Can high voltage from power supply damage harddrive?

[mm]

A more plausible answer is static electric discharge from your
hand. You would not even know you generated that discharge. Many
other reasons also exist.

All disk drives see 'lack of power'. That is the message that tells
every drive (even long before the PC existed) that power has been
turned off. Disk drives work normally. Or if power (voltage) is too
low, the disk drive says, "Oh, this is a power off", and then does the
normal shutdown procedure.

Accurately noted by others. A power supply (with standard and
required functions) will not destroy a disk drive. Many computer
techs without electrical knowledge will simply blame mythical surges –
because that is what most people are told to blame.

mm - it is possible to obtain an identical drive from a used
component house. As others noted, your firmware revision on that new
drive must be identical. Since it costs tens of dollars to buy one,
well, I have had good experiences recovering data this way. But
again, be careful that firmware revisions are identical for drives
with same model number.

Thanks. I'm working on that now. It seems the number on the MCU chip
matters too for Seagate. He's going to send that to me tonight.

If you don't already know it, torx drives are often necessary to
swap the boards.

I have those. Now I can finally use them. ( I've used them a
couple times.)

 

[Arno]

On Sun, 29 Aug 2010 16:52:23 +1000, Franc Zabkar

My friend took it to a friend of his wife, who does data restoration
for a living, according to what I'm told.

He said the drive was "blank", but that there might be some data on it
that the 1000 to 3500 dollar repair could retrieve.

That is BS. Either it is blank or it has data on it.

I forgot to ask
the details of the drive failure, but I'm guessing he tried to start
the computer o;ne day and it wouldn't start, or he was using it and it
couldn't find a file, so he turned ir off and it woudln't restart.

How could it be blank? Even if the MBR is ruined, it's not blank.
Indeed.

How could he tell without a lot of work that it is blank, or that the
MBR or something needed for the whole drive is ruined?

To tell it iis blank is simple. But if it really is blank (i.e.
filled with zero data), then there is no way to restore anything.

Maybe he said it "appeared blank". Why would he say that when he
knows it's full of stuff.
Incomptence?

He must have hooked it up. How can he tell "blank" from bad
electronics?

Blank: Spins up, gets detected, but e.g. a hd /dev/<disk? in Linux
shows only zero data.

Bad electronics: Does not spin up or does not get detected.
No way to tell enything about the data in this case without repair.

The guy didn't charge him anything, so either he did little work, he
doesn't charge when he fails, or he gave my friend a bargain because
he is friends with my friend's wife.

Or he did not charge anything because he has no real clue?

More below.

Tnanks for all of these.

Thank you. I don't have it yet, and haven't even asked for it, until
I'm a little farther along on this process.

That sounds bad, because the circuit board is all that I can replace,
or fiddle with, right?

Right without clean room equipment (moderate level).

If a surge caused the TVS diode to short, I should be able to see that
with on ohmmeter, right? And if it's still good I should be able to
see that.
Yes.

This one was sold inside the Dell in January of 2004. It's date code
is 04231, which I'm figuring is the Julian date and about August 15,
2004. Young enough to certainly have TVS diodes?

No certainity here. Post a picture of the PCB and we can tell you.

Seriously, get that disk away from the "guy", he sounds like
all he can do is run data recovery software without even
understanding what that does.

Arno

 

[mm]

I guess I'm wrong about this because another similar Seagate disk in
ebay had a date of 04385. Unless they mean January 20th of 2005.
Maybe they have 730 12-hour periods, or 1460 6-hour periods....

No certainity here. Post a picture of the PCB and we can tell you.

Seriously, get that disk away from the "guy", he sounds like

The guy hasn't got it anymore. My friend does, and he didn't want to
send it to me until I'm sure i'm going to spend some time and effort
on it. I'll write him about sending me a scan or photocopy.

 

[Yousuf Khan]

A more plausible answer is static electric discharge from your
hand. You would not even know you generated that discharge. Many
other reasons also exist.

Maybe, but this system was highly loaded at the time, possibly verging
on overloaded. I had a 400W PSU running 4 internal HDs, a couple of
optical drives, several USB peripherals, and a video card at that time.
Plus it stayed on 24/7 most of the time. Running it through a PSU
calculator revealed that that was the limit for that PSU. I've since
upgraded to a 600W PSU.

