Can high voltage from power supply damage harddrive?

[mm]

I just talked to my friend who gave me the Dell computer, and he said
that his repair guy told him the reason his harddrive failed may
(likely?) have been a surge or something coming from the power supply.

Is this a cause of harddrive failure?

A common cause?

A conceivable cause but really a guess by the repairman.


FWIW, the repairman sent the HD somewhere to get the data off, but it
was too dead or something to do that, and now they want 1500 to 3500
dollars to do it the hard way.

FWIW, the mobo still seems good, although w/o the harddrive, it just
displays a few lines and displays a one-line message about no SATA
drive, press f2 to do this, f4 to do that.

Could I turn on the computer and connect a voltmeter to the hardrive
power connector** and watch the needle for a few hours (while I do
other things), or do I have to watch for weeks to get a good idea?

**Or do another connector on the same power supply, since the SATA
power connector looks very small?

Thanks.

 

[Yousuf Khan]

I just talked to my friend who gave me the Dell computer, and he said
that his repair guy told him the reason his harddrive failed may
(likely?) have been a surge or something coming from the power supply.

Is this a cause of harddrive failure?

That's possible, but a more likely cause is the exact opposite
situation, where there wasn't enough power coming through the power
supply at the right moment.

A common cause?

A conceivable cause but really a guess by the repairman.

It is a guess, plain and simple.

FWIW, the repairman sent the HD somewhere to get the data off, but it
was too dead or something to do that, and now they want 1500 to 3500
dollars to do it the hard way.

Yes, what he's describing is that the onboard electronics of the drive
are fried. But the data might still be on the mechanical portion of the
hard drive, so a circuit board replacement should do the trick to get
the data again. However, the circuit board replacement costs thousands
of dollars, as he's saying.

FWIW, the mobo still seems good, although w/o the harddrive, it just
displays a few lines and displays a one-line message about no SATA
drive, press f2 to do this, f4 to do that.

Could I turn on the computer and connect a voltmeter to the hardrive
power connector** and watch the needle for a few hours (while I do
other things), or do I have to watch for weeks to get a good idea?

**Or do another connector on the same power supply, since the SATA
power connector looks very small?

I assume that the repairman already may have tried putting the SATA
drive into another system. But you could try it yourself and put it into
your own system to see if it works there.

Yousuf Khan

 

[Arno]

I just talked to my friend who gave me the Dell computer, and he said
that his repair guy told him the reason his harddrive failed may
(likely?) have been a surge or something coming from the power supply.

Is this a cause of harddrive failure?

Yes, but only with a cheap PSU that does not have overvoltage
and surge protection. There, basically anything is possible,
including the PSU killing a drive/computer without a mains surge,
just by running amok.

A common cause?

Cheap PSU: Moderately, especially with above-ground mains
lines and a lightening strike.

Quality PSU: Very, very unlikely. Quality PSUs are known to
occasionally fail on mains surges, but do not produce voltage
spikes on the computer side. They also will switch themselves
off reliably before killing components if their regulator circuitry
fails.

Dell has been known to use cheaper compontnts, but not that
cheap IMO.

Also take into account that a HDD has about a 5% per year (!)
failure rate without any outside influence.

A conceivable cause but really a guess by the repairman.

Indeed. And if that happens, it typically takes the PSU with it.

FWIW, the repairman sent the HD somewhere to get the data off, but it
was too dead or something to do that, and now they want 1500 to 3500
dollars to do it the hard way.

Reasonable. But there are a lot of disreputable data-recovery
outfits out there, "somewhere" is not enough to advise on whether
they are crooks or not.

FWIW, the mobo still seems good, although w/o the harddrive, it just
displays a few lines and displays a one-line message about no SATA
drive, press f2 to do this, f4 to do that.

Entirely reasonable.

Could I turn on the computer and connect a voltmeter to the hardrive
power connector** and watch the needle for a few hours (while I do
other things), or do I have to watch for weeks to get a good idea?

You would need to use a PSU analyzer set-up to be sure. Expensive
and needs considerable knowledge to operate.

I would check the +5V lines (red) and +12V lines (yellow) and
if they have good voltages (+/- 3% of nominal or so) go for it.
If you are afraid of killing more drives, get a new PSU from a
reputable manufaturer (I recommend Enermax, by far the best
quality). But be carefult, DELL cases sometimes do not take
standard PSUs.

**Or do another connector on the same power supply, since the SATA
power connector looks very small?

Other connector on the same wires is best. Direct measurements
on SATA power connectors are not a good idea, to easy to shorten
somethig out.

