samsung 1tb drive wont power on

[rubixcube]

Hey all

Was working on my new pc tonight and something somewhere along the line managed to short -- one of my seagate 500's and my samsung 1tb got wrecked. the seagate shorts out whatever you plug it into (dock, PC).. the samsung doesn't cause as much of a hissy fit but it just wont power up.

The contents of the samsung drive are important (sentimentally) as it contains every single photo I've ever taken digitally (about 7 years worth). Obviously I should know the importance of keeping backups, and I was going to get a new 2TB drive soon, but really when you are just doing something as simple as swapping drives between PCs you dont expect 2 of your drives to be killed in the process.

Anyone have any experience in this and could suggest a solution?

Thanks

[Renza]

how much is the data worth to you? Might be worth getting someone to recover the data for you, theres an OCAU sponsor (DataRetreiver) that does this. Probably wont be cheap but no harm in asking for a quote.

You can try source a board from an identical drive and move the EEPROM chip over and see if that will work.

This is the part where you usually take out the backups and restore them. note the plural. If its important then you need multiple copies if it. You never expect anything to die... lesson is, always have a contingency plan!

[rowan194]

Do you know what caused the shorts? Perhaps using a molex "Y" adaptor which ended up plugged in the wrong way, or a modular PSU which lets you plug things in backwards?

If you're lucky it may just be that the sacrificial components on the HD PCB designed to protect against this scenario (12V applied to the 5V rail) did their job. Sometimes removing (diode for clamping) or replacing (fusible link) these components will bring the drive back to life... but there's also a risk that removing these protective components may cause further electrical damage to the PCB.

If there's damage beyond the protection components you may need a PCB swap as suggested by Renza, which does complicate things. Either way, if the data is important, it's probably worth getting a pro to look at it.

[rubixcube]

Originally Posted by rowan194

Do you know what caused the shorts? Perhaps using a molex "Y" adaptor which ended up plugged in the wrong way, or a modular PSU which lets you plug things in backwards?

If you're lucky it may just be that the sacrificial components on the HD PCB designed to protect against this scenario (12V applied to the 5V rail) did their job. Sometimes removing (diode for clamping) or replacing (fusible link) these components will bring the drive back to life... but there's also a risk that removing these protective components may cause further electrical damage to the PCB.

If there's damage beyond the protection components you may need a PCB swap as suggested by Renza, which does complicate things. Either way, if the data is important, it's probably worth getting a pro to look at it.

I have no idea what caused it, I wasn't using Y adapters, just a regular SATA daisy chain, and my modular PSU prevents things being plugged in backwards.

I can't really afford to have a pro look at it, I don't have any money to my name

What kind of thing should I be looking for on the PCB ? could it be on the underside?

[Jaco]

Originally Posted by rubixcube

I have no idea what caused it, I wasn't using Y adapters, just a regular SATA daisy chain, and my modular PSU prevents things being plugged in backwards.

I can't really afford to have a pro look at it, I don't have any money to my name

What kind of thing should I be looking for on the PCB ? could it be on the underside?

You can usually source a whole new PCB from places like ebay etc, just enter the drives model number and make sure you get the correct one

[rubixcube]

ok so I did a bit of googling and found that I should measure the TVS diodes near the power connector with a multimeter.

heres a pic of mine
http://i.imgur.com/Zs9mq.jpg

busted out the meter, and put it on diode test mode.
the lower one reads a constant 1 in one direction, then 715 in the other, which I am led to believe is 'normal'.
the top one however, is jumping around all over the place then settling at about 001-002 in BOTH directions. Am I safe to assume this means it's buggered and needs to be cut off?

edit: also measured them both on resistance mode at 200, the bottom one doesnt change from the default "1---" in both directions (over range?), the top one reads anywhere from 2-6ohms in both directions.
also, the 0 Ohm resistor above the top diode is reading "1---" when tested.

[rowan194]

Originally Posted by rubixcube

also, the 0 Ohm resistor above the top diode is reading "1---" when tested.