If it was electrostatic discharge, then why that particular hard drive,
and not any of the other three internal drives?

I'm just saying that since my PSU was known to be on the edge of
overload, that's the more likely cause of the problem.

Yousuf Khan

 

[Franc Zabkar]

My friend took it to a friend of his wife, who does data restoration
for a living, according to what I'm told.

He said the drive was "blank", but that there might be some data on it
that the 1000 to 3500 dollar repair could retrieve.

Usually, when the price is that high, it means that there is media or
head damage.

How could it be blank? Even if the MBR is ruined, it's not blank.
How could he tell without a lot of work that it is blank, or that the
MBR or something needed for the whole drive is ruined?

Maybe he said it "appeared blank". Why would he say that when he
knows it's full of stuff.

He must have hooked it up. How can he tell "blank" from bad
electronics?

When a drive powers up, the MCU loads its own internal (masked)
bootstrap code. This code then fetches more code from the serial
EEPROM, plus the "adaptive" data. These adaptives enable the MCU to
locate and fetch the bulk of the drive's firmware from a hidden System
Area (SA) on the platters. These firmware modules would include the
logical-to-physical sector translator, defect list, ATA commands,
SMART data, etc. If the modules cannot be read, then the capacity is
often reported as 0GB, meaning that the drive has powered up in safe
mode. Some drives may also identify themselves using their factory
alias rather than their model number.

If your friend's drive spins up and identifies itself with its correct
model number and correct capacity, then it will most likely have bad
sectors or weak heads, or a corrupted file system. In the latter case,
you would use data recovery software to repair the logical damage.
However, in cases of bad sectors, you would be best to clone as much
of the drive as possible using multipass cloning software such as
ddrescue or dd_rescue. Ddrescue knows how to skip over bad patches in
the media. It can also clone a drive in reverse, effectively disabling
read lookahead caching. Ddrescue clones the "easy" sectors on the
first pass, and then tries for the more difficult ones on subsequent
passes.

Avoid software such as Spinrite or HDD Regenerator. These will
repeatedly hammer a bad sector, potentially accelerating the failure
of a weak head.

I would not use CHKDSK to repair your file system in cases where there
are bad sectors. Instead run it in read-only mode.

The guy didn't charge him anything, so either he did little work, he
doesn't charge when he fails, ...

Many data recovery companies operate on a no data, no fee basis.

HDA?

Head / Disc Assembly

That sounds bad, because the circuit board is all that I can replace,
or fiddle with, right?

You would at least need a special cleanroom, otherwise any dust would
contaminate the drive. Even then, a head stack replacement is not
something I would try. I would advise you not to open your drive. See
the following article for an inside view.

Head Stack Replacement: Questions and Answers:
http://hddguru.com/content/en/articles/2006.02.17-Changing-headstack-Q-and-A/

If a surge caused the TVS diode to short, I should be able to see that
with on ohmmeter, right? And if it's still good I should be able to
see that.

Yes. However, if your board looks like the photo in my other posting,
then a shorted diode would prevent the power supply from starting. You
would see the fans kick once just before the supply shuts down.

This one was sold inside the Dell in January of 2004. It's date code
is 04231 ...

Seagate uses a strange coding.

Seagate date code calculator:
http://www.bugaco.com/calculators/seagate_date_code.php

"04231 corresponds to 2003 December 6th".

- Franc Zabkar

 

[westom]

I'm just saying that since my PSU was known to be on the edge of
overload, that's the more likely cause of the problem.

Overloaded power supplies are not damaged. Supplies simply
shutdown. In fact, Intel demands a maximum load - a short circuit -
be applied to every supply without damage. Intel's requirement even
says how thick the wire must be to short out all outputs from each
supply. And no supply must be damaged. As was true with power
supplies long before the PC even existed.

 

[westom]

I get it. I'll keep this in mind.
I'm looking for a PS that will fit this Dellcomputer. I found some
prospects last night but then I got tired.

If any power supply outputs 20 volts on it 5 volt output, then a
human is responsible for buying a defective supply. No power supply
must ever damage the electronics. Others posters said that and said
why. But then so many computer assemblers buy supplies only on
dollars and watts, then foreign suppliers have discovered a market
ripe for scamming. A market full of electrically naive customers.

