Burnt diode/resistor/capacitor/other in Seagate ST380013AS ?

[Pecm]

Hi, I am new to this forum, I wonder if you could help me:
I have a Seagate 80Gb SATA ST380013AS barracuda, and suddenly it stopped working. Now, everytime I plug the the power cable to the motherboard, the pc won't even turn on (some short-circuit protection system from the power source, for sure)

Examining the Hard disk PCB (logical board) I have found a component with a suspiciously toasted/cracked look

Here's a foto (the component is signaled in red, you may have to click the pictures to see them in full detail):

Higher resolution HERE

..very near the power cable, as you can see
Unfortunately, and God knows why, I tried to take the screws out with pliers, and *bam*, another stinky microbian SMD met his maker... :P ; in fact, it just disappeared He is (was) near a big squared chip with the ST logo.


Here are some more fotos with that one also signaled (as well as the 1st one. ) :

Higher resolution HERE

Higher resolution HERE

Higher resolution HERE


And so, the inevitable question of course is, does anyone knows what type of components they are? capacitors, resistors, diodes, voltage spike protectors, ... ? and do you know their values as well, so I can weld equivalent ones? Or at least do you know any schematic of this Seagate, because I searched the Net a LOT, and found nothing!
I have to recover my data.
Please, please, PLEASE, help.

Thanks in advance

[fzabkar]

You should have used a Torx screwdriver. However, the irony is that you probably don't need to remove the board.

What you do need, however, is a digital multimeter (~$10), a pair of side cutters, and a soldering iron.

The following thread has detailed instructions and photos:
http://forums.seagate.com/stx/board/mes ... 109#M18109

You can remove the capacitor near the power connector. The drive will work without it, but it would be better if you replaced it. The capacitor's function is to filter high frequency electrical noise. That said, the usual culprit in cases of short circuits are the TVS diodes. The part labelled "BUX" is the 12V TVS diode, and the smaller part to the left of the capacitor is the 5V diode. You can also remove the shorted diode, provided that you can trust your PSU. The drive will work without it, albeit without protection on the affected supply rail.

As for the other part near the MCU, it is probably a bypass capacitor. Measure the ones above it with a capacitance meter (your DMM may have this function). What the hell, try running without it.

If all else fails, obtain a replacement board and transfer the 8-pin serial EEPROM chip (25P05AV) from patient to donor. It stores drive specific calibration data determined at the factory.

[Pecm]

Hi, thanks for your response mate

So, bottomline, you said that the cracked capacitor isn't the reason for the apparent HDD short-circuit because he just filters voltage spikes to the ground, as well as the other one near the big chip (MCU, you called it), and both aren't vital.
And it's the 'BUX' that is giving me problems, so I have to remove it. What about the 5v diode, can I leave it? and I assume you were referring to the nearest black component to the cracked capacitor? or the brown one that's between the craked and the BUX? (you said "smaller part to the left of the capacitor", but that depends to which foto are you referring to, because they have differente orientations)

Do you happen to know the values that those capacitors have, give or take? 10pF, 10nF,.... ?
I would replace that bypass cap for a non smd, but that would be the same as soldering a electrical valve tube to a wrist digital watch, just to replace a transistor not to mention the electronic noise that something clumsy like that would generate (but still, it could improve MCU's safety, right?)

Finally, do you know if a standard diode can replace the TVS? (I'm guessing it doesn't...)
Thanks again.

Best regards,
Pedro

[fzabkar]

You misunderstood me.

If there is a short on a particular supply rail, then any part that is connected across that rail could be responsible.

You first need to measure the resistances between ground and each of the +5V and +12V rails. Do this with your multimeter set to the 200 ohms range. If there is a short, then remove the suspects one by one until the short goes away.

Here is the pinout of the SATA power connector:
http://pinouts.ru/Power/sata-power_pinout.shtml

I don't see any problem with using a leaded replacement for your capacitor, other than size.

However, I would try to determine which component is missing alongside the MCU. If it's a bypass cap, then running without may be OK. If it's a resistor or timing cap, then it will be critical.

