How to recover a crashed laptop hard disk (windows NTFS)

[Franklin]

This was in the first post:

I did find a good video from PC World at
http://www.pcworld.com/video/catid,1610-page,1/video.html
titled "How to Resurrect a Crashed Hard Drive - PC World Video".

This video has LOTs of suggestions which were not covered in this
thread yet.

That video maye misleading because there is a far better approach for
fixing certain IBM hard drives. If you have one of the old Deskstar
drives then check out the following repair guide.

http://web.ukonline.co.uk/morgan.tate/

 

[Zak Hipp]

I don't understand this one. Doesn't the device draw the amps it needs
regardless of the supply's maximum output rating? I can see it making
sense if the original power supply was overloaded/overrated and not
really delivering, but the video implies that the power supply "pushes"
the current rather than allowing the device to draw it.

Watts = Amps * Volts (W=IV)

If a maximum rating of a supply is 100 Watts and the desired electron pressure to be maintained is 10 Volts then a
maximum of 10 Amps can be drawn. Therefore if 15 Amps is drawn then the electron pressure must fall to 6.667 Volts. If
10 volts is required for correct functioning of a circuit drawing 15 Amps then a power supply of 150 watts is required
(W/I=V). I have no idea if this sort of thing works but I understand the reasoning. A component may have stepped outside
the desired specification, say a resistor with a lowered value, allowing a higher current to flow and exceeding the
ability of the power supply to maintain correct voltage levels.

Zak Hipp

 

[Larry Sabo]

Nope, the supply may not be able to supply enough current at the rated voltage.

Exactly.

Motors are notorious for drawing more current at start-up than once
started, i.e. surge current. If the *maximum* supply current was
insufficient, because more devices have been added to the system than
the original supply was rated for, a stuck drive could tax it beyond
its ability to deliver, i.e. it wouldn't produce the start-up torque
required to overcome the stiction.

Larry

 

[bluerhinoceros]

That video maye misleading because there is a far better approach for
fixing certain IBM hard drives. If you have one of the old Deskstar
drives then check out the following repair guide.

http://web.ukonline.co.uk/morgan.tate/

My hammer is missing one side of its claw. Will it still work?

 

[Bear Bottoms]

My hammer is missing one side of its claw. Will it still work?

Depends on how deeply the nail is buried.

 

[Jumpster Jiver]

I had already realized it was spinning because when I pressed on the top of
the hard disk drive, I could hear a whirr as I made something touch
something else.

Most laptop hard drives have a label that says DO NOT PRESS HERE!

A hard drive depends on a 100% dust free interior to work correctly.
By causing the cover to rub on the spindle that holds the disc in place,
even for a moment, you have likely ground off a few tiny metal particles.
These particles will probably end up between the heads and the disc.
This will scratch the disc and damage the heads.
That hard drive is no longer reliable.
If you get the system up and runnig you should retrieve all the data you
can off that drive then destroy it (to prevent someone else from
stealing your information.)

 

[Arno Wagner]

Motors are notorious for drawing more current at start-up than once
started, i.e. surge current. If the *maximum* supply current was
insufficient, because more devices have been added to the system than
the original supply was rated for, a stuck drive could tax it beyond
its ability to deliver, i.e. it wouldn't produce the start-up torque
required to overcome the stiction.

True, but have a look into a HDD manual for the startup current.
The highest rating I have seen is 2.5A at 12V. That is the maximum
rating and it gets limited by the motor controller. There will be
no spike above it. This is not a DC motor directly connected
to the 12V line, where there is a huge spike at startup.

Arno

 

[Arno Wagner]

Watts = Amps * Volts (W=IV)

If a maximum rating of a supply is 100 Watts and the desired
electron pressure to be maintained is 10 Volts then a maximum of 10
Amps can be drawn. Therefore if 15 Amps is drawn then the electron
pressure must fall to 6.667 Volts.

Nice try. This would be for a wattage-regulated output. This is
a voltage regulated one, i.e. there is active circuity that
does maintain the elecron pressure at all cost. It will deliver
constant voltage, until the a second circuit becomes active,
namely the overcurrent protection. Then the voltage will drop
to zero and you typically have to power-cycle it. The way this
works is that the PSY can actually deliver much more power for
a short time (miliseconds), but would destroy itself by overheating.

If 10 volts is required for
correct functioning of a circuit drawing 15 Amps then a power supply
of 150 watts is required (W/I=V).
Indeed.

I have no idea if this sort of
thing works but I understand the reasoning. A component may have
stepped outside the desired specification, say a resistor with a
lowered value, allowing a higher current to flow and exceeding the
ability of the power supply to maintain correct voltage levels.

Not how it works except in very rare failure modes, were the PSU has
lost enough of its rated power that it cannot trigger the overcurrent
protection at its rated output voltege anymore. Or if it does not have
overcurrent protection in the first place. A very badly designed
and dangerous (fire hazard) PSU might not have it.

Arno

 

[Morten Skarstad]

Erica Eshoo skrev:

If the drive spins up ...
- Start the PC with a boot disk & back up data.
- Ultimate Boot CD for Windows v3.05 (6/23/2007)
- http://www.UBCD4Win.com
- http://find.pcworld.com/57857

No. Use a Linux boot CD.

I know it sounds weird, but I have tried this myself when a relative put
her hopes into me being able to recover the data from a laptop with a
broken hard drive. The file system was NTFS, so my first thought was
that a "proper" Windows based solution was the way to go. I was wrong.

My first attempt was done using BartPE (UBCD4Win is based on BartPE).
First, hardware support was pitiful. Neither of the two NICs were
recognized, so I had nowhere to put the recovered data. Second, whenever
it hit a bad sector it ground down to a seeming halt trying to read
unreadable data, then it threw out numerous error messages and then it
aborted. The hardware support-issue may be improved with UBCD4Win with
or without some sort of drivers-plugin (I haven't tried them, but I
believe they are available), but I don't see any way for a product based
on Windows very own system files to be able to circumvent the problems
inherent in Windows itself.

The solution turned out to be Knoppix. NICs were detected just fine, so
I could mount a share on another computer to dump the data. And physical
errors were handled both faster and more gracefully; Bad files were
skipped, and the rest were copied without user intervention. The entire
operation was a piece of cake, and loss of data was neglible.