No minimally acceptable supply - long before the IBM PC existed, can
increase any output voltage to destructive levels. But to sell a
supply at $20 less money with higher profits, these 'dumped' supplies
are missing essential functions.

 

[Rod Speed]

Yousuf Khan wrote

westom wrote
Unlikely.

Nope.

Maybe, but this system was highly loaded at the time, possibly verging
on overloaded. I had a 400W PSU running 4 internal HDs, a couple of
optical drives, several USB peripherals, and a video card at that
time. Plus it stayed on 24/7 most of the time. Running it through a
PSU calculator revealed that that was the limit for that PSU.

That doesnt prove much, those calculators dont really tell you much about what is needed power supply wise.

I've since upgraded to a 600W PSU.

If it was electrostatic discharge, then why that particular hard drive, and not any of the other three internal
drives?

One might be more sensitive to static.

I'm just saying that since my PSU was known to be on the edge of overload, that's the more likely cause of the
problem.

Its not actually that likely, but its certainly possible that you had a supply that
doesnt react well to being overloaded and over voltages the rails in that situation.

 

[Rod Speed]

westom wrote

Overloaded power supplies are not damaged. Supplies simply shutdown.

Thats what is supposed to happen, but doesnt always with bad designs.

In fact, Intel demands a maximum load - a short
circuit - be applied to every supply without damage.

Doesnt mean that the worst designed power supplys comply with that requirement.

Intel's requirement even says how thick the wire
must be to short out all outputs from each supply.

Doesnt mean that the worst designed power supplys comply with that requirement.

And no supply must be damaged. As was true with
power supplies long before the PC even existed.

Doesnt mean that the worst designed power supplys comply with that requirement.

The Intel spec also says that no mains surges should ever damage anything
downstream of the power supply too. In the real world that does happen anyway.

The Intel spec also says that no power supply failure should ever damage anything
downstream of the power supply too. In the real world that does happen anyway.

 

[Arno]

Overloaded power supplies are not damaged. Supplies simply
shutdown. In fact, Intel demands a maximum load - a short circuit -
be applied to every supply without damage. Intel's requirement even
says how thick the wire must be to short out all outputs from each
supply. And no supply must be damaged. As was true with power
supplies long before the PC even existed.

So far for theory. In practice ElCheapo designers could not care
less what Intel demands and the typical cheap PSU (and some more
expensive ones) will not survive a short circuit and may even
die when operated at 100% load for more than a few minutes.

Arno

 

[Arno]

On 8/30/2010 12:56 PM PT, Arno typed:

Cool. I might get another SeaSonic PSU for my future PSUs. That is
assuming they are still good. 2/2006 is over four years ago.

Antec Basiq 500 Watt ATX Power Supply (BP500U) [started making fan
noises and I only had it since 1/23/2010 -- ugh!]).

Antec is prettied-up ElCheapo with a high price tag. I had
one server-grade EPS from them fail. Opened it and it turned
out to be an insufficiently cooled component, in a configuration
that I can only call either "incompetent" or "designed to fail".
The thing was running at about 40% permenant load and about 80%
load on startup (lots of disks). In addition the overall
impression was that of very cheap build design and component
selection.

I will not buy Antec again.

Great. So what are good brand of these days that isn't so pricey these
days for home users?

You get what you pay for. I cannot really recommend
any manufaturer except Enermax. It seems they are
also the only ones that do design and manufacturing
all in-house.

Arno

 

[Arno]

On 8/31/2010 6:49 AM PT, Arno typed:

Hmm, prices vary a lot on
http://www.newegg.com/Product/Produ...der=BESTMATCH&Description=Enermax+psu&x=0&y=0
... I hope I don't need 900 watts any time soon. Yes, I have lots of
hardwares and play computer games (those darn video cards [even single
card] suck a lot!).

