaced the battery, i got a DMM and measured the Power supply
at 5.14V. The BIOS does not have any pc health monitoring stuff, i'm
about to hook an old drive i found in the closet to it and see what
happens.

 

[Chris Milne]

This hard drive has linux on it (its an old one)...with memtest86 so
ill get that out of the way too...its running now. Im going to put
windows on here and install intel's motherboard monitoring software.

Chris

 

[Chris Milne]

I forgot to post the 12 V setting, its at 11.65

Memtest has been running for 3 hours, no errors (89% complete). The
hard drive attached doesnt seem to have died yet. I got an external
USB 2 enclosure and attached the suspected dead drive and it did not
show up (another drive that works worked fine). The dead drive is not
clicking and is spinning, just refuses to show up in windows.

 

[Rod Speed]

I forgot to post the 12 V setting, its at 11.65

That doesnt prove much about what it might sag to etc tho.

Corse the two drives clicking might have been a coincidence too.

 

[Chris Milne]

Would be a heck of a coincidence...one is 3 to 4 years old the other
was straight out of the box, 2 hours old. Died within min's of each
other. Could have been some crazy power supply surge that corrected
itself?

Don't know...installing XP on this old HD and let it run a while before
i connect anything important, in the mean time im going to use the USB
enclosure on teh last good drive and make sure i back everything up on
dvd's.

 

[Rod Speed]

Would be a heck of a coincidence...

Sure, but once you have ruled out the other obvious possibilitys
like physical abuse, temperature and you get the same thing with a
different power supply, you have to consider that unlikely possibility.

one is 3 to 4 years old the other was straight out of the box, 2 hours old.

Yes, but the Seagate does now look like it has died
since it doesnt work in a different system, the external
box, so its almost certainly just another infant mortality.

Died within min's of each other. Could have been
some crazy power supply surge that corrected itself?

Yes, that's certainly possible. Basically the regulator goes
mad and over voltages the rails and then starts working
again. That wont necessarily kill everything powered from it.

Time will tell on that, particularly if it kills the replacement Seagate later again.

Don't know...installing XP on this old HD and let
it run a while before i connect anything important,

I'd personally replace the power supply to be safe, but then I can
afford to play safe like that and am happy to pay for that sort of
insurance when it doesnt cost the much like with a power supply.

And I have a decent max/min multimeter that I can use
to monitor for that sort of failing power supply too.

in the mean time im going to use the USB enclosure on teh last good drive

Yeah, wise move.

and make sure i back everything up on dvd's.

I always do that all the time anyway, at least
the stuff that isnt readily replaceable anyway.

I dont even do anything with the hardware unless that is up to date.

 

[Arno Wagner]

I forgot to post the 12 V setting, its at 11.65

Memtest has been running for 3 hours, no errors (89% complete). The
hard drive attached doesnt seem to have died yet. I got an external
USB 2 enclosure and attached the suspected dead drive and it did not
show up (another drive that works worked fine). The dead drive is not
clicking and is spinning, just refuses to show up in windows.

Hmm. Maybe it was a coincidence. Unlikely, but possible....

Arno

 

[w_tom]

You have just identified the problem. 12 volts is too low. Also the
5 volts being on the high side implies the power supply has hit its
limit.

Chances are the drives are not physically damaged but data may be
corrupted. Get manufacturer diagnostics. With only that one drive
attached, execute those diagnostics. This eliminates complications of
thing like Vista. With only one drive attached, first confirm all four
voltages - purple, orange, yellow, and red wire. Those numbers must
exceed 3.23, 4.87, and 11.7. Then execute the manufacturer diagnostic
to discover what is and is not working in each drive.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 - these are
not good numbers. 12 volts is woefully too low and the 5 volts
suggests the power supply power supply is not large enough for that 12
volt load.

 

[w_tom]

You have just identified the problem. 12 volts is too low. Also the
5 volts being on the high side implies the power supply has hit its
limit.

