Are the connections to hard drive electronics soldered.

  • Thread starter Thread starter mm
  • Start date Start date
M

mm

Earlier, I think we were talking about changing the electronics on a
hard drive. A video I watched said that only the torx screws have to
be removed (and I guess a connector unplugged) and thus implied that
no desoldering was necessary.

Is that true?

I know how to desolder and solder, but if there is no soldering, that
would make it a lot easier. I need to know how much to encourage my
buddy to let me try this.

In this case it's a Seagate ST3010026AS hd. 120G, SATA.

Thanks.
 
mm said:
Earlier, I think we were talking about changing the electronics on a
hard drive. A video I watched said that only the torx screws have to
be removed (and I guess a connector unplugged) and thus implied that
no desoldering was necessary.
Is that true?
Yes.

I know how to desolder and solder, but if there is no soldering,
that would make it a lot easier. I need to know how much to
encourage my buddy to let me try this.

You can void the warranty.
 
mm said:
Great. I guess I assumed it was soldered, and I wrote about it to
someone I thought knew and he didnt' contradict me, or agree.


The warranty is long gone. :)

Thanks a lot.

So what "electronics" are you going to change? Obviously you need the
IDE interface logic board which also contains the controller circuits.
Since those get tuned to the particular drive along with burning in the
table of masked hardware-level bad spots on the platters, just what did
you think you would accomplish by replacing the PCB on the hard disk?

If you get a replacement PCB that is the same exact model and used for
the same exact model of hard disk assembly then the switcheroo *might*
work. I've done this in the past. It's worked maybe around 1 times out
of 6. That was back when I had access to LOTS of dead drives from which
to steal parts and had the time and inclination to waste on trying to
salvage an old drive because, say, one of the surface-soldered regulator
diodes blew up for the motor.

After many years, I figure if the files aren't backed up then they
weren't important and are either okay to lose or are reproducible. The
value of your time to salvage a drive exceeds the cost of getting a new
drive. If it's the data on the platters you're after and you don't have
backups (rolls eyes) then good luck replacing the PCB and expecting it
to work as-is.
 
So what "electronics" are you going to change? Obviously you need the
IDE interface logic board which also contains the controller circuits.
Since those get tuned to the particular drive along with burning in the
table of masked hardware-level bad spots on the platters, just what did
you think you would accomplish by replacing the PCB on the hard disk?

I used to think that all the firmware, calibration data, and defect
lists were stored on the PCB. However, both the factory defects
(P-list) and grown defects (G-list), and the bulk of the firmware, are
actually stored on a reserved area of the platters. The calibration
data, however, are nowadays stored in a serial EEPROM. When swapping
boards, you need to transfer the ROM or its contents from patient to
donor.
If you get a replacement PCB that is the same exact model and used for
the same exact model of hard disk assembly then the switcheroo *might*
work. I've done this in the past. It's worked maybe around 1 times out
of 6.

That could be because the drive technology predated "adaptive" data,
or it could be because the adaptives were in close tolerance.
That was back when I had access to LOTS of dead drives from which
to steal parts and had the time and inclination to waste on trying to
salvage an old drive because, say, one of the surface-soldered regulator
diodes blew up for the motor.

Not sure what you are talking about here. :-)

- Franc Zabkar
 
I used to think that all the firmware, calibration data, and defect
lists were stored on the PCB. However, both the factory defects
(P-list) and grown defects (G-list), and the bulk of the firmware, are
actually stored on a reserved area of the platters. The calibration
data, however, are nowadays stored in a serial EEPROM. When swapping
boards, you need to transfer the ROM or its contents from patient to
donor.

So I'll have to desolder it and solder it to the replacement board?
You guys will help me identify the ROM, right?

Without good calibration data, what won't be calibrated? Like how far
to move the head to get to a cyclinder?

I think one of 6 is pretty good odds, considering the other
alternative is, he's told, to spend I think it was 2000 dollars to
take apart the drive.