Hehe, no, you are unlike to need 900W unless you have
3-4 graphics cards or two very, very power hungry ones.
You can to a rough calculation on what you need as follows:

- Mainboard: 20W
- CPU: TDP + 20% (regulator-loss)
- Add-in cards: each 10W
- Graphics card: Sometimes given by manufactuer, otherwise google
for test results
- Disks 3.5": each 15W
- Disks 2.5", also SSD: each 5W

As an example, my system would be
- Mainboard 20W
- Phenom II X4 945 95W TDP => 114W
- Network card: 10W
- Radeon 4850: about 250W (http://techreport.com/articles.x/14967/10)
- 2 x 3.5" disks: 30W
- 1 SSD : 5W

Altogether: about 430W. The 500W Enermax PSU in the system
is entirely fine and very rarely goes to full fan speed.
Low activity load (desktop) gives a mains-side load of 225W,
which comes down to about 180W on the computer side, stioulating
80% efficientcy.

Note: If you do this with a lower quelity PSU, add at least 20%
safety magin. Enermax PSU routinely display built-in margins
of 25-30%, so you do not need it there. For some other quality
PSUs you may get similar margins, but do not assume it without
conclusive evidence, e.g. a load test from a reputable source.

Arno

 

[mm]

On 8/31/2010 6:49 AM PT, Arno typed:

Hmm, prices vary a lot on
http://www.newegg.com/Product/Produ...der=BESTMATCH&Description=Enermax+psu&x=0&y=0
... I hope I don't need 900 watts any time soon. Yes, I have lots of
hardwares and play computer games (those darn video cards [even single
card] suck a lot!).

Hehe, no, you are unlike to need 900W unless you have
3-4 graphics cards or two very, very power hungry ones.

I have a video card that also has a tv output. I think I can turn the
tv output off except on the rare occasions I use it.

Do you think it would use less power if it were off? More than a
watt?

I ask because during the really hot weather, 95+, my cpu and sometimes
my mobo get very hot and I even may have to turn off the computer.

Also, the video card seems to be failing. About 4 times in the last 3
months the picture has been illegible, full of specks, with a pink
ghost image of the text, and I've had to turn the computer off for
that too.

Do you think turning off the tv output make a difference?

(Every extra browser tab raises the temp a little, although closing
Eudora doesn't seem to make a difference.)

 

[Simon Brown]

On 8/30/2010 12:56 PM PT, Arno typed:

Cool. I might get another SeaSonic PSU for my future PSUs. That is
assuming they are still good. 2/2006 is over four years ago.

Antec Basiq 500 Watt ATX Power Supply (BP500U) [started making fan
noises and I only had it since 1/23/2010 -- ugh!]).

Antec is prettied-up ElCheapo with a high price tag. I had
one server-grade EPS from them fail. Opened it and it turned
out to be an insufficiently cooled component, in a configuration
that I can only call either "incompetent" or "designed to fail".
The thing was running at about 40% permenant load and about 80%
load on startup (lots of disks). In addition the overall
impression was that of very cheap build design and component
selection.

I will not buy Antec again.

Great. So what are good brand of these days that isn't so pricey
these days for home users?

You get what you pay for. I cannot really recommend
any manufaturer except Enermax. It seems they are
also the only ones that do design and manufacturing
all in-house.

Arno

That is a completely silly claim.

 

[Rod Speed]

mm wrote

Great. So what are good brand of these days that isn't so pricey
these days for home users?
You get what you pay for. I cannot really recommend
any manufaturer except Enermax. It seems they are
also the only ones that do design and manufacturing
all in-house.
Hmm, prices vary a lot on
http://www.newegg.com/Product/Produ...der=BESTMATCH&Description=Enermax+psu&x=0&y=0
... I hope I don't need 900 watts any time soon. Yes, I have lots of
hardwares and play computer games (those darn video cards [even
single card] suck a lot!).

Hehe, no, you are unlike to need 900W unless you have
3-4 graphics cards or two very, very power hungry ones.

I have a video card that also has a tv output. I think I can turn the
tv output off except on the rare occasions I use it.

Do you think it would use less power if it were off?
Nope.

More than a watt?
Nope.

I ask because during the really hot weather, 95+, my cpu and sometimes
my mobo get very hot and I even may have to turn off the computer.

Also, the video card seems to be failing. About 4 times in the last 3
months the picture has been illegible, full of specks, with a pink ghost
image of the text, and I've had to turn the computer off for that too.

Do you think turning off the tv output make a difference?
Nope.

(Every extra browser tab raises the temp a little,

Thats mad.

 

[mm]

Thats mad.