[Pecm]

Ok, thanks, I'll see what I can do and post here the results

However, I would try to determine which component is missing alongside the MCU. If it's a bypass cap, then running without may be OK. If it's a resistor or timing cap, then it will be critical.

Perhaps the MCU datasheet has some schematic featuring that missing component. The strange thing is that I can't find that datasheet in ST website. I've searched all the serials printed in his package, but the search engine of their site doesn't find anything. Oh well, I'll try google .

Best regards,
Pedro

[pepe]

I doubt u would find anything...

[Pecm]

You're right, not even in google.... what the hell, I would think that chip doesn't exist, if I wasn't staring at it right know. Is he that secret for not having a datasheet? (well, not for consumers anyway)

[drc]

Yes

[Pecm]

Bummer

[fzabkar]

Perhaps the MCU datasheet has some schematic featuring that missing component. The strange thing is that I can't find that datasheet in ST website. I've searched all the serials printed in his package, but the search engine of their site doesn't find anything.

I suspect that hard drive MCUs are custom made components. There will be a CPU core (eg ARM) with various I/O modules, eg DACs, ADCs, UARTs, counter/timers.

If the datasheet is available at all, you will probably be asked to sign a non-disclosure agreement (NDA), or you will be advised that Seagate hold the rights to the design.

[Pecm]

If the datasheet is available at all, you will probably be asked to sign a non-disclosure agreement (NDA), or you will be advised that Seagate hold the rights to the design.

Didn't found anything


Ok, so here's what I got so far: I desoldered the cracked capacitor, pluged power to the drive, and nothing, no noise, no 'clicks';
(here's again a picture of the drive, with the cracked capacitor signaled in red, before any desoldering):

Higher resolution HERE

...then I desoldered the TVS in the upper left corner (the bigger one, black), still nothing. Next I desoldered the smaller TVS (right under the place where the cracked capacitor was), still a useless paperwheight. Finally I desoldered the remaining capacitor (in vertical position), and again, nothing >:(

Just to be sure, I tested the both TVS (they are diodes) and they conducted current just one way, meaning they are fine, and the capacitor charged and discharged, so it is fine too.
The culprit must be in other place, but where? none of the other components has burn marks/cracks.

Before desoldering, I tested the conductivity between power rails, and I got some strange results: from 3 pins that are supposed to be GND, 1 isn't connected to the others, the same to the 3 12V pins. And one GND pin (nr. 4) is connected to the 3,3V rail
It's because it's damaged, you're thinking... well, the same thing is happening to the other functioning drive. In fact, the browser from where I'm writing this, as well as the Duran Duran track that I'm listening now, are stored on that other drive.

I analysed the bad drive a little more, when I noticed something strange: all the pins, pads, vias, components of the PCB were electrically connected every capacitor, to every pin of every chip. It's as if the air itself was a good electrical conductor, so everything is connected/shunted. Oh, and the same thing happens to the good drive.. what the hell?!

[Pecm]

EDIT: only the second capacitor was good, the first (cracked) is virtually a conductor

[fzabkar]

I respectfully suggest that you get some hands on help.

The SATA pinouts are 15, 14, 13, ... 1 running from left to right in the first photo.

It is evident from the photo that pins 15, 14, 13 (+12V) are connected together, as are pins 9, 8, 7 (+5V), and 6, 5, 4 (Ground), and 2 & 1 (+3.3V, probably not used).

BTW, the 5V TVS diode is the black device in parallel with the chipped capacitor. In some photos it is to the left, in others the photo is rotated. I hope that didn't throw you. In any case, the references I gave you should have identified the parts.

It would be a good idea to measure each of the components you have removed.

And, BTW, on end of most of the capacitors would be connected to ground.

As for connectivity with "every pin of every chip", it sounds like you need a tutorial on how to use a multimeter. There is a link to an online tutorial in the Seagate forum thread.

[Pecm]

The SATA pinouts are 15, 14, 13, ... 1 running from left to right in the first photo.