Arno

 

[Arno]

That's possible, but a more likely cause is the exact opposite
situation, where there wasn't enough power coming through the power
supply at the right moment.

That would not kill a drive. All modern HDDs have voltage
sensors and will just shut down if given too low voltages.

Arno

 

[JD]

Yes, but only with a cheap PSU that does not have overvoltage
and surge protection. There, basically anything is possible,
including the PSU killing a drive/computer without a mains surge,
just by running amok.


Cheap PSU: Moderately, especially with above-ground mains
lines and a lightening strike.

Quality PSU: Very, very unlikely. Quality PSUs are known to
occasionally fail on mains surges, but do not produce voltage
spikes on the computer side. They also will switch themselves
off reliably before killing components if their regulator circuitry
fails.

Dell has been known to use cheaper compontnts, but not that
cheap IMO.

Also take into account that a HDD has about a 5% per year (!)
failure rate without any outside influence.


Indeed. And if that happens, it typically takes the PSU with it.


Reasonable. But there are a lot of disreputable data-recovery
outfits out there, "somewhere" is not enough to advise on whether
they are crooks or not.


Entirely reasonable.


You would need to use a PSU analyzer set-up to be sure. Expensive
and needs considerable knowledge to operate.

I would check the +5V lines (red) and +12V lines (yellow) and
if they have good voltages (+/- 3% of nominal or so) go for it.
If you are afraid of killing more drives, get a new PSU from a
reputable manufaturer (I recommend Enermax, by far the best
quality). But be carefult, DELL cases sometimes do not take
standard PSUs.


Other connector on the same wires is best. Direct measurements
on SATA power connectors are not a good idea, to easy to shorten
somethig out.

Arno

Was a surge protector used for this computer?

If not, the user was a fool. Electric supplies are
always oscillating up and down and
need taming by the surge protector.

 

[mm]

Yes, but only with a cheap PSU that does not have overvoltage
and surge protection. There, basically anything is possible,
including the PSU killing a drive/computer without a mains surge,
just by running amok.


Cheap PSU: Moderately, especially with above-ground mains
lines and a lightening strike.

New York City, East 57th St. No telephone poles for miles. Everything
from the power substation up to the building is below ground.
Everything to the power substation also.

11th story of his apartment building -- Is that good or bad? -- but
the mains are in the center of the building, near the elevators --
Does that help? (His building has the same elevator motors and
relays that were put in when the building was new, around 1935 iirc.
They are beautiful.)

Quality PSU: Very, very unlikely. Quality PSUs are known to
occasionally fail on mains surges, but do not produce voltage
spikes on the computer side. They also will switch themselves
off reliably before killing components if their regulator circuitry
fails.

Dell has been known to use cheaper compontnts, but not that
cheap IMO.

Also take into account that a HDD has about a 5% per year (!)
failure rate without any outside influence.


Indeed. And if that happens, it typically takes the PSU with it.

Well, it still works so far, but it has little load, only the mobo and
the video card. Oh, I said that.

Reasonable. But there are a lot of disreputable data-recovery
outfits out there, "somewhere" is not enough to advise on whether
they are crooks or not.

He's trusting his repairman. Where he got him I don't know, but
because his wife insisted on living in NY instead of Brooklyn, he
lives in an expensive building. Of course everything on the East
Side north of 23rd or 42nd is expensive, and I don't know if he got
the referral from his neighbors (who have more money than he does) and
of course rich people get ripped off too.

Entirely reasonable.


You would need to use a PSU analyzer set-up to be sure. Expensive
and needs considerable knowledge to operate.

Well, forget that.

I would check the +5V lines (red) and +12V lines (yellow) and
if they have good voltages (+/- 3% of nominal or so) go for it.
If you are afraid of killing more drives, get a new PSU from a
reputable manufaturer (I recommend Enermax, by far the best
quality). But be carefult, DELL cases sometimes do not take
standard PSUs.

Ugh. But I'll do my best to measure. And I'll see what Dell sells,
or what is sold in Dell's name, for this model if possible.

Other connector on the same wires is best. Direct measurements
on SATA power connectors are not a good idea, to easy to shorten
somethig out.

Yeah. I think there are other connectors on the same wire.

 

[mm]

That's possible, but a more likely cause is the exact opposite
situation, where there wasn't enough power coming through the power
supply at the right moment.


It is a guess, plain and simple.


Yes, what he's describing is that the onboard electronics of the drive
are fried. But the data might still be on the mechanical portion of the
hard drive, so a circuit board replacement should do the trick to get
the data again. However, the circuit board replacement costs thousands
of dollars, as he's saying.