Chances are the drives are not physically damaged but data may be
corrupted. Get manufacturer diagnostics. With only that one drive
attached, execute those diagnostics. This eliminates complications of
thing like Vista. With only one drive attached, first confirm all four
voltages - purple, orange, yellow, and red wire. Those numbers must
exceed 3.23, 4.87, and 11.7. Then execute the manufacturer diagnostic
to discover what is and is not working in each drive.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 - these are
not good numbers. 12 volts is woefully too low and the 5 volts
suggests the power supply power supply is not large enough for that 12
volt load.

 

[Arno Wagner]

You have just identified the problem. 12 volts is too low. Also the
5 volts being on the high side implies the power supply has hit its
limit.

Actually it is not. 12V is a 10% tolerance voltage (new ATX standard
says 5%, but allows a peak deviation of 10%). That means 10.8V
.... 13.2V are fine. Even with 5%, 11.4V ... 12.6V is the acceptable
range.

Chances are the drives are not physically damaged but data may be
corrupted. Get manufacturer diagnostics. With only that one drive
attached, execute those diagnostics. This eliminates complications of
thing like Vista. With only one drive attached, first confirm all four
voltages - purple, orange, yellow, and red wire. Those numbers must
exceed 3.23, 4.87, and 11.7.

Definitely not. According to the ATX standard, 3.3V, 5V, 12V are all
5% tolerance voltages (with 10% on the 12V under "peak loading"
conditions). Lowest acceptable levels are 3.14V, 4.75V and 11.4V.

Then execute the manufacturer diagnostic
to discover what is and is not working in each drive.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 - these are
not good numbers. 12 volts is woefully too low and the 5 volts
suggests the power supply power supply is not large enough for that 12
volt load.

These are perfectly fine numbers. 12V is within 3% and 5V is
within 2.4% of the target voltage, i.e. far below the allowed
5% tolerance. These voltages will not cause any problem.

Arno

 

[McSpreader]

You have just identified the problem. 12 volts is too low.
Wrong.

Also the
5 volts being on the high side implies the power supply has hit
its limit.
Wrong.

Those numbers must exceed 3.23, 4.87, and 11.7.
Wrong.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 -
these are not good numbers.

Wrong.

 

[Arno Wagner]

You have just identified the problem. 12 volts is too low. Also the
5 volts being on the high side implies the power supply has hit its
limit.

Chances are the drives are not physically damaged but data may be
corrupted. Get manufacturer diagnostics. With only that one drive
attached, execute those diagnostics. This eliminates complications of
thing like Vista. With only one drive attached, first confirm all four
voltages - purple, orange, yellow, and red wire. Those numbers must
exceed 3.23, 4.87, and 11.7. Then execute the manufacturer diagnostic
to discover what is and is not working in each drive.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 - these are
not good numbers. 12 volts is woefully too low and the 5 volts
suggests the power supply power supply is not large enough for that 12
volt load.

This is nonsense. The voltages are well within the allowed tolerances,
fairly typical and are not the cause of the problem.

Please read the respective standards before posting such grossly
wrong information.

Arno

 

[Rod Speed]

You have just identified the problem. 12 volts is too low.

Wrong, the legal minimum is 11.4

Also the 5 volts being on the high side
implies the power supply has hit its limit.

Not a clue, as always.

Chances are the drives are not physically damaged but data may be
corrupted. Get manufacturer diagnostics. With only that one drive
attached, execute those diagnostics. This eliminates complications
of thing like Vista. With only one drive attached, first confirm all
four voltages - purple, orange, yellow, and red wire. Those numbers
must exceed 3.23, 4.87, and 11.7.

Wrong, as always. Its actually 3.14, 4.75, 11.4 and
the 12V rail is allowed to vary by 10% on peak load too.

Those are the numbers from the ATA standard, not your arse.

Then execute the manufacturer diagnostic to
discover what is and is not working in each drive.

Again, 12 volts at only 11.65 and 5 volts so high at 5.12 -
these are not good numbers. 12 volts is woefully too low

Not a ****ing clue, as always.

and the 5 volts suggests the power supply power
supply is not large enough for that 12 volt load.

Not a ****ing clue, as always.

And have fun explaining how come the Seagate doesnt
work in an external box that powers another drive fine too.