But true. Before it got hot out, (and also in the house, since my AC
is broken), I might have 60 or 70 tabs open with no problem. When it
was 98 out and maybe 86 or 88 in the house, 10 tabs would be enough to
overheat the CPU**. If I closed the browser the temperature went
down.


**Well, I wasn't sure what "too hot" was, but if it went from a normal
146F up to 164, I closed Firefox or turned off the whole computer.

 

[Rod Speed]

mm wrote

But true.

With what browser ?

Before it got hot out, (and also in the house, since my AC is broken),
I might have 60 or 70 tabs open with no problem.

Fark, why do you have so many ? I find it VERY hard to believe that with
say 60 tabs open, you can actually see a temperature rise with 61 open.

When it was 98 out and maybe 86 or 88 in the house, 10 tabs would be enough
to overheat the CPU**. If I closed the browser the temperature went down.

That last is nothing like the first claim.

**Well, I wasn't sure what "too hot" was, but if it went from a normal
146F up to 164, I closed Firefox or turned off the whole computer.

Firefox can be rather resource hungry.

 

[westom]

So far for theory. In practice ElCheapo designers could not care
less what Intel demands and the typical cheap PSU (and some more
expensive ones) will not survive a short circuit and may even
die when operated at 100% load for more than a few minutes.

True. Which is why fools buy computers assembled by the
electrically naive. Which is why fools buy supplies only on dollars
and watts. Why so many buy 500 watt supplies that are electrically
equivalent to a 350 watt supply in a brand name computer.

The informed computer assembler knows what many required power
supply functions are so that computers are never damaged by the
supply. And so that the load never damages a supply. Some examples of
what so many Certified A+ computer techs never learn:
Acoustics noise 25.8dBA typical at 70w, 30cm
Short circuit protection on all outputs
Over voltage protection
Over power protection
100% hi-pot test
100% burn in, high temperature cycled on/off
PFC harmonics compliance: EN61000-3-2 + A1 + A2
EMI/RFI compliance: CE, CISPR22 & FCC part 15 class B
Safety compliance: VDE, TUV, D, N, S, Fi, UL, C-UL & CB
Hold up time, full load: 16ms. typical
Dielectric withstand, input to frame/ground: 1800VAC, 1sec.
Dielectric withstand, input to output: 1800VAC, 1sec.
Ripple/noise: 1%
MTBF, full load @ 25°C amb.: >100k hrs

Power supplies that would cause disk drive failure are typically
missing these and many other functions. Power supplies even 40 years
ago were not damaged by the load. All power supplies from the
monster sized to single chip supplies were not damaged by the load
even 40 years ago. It was industry standard even that long ago.

Supply failure is traceable to a computer assembler without
electrical knowledge. Who buys a supply only on dollars and watts. A
problem becoming severe in North America. Basic electrical knowledge
has become so minimal that 60% of new Silicon Valley employees come
from India or China. Finding enough Americans with basic electrical
knowledge has become that difficult.

The load must never damage a supply. And a supply must never damage
its load (ie disk drive). Overloading must never damage a supply.
But does when the supply is missing essential functions - when sold on
price, When the consumer does always demand technical specs.

 

[westom]

Do you think turning off the tv output make a difference?

Power consumers are the maybe 80 computers on a graphics card. That
TV output is a trivial power consumer.

Its not so much about power. Its mostly about a sudden current
demand. For example, the original Intel Pentium would demand from
less than one amp to tens of amps in microseconds. That sudden power
demand is critical - and why video cards need that dedicated power
connection.

All computers work perfectly ideal inside a 100 degree F room.
High temperature does not harm hardware. High temperature causes
signal timing and voltage threshold changes. Also known as a
computer crash.

Routine was to operate Intel CPUs even at 140 degrees C - by simply
slowing the processor. Yes, hotter than boiling water. Those
temperatures do not damage semiconductors. It causes timing changes
that cause computer crashes.

How to find defective hardware that is working today but will fail
under warranty months or years later? Operate it in a 100 degree F
room. Also execute comprehensive hardware diagnostics. A computer
that works normally at 70 degrees F and that is defective can be
located by simply operating at the normal operation temperature of 100
degrees F. How to find defects before a warranty expires. What we
always do with hardware that can never fail.

 

[Rod Speed]

This clown has never had a clue about anything at all, ever.