So, if I could get an identical drive, I could replace the circuit
board myself??? I have 40 years part-time experience in electronic
repairs including soldering.

FTR I don't have the drive, but I could get it from my friend.

I assume that the repairman already may have tried putting the SATA
drive into another system. But you could try it yourself and put it into
your own system to see if it works there.

Here, I just mean that to test the power supply, I could watch the
voltage on a voltmeter, with a needle, to see if it varies.

 

[Yousuf Khan]

That would not kill a drive. All modern HDDs have voltage
sensors and will just shut down if given too low voltages.

Arno

I'm pretty certain a lack of power was what killed one of my old 500GB
drives a few years back. Up until that time, it was working perfectly,
there were no SMART errors, etc. Then one day, I added another hard
drive into the system, and this one just stopped functioning
simultaneously. The only plausible answer is that the PS wasn't putting
enough power out to run all of the drives at the same time.

Yousuf Khan

 

[DevilsPGD]

Was a surge protector used for this computer?

If not, the user was a fool. Electric supplies are
always oscillating up and down and
need taming by the surge protector.

Surge protectors really aren't needed for modern computers, the PSU
should have far more capable surge protection built-in as part of it's
design.

More often than not a surge protector won't catch a surge in time
anyway, although it will usually do a decent enough job of preventing a
fire if some connected device shorts out.

 

[DevilsPGD]

In message <[email protected]> mm

So, if I could get an identical drive, I could replace the circuit
board myself??? I have 40 years part-time experience in electronic
repairs including soldering.

Maybe. Until recently this was usually possible if you could get an
absolutely identical drive, not only the model number but the software
version too in some cases.

However, modern drives are apparently storing some data on the
controller circuitry itself that makes this type of swap out more
difficult.

Still, if the data is worth the gamble, and if it's a bit of an older
drive you might find replacements on eBay cheap, so it's worth trying.

 

[Rod Speed]

I just talked to my friend who gave me the Dell computer, and he said
that his repair guy told him the reason his harddrive failed may (likely?)
have been a surge or something coming from the power supply.

Is this a cause of harddrive failure?

Yes, it can happen.

A common cause?
Nope.

A conceivable cause but really a guess by the repairman.

Depends on whether he worked out that the rails to the hard drive had
been over voltaged like with some visible damage on the drive logic card.

FWIW, the repairman sent the HD somewhere to get the data off, but it
was too dead or something to do that, and now they want 1500 to 3500
dollars to do it the hard way.

FWIW, the mobo still seems good, although w/o the harddrive, it just
displays a few lines and displays a one-line message about no SATA
drive, press f2 to do this, f4 to do that.

Could I turn on the computer and connect a voltmeter to the hardrive power
connector** and watch the needle for a few hours (while I do other things),

Yes, but that may not prove anything, some power
supply glitches can be too short term to see on a meter.

or do I have to watch for weeks to get a good idea?

You really need to have a more fancy multimeter that records the peaks seen on the rails.

Even then it doesnt prove the power supply isnt faulty,
it can happen only very rarely, with a mains surge etc.

**Or do another connector on the same power supply,
since the SATA power connector looks very small?

Yes, its fine to use one of the other molex power connectors thats easier to connect the meter probe to.

 

[Rod Speed]

So, if I could get an identical drive, I could replace the circuit board myself???

Yes. But the recovery operation should have tried that.

I have 40 years part-time experience in electronic repairs including soldering.

FTR I don't have the drive, but I could get it from my friend.

Here, I just mean that to test the power supply, I could watch
the voltage on a voltmeter, with a needle, to see if it varies.

That wont see the shortest spikes and you wont be watching it when the mains surges either.

 

[Yousuf Khan]

So, if I could get an identical drive, I could replace the circuit
board myself??? I have 40 years part-time experience in electronic
repairs including soldering.

Yes, that's conceivable. But remember if it doesn't work, you got two
dead drives instead.

Here, I just mean that to test the power supply, I could watch the
voltage on a voltmeter, with a needle, to see if it varies.

Not worth the effort, the voltage will vary, without much doubt. But
what you can learn from watching those variations is dubious. Whatever
caused that drive to fail has long since passed by.

Yousuf Khan

 

[Franc Zabkar]

I just talked to my friend who gave me the Dell computer, and he said
that his repair guy told him the reason his harddrive failed may
(likely?) have been a surge or something coming from the power supply.

Is this a cause of harddrive failure?

A common cause?

Yes, it is a very common cause, especially in external enclosures
where people apply a 19V - 20V laptop power adapter in place of the
original 12V supply.