Also, maybe it only worked 1 out of 6 times becase in the other 5 the
heads or the "tone-arm" was messed up. That is, the problem was never
in the electronics.
That could be because the drive technology predated "adaptive" data,
or it could be because the adaptives were in close tolerance.

I have the inclination to help my friend.
Not sure what you are talking about here. :-)

- Franc Zabkar

You said before, Van,
: I figure if the files aren't backed up then they
:weren't important and are either okay to lose or are reproducible.

That's not a decision *I* can make. My goal is to do a favor for my
friend. I can tell you that he doesn't think they are okay to lose.

: The
:value of your time to salvage a drive exceeds the cost of getting a new
:drive.

We're not trying to save the drive but the contents of it.

Thanks, Franc and Rod and Van.
 
mm said:
That's not a decision *I* can make. My goal is to do a favor for my
friend. I can tell you that he doesn't think they are okay to lose.

How could your friend think his files were valuable when he took no
measures to keep backups of them? Only infants are unaware that
mechanical devices do fail. Now he wants you to do the recovery because
that would be cheap - and cheap is the value he has put on his files.
He should know (or you should've told him) that your recovery is risky,
will probably fail, and could make the situation even worse (by
corrupting the files so a later lab recovery is even less likely).

Make sure your friend has absolutely no intent to attempt a lab recovery
of his files if your attempt fails. It's your attempt or tossing the
drive in the trash. This friend should not expect to later send in the
drive to a lab if your attempt fails.

So is this a really old drive? If so, have you found another hard disk
for the SAME brand and model *and* which has the same version PCB on it?
Don't assume the same PCB is going to remain used throughout the sales
life of a particular hard drive. The drive assembly might stay the same
but they switch to a different version of the PCB sometime down the
road.

While it was mentioned the EPROM is soldered, I really doubt you have
the wave soldering equipment to remove and then solder back a
replacement chip. Have a look at this photo of the PCB for a hard
drive:

http://cobolhacker.com/images/content/gto_hard_drive.jpg
http://public.blu.livefilestore.com...stv4NigfudPJraWV-JXtKdgg6itRQXGg/P9060241.jpg

Look at the itty bitty pins on the chips and how close together they
are. Where are you going to find the equipment to desolder one of these
(without generating too much heat that damages the PCB or the chip).
Remember that you'll be heating the chip TWICE: once to desolder and
again to solder (onto the replacement PCB). There are some discrete
components, like resistors and diodes, that have just 2 solder points on
the tiny components and a micro-tip soldering iron will work if you are
very good at not using too much heat. Considering the size of the work,
you probably want to have an illuminated magnifying lamp so you can see
what you are doing - and a very steady hand. However, for the
hundred-pin chips that are also wave soldered onto the PCB, you really
think you can desolder a chip without damaging it and then apply heat
again without damage to solder it onto another PCB?

In the top pic, you can see a small square chip where a blue wire was
nearby soldered. That style chip can be, ahem, easily desoldered and
soldered. Presumably you have a solder sucker, solder wick, and a
micro-tip soldering iron and have the knack for soldering (e.g.,
remembering that solder flows toward the heat source and a very very
steady hand). There are some other components that are large enough for
hand [de]soldering, like the rectifier diode. Although really tiny, the
resistors and diodes can usually be [de]soldered by hand with proper
care. But as for those hundred-pin chips, well, good luck. Hope your
friend is willing to reimburse you for the proper equipment.