It is evident from the photo that pins 15, 14, 13 (+12V) are connected together, as are pins 9, 8, 7 (+5V), and 6, 5, 4 (Ground), and 2 & 1 (+3.3V, probably not used).

I thought it was 1 -> 15 left to right

Anyway, I measured all again, and although I know that I should have 5 groups of interlaced power and ground, each one with 3 pins, in reality what I have here is this:

Code:

                     spin
     12V        GND   up   GND         5V              GND               3,3V
___________      __        __      ___________      ___________      ___________
15   14   13     12   11   10      9    8    7      6    5    4      3    2    1
|     |    |      |         |      |    |    |      |    |    |           |    |

\_________/       \__________________________________________/             \___/
connected                         connected                             connected



taken from http://en.wikipedia.org/wiki/SATA#Standard_connector

...meaning, pins 15, 14, 13 are connected, 12, 10, 9, 8, 7, 6, 5, 4 somehow connected; 1 & 2 connected just to each other, and 3 & 11 (spin-up) aren't connected to anything. And the 15-14-13 group (12V) is *almost* connected to the 12-10-9-8-7-6-5-4, very low resistance between them
That means that not only the 3,3V rail is shorted with the ground, but also there is very few resistance between ground and 12V.
The funny (not so) is that I obtain the exact same readings from another working SATA drive!

As for connectivity with "every pin of every chip", it sounds like you need a tutorial on how to use a multimeter. There is a link to an online tutorial in the Seagate forum thread.

Actually I know as bit of electronics and used a multimeter many times before, but I can understand why you wouldn't think so, because of such strange readings. Trust me, they are very strange to me too.

I don't have a multimeter, but I made something to test conductivity just like a multimeter would do:

Code:

                             / /             _______
      _________/\/\/\/\_____|\|_____________|    +  \
     |                      |/|             |_______/
  ___|___
    ---
     |                                     _______
     |____________________________________|    -  \
                                          |_______/

The diode is a LED. As you can see, every time I place between the probes a wire, conductor, shorted component, or even touch the probes toghether, the LED lits. As a multimeter would make a sound if you would do the same thing with it's probes.

But I admit, it's not as exact determining if there's still a small resistance or it's really shorted, as a multimeter would do.
About the readings where everything in the PCB was apparently connected to everything (caps, resistors, chips) the same thing happens with a working SATA and IDE drives that I have here, but I noticed *almost* imperceptible light variations in the LED, so it isn't like everything is shorted afterall (how could it, even with the working drives?), but anyway I would expect much bigger variations, or even the LED off most of the time.

(I know you are going to tell me to get a real multimeter... )

I think that the problem is somewere else, in a damaged component but without any burns on it.

It would be a good idea to measure each of the components you have removed.

I did, the TVSs are working, because they only alow current to flow in one way and not the other (and the correct way, with the negative of the probe touching the TVS end with the stripe), the capacitor charges and discharges (the LED's light faded for about 1 second). The other cap, (cracked) isn't a capacitor anymore, but a conductor

And, BTW, on end of most of the capacitors would be connected to ground.

You mean it's another way to identify where the ground rails are? I know.

[Pecm]

EDIT: where I say, below the SATA schematic, "That means that not only the 3,3V rail is shorted with the ground", I mean obviously the 5 V rail. Sorry.

(where is the "edit post" button? )

[pepe]

U can edit posts within 15 min or so...

[kaziulaz]

Tem lies me in garage that HDD if you still want continually to know some parameter of some element this ask and I will take out him and I will measure.There was damege TVS but HDD was work as i remmember so rest elements shoult be nice.

[Russwinters]

These boards have external eeprom, why not just replace board and transfer eeprom?


yes repairing is more fun, but if you want the data quickly, then just use the fastest method possible.



It is good to learn repair though, so in those cases you cannot find matching board.



Regards,

[Pecm]

Maybe I will do that, as long as I can get a used and identical board. That, or replace some components that are the size of big grains of sand
I had several other things to do and didn't fix this thing yet, but when I do, I'll post it here.

[kaziulaz]

That capacitor that you mark red arrow (near source 5v) have 10uF