In most cases there is an easy no-cost DIY fix. It involves using a
pair of flush cutters to remove a shorted TVS (transient voltage
suppression) diode.

These articles should help you identify the various components:

HDD from inside Main parts:
http://hddscan.com/doc/HDD_from_inside.html

WD Pc easy Faultdiagnostic troubleshooting must try it ...
http://forum.hddguru.com/easy-faultdiagnostic-troubleshooting-must-try-t12319.html

My notes may also help:
http://www.users.on.net/~fzabkar/HDD/HDD_ICs.txt
http://www.users.on.net/~fzabkar/HDD/TVS_diodes.txt

If you upload a detailed photo of the component side of the drive's
PCB, maybe I could help you locate the faulty component. That said, if
the drive spins up, then the PCB is probably OK. The fault will
instead be inside the HDA.

If you wish to replace the TVS diode, then you can use an SMBJ12A in
place of the 12V diode, and an SMAJ5.0A in place of the 5V diode. Both
parts are available from Farnell, Mouser, Digikey.

Be aware that early model drives are not protected by TVS diodes.
Also, when such a diode is present, there will sometimes be a fuse
(Samsung), or zero-ohm resistor (WD), or inductor (Seagate), or
polyswitch (Hitachi) in series with the supply. These often go
open-circuit and will need to be bridged with a short piece of wire.
Alternatively, you could use a 2A smt fuse:

Littelfuse Surface Mount Fuses, N = 2A, S = 4A:
http://www.farnell.com/datasheets/48294.pdf

- Franc Zabkar

 

[Franc Zabkar]

So, if I could get an identical drive, I could replace the circuit
board myself??? I have 40 years part-time experience in electronic
repairs including soldering.

In modern drives, you will need to transplant the 8-pin serial EEPROM
chip, or its contents, from patient to donor. This chip stores unique,
drive specific calibration data. Identical drives manufactured in the
same factory on the same day will have different ROM contents. In
particular, there are "adaptive" data for each head.

The following article explains why drives have adaptive data.

HDD from inside: Tracks and Zones. How hard it can be?
http://hddscan.com/doc/HDD_Tracks_and_Zones.html

Essentially the reason is that no two heads are physically identical.
HD manufacturers try to fit as much data as possible onto each
platter. To do this, they take advantage of any head that performs
better than the average.

For example, some heads will have a better frequency response than
others, which means that you can cram more bits on each track. This
technique is called Variable Bits Per Inch (VBPI).

Giant magnetorestive (GMR) heads use a separate element for writing
and another for reading. The separation between these two components
varies from head to head. Once again manufacturers optimise data
density by implementing Variable Tracks Per Inch (VTPI).

Each HD therefore needs to be calibrated to account for VBPI and VTPI,
otherwise the drive doesn't know where to find the tracks, or the data
within the track. When a drive powers up, it needs to retrieve the
bulk of its firmware from a reserved area (System Area) on the
platters. If it can't read these data, then it clicks. This is what
usually happens after you swap a board without transferring the
calibration information.

That said, you may be lucky, and the tolerances between patient and
donor may be close enough to allow you to access your data without a
ROM transfer.

- Franc Zabkar

 

[Arno]

I'm pretty certain a lack of power was what killed one of my old 500GB
drives a few years back. Up until that time, it was working perfectly,
there were no SMART errors, etc. Then one day, I added another hard
drive into the system, and this one just stopped functioning
simultaneously. The only plausible answer is that the PS wasn't putting
enough power out to run all of the drives at the same time.

Well, temprary failure is of course possible, but if will simply
refuse to work and stay undamaged. There are other explanations
for a drive dying when adding another. Basically mechanical damage,
the first spin-down after along time and it was not able to spin-up
for a long time, electrostatic damage, etc..

Arno

 

[Arno]

Was a surge protector used for this computer?

If not, the user was a fool. Electric supplies are
always oscillating up and down and
need taming by the surge protector.

That is BS. First, a surge protector would not help here at all.
Second, changes in AC frequency are harmless for most equipment
including computers. You clock may be off by a bit. But at
least in Europe these variances are tiny.

What a surge protector is for is to catch severe overvoltage
spikes induced by ligthening stikes close to or into above
ground mains lines. These go into the > 1000V range
and can not only kill PSUs, they can also induce overvoltage
on the low-volt side in badly designed PSUs. It is the task
of the surge protector to limit the voltge to something
like 400V (235V AV) or 200V (115V AC), which a PSU can
usualy withstand for some miliseconds (at least if reasonable
quality).

Note that above-ground mains lines are a sign of an
underdeveloped or very old grid and are not in use anymore
in modern installations. With below ground mains, the
surge risk is very, very small.