Yes, the chips are soldered. Yes, it's possible to desolder and
resolder the chips without damage and without shorting across the pins.
But is is DOABLE by *you*? You might get lucky if the EPROM isn't
soldered and instead is pushed into a carrier socket, like below:

http://blog.makezine.com/FD4RJI9BPXEXCFHR82.medium.jpg

I haven't seen this on a hard disk PCB but then I haven't seen every
hard disk. The lucky part is the EEPROM is an 8-pin chip, I think.
From the 2nd pic above, there are a couple of these package types at the
bottom left and right (e.g., just to the right of the EtronTech memory
chip is an 8-pin package). That's probably big enough and the style
that you can desolder and solder. I don't which does what because I
didn't bother to go look up their datasheets to see what they are. You
might want to read some forum posts by others doing similar low-level
hardware mods on hard drive PCBs, like:

http://forums.seagate.com/t5/Barrac...or-ST3500320AS-Firmware-SD35/m-p/45034#M18109
http://forum.hddguru.com/burnt-diode-resistor-capacitor-other-seagate-st380013as-t14919.html

The folks in those forums are way ahead of me. Personally, and since
this is not your files to recover, I'd back out of doing your friend a
favor and end up instead getting into hot water with that friend when
the attempt fails and the drive gets screwed up worse so now the lab
might not be able to recover anything. Let the friend think about it
for awhile and eventually the dead brick will just sit on their shelf
and those oh-so important files will be forgotten. Meanwhile, you
haven't rattled your friendship with a valiant but failed exercise.
 
So I'll have to desolder it and solder it to the replacement board?
You guys will help me identify the ROM, right?

You need to do some basic troubleshooting before you embark on a board
swap.

IIRC, you previously stated that the drive was "blank". That's a very
vague description. If the drive spins up, but reports a capacity of
0MB, then the problem is most likely inside the HDA (head/disc
assembly). OTOH, if it identifies with its correct model number and
capacity, then the problem could be logical rather than physical, in
which case you would clone the drive and use data recovery software on
the clone.

In any case, if this is your board ...

http://www.ixbt.com/storage/barr7200-7-sata/pcb-big.jpg

.... then the serial EEPROM is the 8-pin device to the left of the
Samsung SDRAM.

My notes should help you identify the components:
http://www.users.on.net/~fzabkar/HDD/HDD_ICs.txt
http://www.users.on.net/~fzabkar/HDD/TVS_diodes.txt

- Franc Zabkar
 
You need to do some basic troubleshooting before you embark on a board
swap.

IIRC, you previously stated that the drive was "blank". That's a very
vague description. If the drive spins up, but reports a capacity of
0MB, then the problem is most likely inside the HDA (head/disc
assembly). OTOH, if it identifies with its correct model number and
capacity, then the problem could be logical rather than physical, in
which case you would clone the drive and use data recovery software on
the clone.

In any case, if this is your board ...

http://www.ixbt.com/storage/barr7200-7-sata/pcb-big.jpg

... then the serial EEPROM is the 8-pin device to the left of the
Samsung SDRAM.

Yes, someone asked and I do have solder-wick and a solder suction pen
and bulb, and a fine-tip soldering iron, etc. but...

What happens if I just install the board without changing the eeprom,
and if it doesnt' work well, change the eeprom after that? Will I
have made things worse in between?

Thanks a lot.
 
mm wrote
Yes, someone asked and I do have solder-wick and a solder
suction pen and bulb, and a fine-tip soldering iron, etc. but...
What happens if I just install the board without changing the
eeprom, and if it doesnt' work well, change the eeprom after
that? Will I have made things worse in between?

Nope, should be fine.
 
What happens if I just install the board without changing the eeprom,
and if it doesnt' work well, change the eeprom after that? Will I
have made things worse in between?

That would be OK for this particular drive. However, I have been told
that there are some Samsungs or Hitachis where the MCU updates the
EEPROM. In such cases there is a chance of making the problem worse.

- Franc Zabkar
 
That would be OK for this particular drive. However, I have been told
that there are some Samsungs or Hitachis where the MCU updates the
EEPROM. In such cases there is a chance of making the problem worse.

- Franc Zabkar

Good to know. Thanks, and thanks Rod. You both said what I wanted to
hear.
 
Back
Top