What happens when the multimeter contacts aren't touching anything? Does it show 1--- or something else? A 0 ohm resistor (or rather a fuse) should definitely be showing a very low range value (like 0.0xxx), so it's possible it's gone open circuit.

Did you find anything for your specific model of HD? Generally, you'd bridge the 0 ohm resistor, and possibly remove one or more of the TVS diodes.

[rubixcube]

Originally Posted by rowan194

What happens when the multimeter contacts aren't touching anything? Does it show 1--- or something else? A 0 ohm resistor (or rather a fuse) should definitely be showing a very low range value (like 0.0xxx), so it's possible it's gone open circuit.

Did you find anything for your specific model of HD? Generally, you'd bridge the 0 ohm resistor, and possibly remove one or more of the TVS diodes.

when the contacts are in air it shows 1---

is it possible the TVS diode AND the zero ohm are both cooked? Or the diode is affecting the resistor reading? The diode closest to the resistor was the one giving shonky readings.

Bridging the 0-ohm resistor should be easy, just a small piece of solid wire from one side to the other yes? I have a low power soldering iron that shouldn't damage any heat sensitive components. How do I go about removing the diode, too? Even if it might not be fried (but I suspect it is), I figure it doesn't matter if I remove it since all it does is provide protection. I'm planning on using a different PSU and buying a spare drive to immediately clone the contents.

Yeah I found a couple of things for the samsung but nobody seems to provide a definitive result. Just "thanks solved!" or "yeah I cut off a diode"

[rowan194]

I'd guess "1---" on the display is signalling that the multimeter can't autorange as it can't sense any resistance at all, which would mean your 0 ohm resistor is open circuit (blown)

I'd probably be a bit more conservative here: rather than simply bridging it with a bit of wire, I would place a normal fuse between the two contacts, say a 2A or 2.5A. (Try fast blow first, then slow if it blows immediately.) It will be a little unwieldy as you'll have to run two wires from those tiny pads up to a relatively large fuse, but it's a bit safer than a straight bridge... if there's a short elsewhere on the PCB there is no further protection so you may find a chip smoking or your power cable insulation melting.

You should be able to remove the diode by alternately heating the ends of it (couple of seconds each, repeat) and gently pulling the diode up with some tweezers or needle nosed pliers. If that doesn't work, you could try cutting it through the middle with wire snippers, but take care as this could damage the pads or any tracks under it.

Since it's in the "power" section the PCB tracks are likely to be thicker, which will sink more heat, so you may need some additional time to heat the pad enough that the solder melts.

[rubixcube]

Fuses like this one?
http://jaycar.com.au/productView.asp...e&form=KEYWORD

I'll head to jaycar on the weekend and check it out. Might get some flush cutters too if I can't get the diode off. Once its off, does the gap it leaves need to be bridged?
Sincere thanks for your help btw

Edit:
Oh, seriously dumb question coming up..
Pretty sure I'm measuring in the right place...
Zero ohm resistor measurements are taken either side of the tiny 0 on the solder spots yes?
Excuse my ignorance. The diodes were easy because they are big

[rowan194]

Yep, you measure with the probes sitting on each solder spot.

DON'T bridge the diodes as that will most likely create a short circuit and will immediately blow your fuse. Just leave them off.

[rubixcube]

Originally Posted by rowan194

Yep, you measure with the probes sitting on each solder spot.

DON'T bridge the diodes as that will most likely create a short circuit and will immediately blow your fuse. Just leave them off.

excellent. I'll head to jaycar sunday and let you know how i go. thanks mate

[rubixcube]

So, here's the contraption and the location its going.
http://imgur.com/a/Tw6Ws
Correct?
Wanted to check before I went ahead.
Thanks

[mike-s]

Personally I'd just pony up the money for a new pcb and avoid the risk of nuking the drive due to the lack of backup, but that's me I guess. Have done it a couple of times with success in the past.

Will watch with interest however. Good luck.

[rowan194]

Originally Posted by rubixcube

So, here's the contraption and the location its going.
http://imgur.com/a/Tw6Ws

Yep, looks good. Don't forget to remove the diodes. Also minimal contact of the soldering iron on the PCB, just enough to melt it.