Arno

 

[Arno]

In message <[email protected]> JD <[email protected]> was
claimed to have wrote:

Surge protectors really aren't needed for modern computers, the PSU
should have far more capable surge protection built-in as part of it's
design.

Ok, sorry, have to go into lecture mode here...


Depends on your situation. If you have outdated above-ground mains
lines, a high-powered surge-protector is a good idea. The one in
the PSU will protect the PSU, but likely die in the process.
The problem here is that they are typically too small to survive
a surge and typically are not thermally moniroted. So there is
a fuse in addition (whcih you should _not_ replace, unless
you replace the varistor that does the surge prtection as well!),
that blows on a surge. The thing is that varistoirs start to leak
current after surges and start to heat up. This can cause fires
and make them ineffective as protection.

If you have a good quality external surge protector (i.e. one with
thermal and functional monitoring), it will also die after a few
surges, but it will
a) signal this and
b) interupt mains voltage via thermal fuses
c) have replacable normal fuses
d) have varistors that can take several surges before
starting to leak enough to need replacement

You can then replace the surge protector instead of the PSU
or even only the varistor module in some models.

More often than not a surge protector won't catch a surge in time
anyway,

Both gas-arc type and varistor type surge protectioon has
basically no reaction time. They do have a minimal activation
voltage and lower quelity electronics may get damaged before
that voltage is reached.

although it will usually do a decent enough job of preventing a
fire if some connected device shorts out.

That is what fues are for. A surge protector does not hekp here.

Arno

 

[Arno]

Yes, but only with a cheap PSU that does not have overvoltage
and surge protection. There, basically anything is possible,
including the PSU killing a drive/computer without a mains surge,
just by running amok.


Cheap PSU: Moderately, especially with above-ground mains
lines and a lightening strike.

[/QUOTE]

New York City, East 57th St. No telephone poles for miles. Everything
from the power substation up to the building is below ground.
Everything to the power substation also.

Good. Very low surge risk there.

11th story of his apartment building -- Is that good or bad? -- but
the mains are in the center of the building, near the elevators --
Does that help? (His building has the same elevator motors and
relays that were put in when the building was new, around 1935 iirc.
They are beautiful.)

Should not matter at all. The building has lightening protection
which also protects the power lines within. I imagine that in
New York these are inspected regularly, as you reliably get
building fires on lightening stikes if they do not work.

Sode note: Yes, classic more expensive electical installation
ccan be quite beautiful and work for a long, long time.

Well, it still works so far, but it has little load, only the mobo and
the video card. Oh, I said that.

Then I would rule out the PSU as cause. More likely the disk
died by itself.

He's trusting his repairman. Where he got him I don't know, but
because his wife insisted on living in NY instead of Brooklyn, he
lives in an expensive building. Of course everything on the East
Side north of 23rd or 42nd is expensive, and I don't know if he got
the referral from his neighbors (who have more money than he does) and
of course rich people get ripped off too.

Trust does not imply competence om the target. Sorry. Although
the price-quote is in range of hwat you would get form a
reputable company like Kroll-Ontrak.

Well, forget that.

Indeed ;-)

Ugh. But I'll do my best to measure. And I'll see what Dell sells,
or what is sold in Dell's name, for this model if possible.

Yeah. I think there are other connectors on the same wire.

There should be.

Arno

 

[Arno]

On 8/28/2010 11:03 PM PT, Rod Speed typed:

It has happened to me back in summer of 2005 with an old ASUS A7V333
motherboard/mobo. (lock ups and no more boot ups) and Quantum Fireball
Plus 30 GB IDE HDD (circuit controller died). They were damaged by a
recent PSU melt down. The mobo's circuit board had brown/orange coloring
on its circuit board, a horrible burning odor/smell, etc. The old HDD
also went dead too. I never bothered to recover that drive's datas
since I had a backup.

It is possible, but it requires a) the PSU regulator failing into
overvoltage and b) no output-overvoltage protection. Both failing
is very rare, hovever cheap PSUs often come without the protection
circuitry. Reason: Price. Let me give you an idea (say 400W):

- ElCheapo : $20
- Mid-range : $80
- Enermax with all protection: $120
- Industrial with 10 year lifetime: $300-$500

Now, at the lower end, using cheaper components alone does not cut
it anymore, you have to use a lot less components. That means
cutting out protection circuirty, using cheaper, less reliable
regulators, cheaper transformers with lower insulation rating,
etc.. Also your ElCheapo 400W unit is more likely a 250-300W unit,
while a 400W Enermax is more likely 500-700W.

Arno