Anatomy of a OCZ SSD smoked drive.

[josh1980]

just one question

why the sata connector is like that???

My guess is that the wires were wrong on the power supply. If you look closely at the SATA power, pins 7 to 9 are cooked. If you look real closely you can see that the coating on the PCB was burned off and the copper is showing. Pins 7, 8, and 9 are all +5VDC. Being that it looks like current ran from pin 7 to 9(or vice versa) I'd say that pin 7 and/or 9 were possibly wired wrong. Perhaps power went in through the SATA power and out through the SATA data cable via ground?

In any case, I'd put money in it being a power supply issue since these drives were used before. As much as I'd love to see the wires on the SATA connectors, I'm not sure we'd be able to identify which cable was connected to which pins inside the connector. For molex connectors they'd be very easy to identify.

[NEOAethyr]

If you look closely you'll see that the middle of the power connector is fried too.

But even so, why in the world would it be able to travel down the sata data connector path :\ ???

That and the grounding sticks out to me, the resin on the pcb looks as if it pealed off the grounding areas.
Looks like power was sent to ground and data ground as well, shouldn't even be the same thing... unless you're shielding the cable and if that were the case you shield a cable by connecting the shielding to the grounding src, not the grounding dest...

I mean since when does a sata convertion or pata conversion chip take a 5v/3v lead and short it to it's data ground?

It's just unheard to me, doesn't make any sense.
The prob looked like it started with those 2 chips to the side there.
8pin chips looked like it had something to do with power.
The chip above it looks like a fried tantalum cap...
And around it some smd resistors and another tantalum cap.

The power connection where it fried is in the middle, not the far corner where the 12v or 3v would be.
Something to do with the 5v or the ground as those are the only wires close to that location normally.

Possible I supposed a wire was in the wrong position.

I've done that before on a pata drive, killed it (had 12v swaped with 5v on that one connector when I was adding ferrite beads to the cables).
But that only took out a small smd part by the 5v that I was able to break off, then redo the cable and the drive ran fine again faster then ever, ran straight 12v since.
Cool mod, but something I probably won't even try to attempt to duplicate.
What I mean is, the only parts on that drive of mine that died were by the location of the swaped power cables (5v specifically).
It didn't travel down the data lines...
It was a seagate 7200.8 400gig that my mom still uses to this day.

Though I have never swaped the ground with say the 5v...
That would of been to obvious for me not to notice a wiring screw up like that.
I don't put it past a psu maker to do that, but you'de gotta be blind or something to do wire a connector like that.

I wouldn't mind seeing the sata power connectors on that psu.
Maybe a hot wire was swaped with ground, I dn.

Edit:
@josh1980, I could identify the sata power connectors if they were wired wrong...
Each wire uses 2 or 3 pins, I've wired a few by hand before too.
It's not much diffrent from a ide power connector, just that the side angle ones are a pos and hard to wire up...

[Tiltevros]

im not taking the side from noone but i cant understand why those 2 big companies through the ball in most of the cases to the buyer?

for me +1 to silverstone for giving a full statement about the problem to the end user and -1 to OCz for not even take a look to the problem.

Computurd i hope u find a good end with this, and i hope not to see another case like this.

[masterg]

You should take the PSU if you still have it and grab a digital multimeter and measure the voltages across the lines yourself.

[josh1980]

There just isn't enough information to really blame any of it on any given vendor. Most likely, it was the PSU since that was the new component. However, there could have been a strand of wire that fell inside the SSD. There could have been a stray strand of wire from manufacturing that was left inside the SSD and hadn't shorted anything out until now. We really will never know for certainty. Jonnyguru reviewed the same model and he had very good comments about the PSU. Although the PSU itself was well made from Jonnygurus review that's not to say mistakes aren't made. Honestly, to me, this was nothing more than a fluke accident. Smoking an SSD like this isn't common and the PSU passed it's diagnostics per Silverstone. I don't blame OCZ for not doing the test though. Those tests can be quite expensive. The equipment alone for performing the diagnostic test to determine the cause can run into the millions. If this were a recurring issue with OCZ, I'd bet money they'd be asking for it back so they could analyze it. Being that it was a one time event theres relatively little value to be gained(from the company's perspective). Honestly, I'm not sure if I would have called them up after I had seen it get smoked. Any evidence of what happened likely vaporized and went up the OP's nose .

@NEOAethyr

I'm just saying that once the PSU wires go into the black sata connector we wouldn't be able to "see" where each colored wire was connected. They might have looked correct on the outside, but since we can't see inside the sata connector we wouldn't be able to say for 100% that a wire wasn't connected improperly.

[NEOAethyr]

Well we probably could tell from the colors of the wires but if a wrong color wire say red for ex. was hooked into the 12v then we wouldn't really know.


Now that I think about it though dudes.
I bet ya those rom chips in those ssd's are 100% ok.
Rom chips these days are extremely durable.
Well most are...

You can short the heck out of them and they'll still work when hooked up right.
Ex. the older sst plcc bios chips, put one in the socket backwards, it heats up alot.
Put it back in normally and it still works .

You can take magnet to a usb stick's rom chip and it won't clear a single bit.
At least when I tried it - it didn't.

A small mount of parts could be replaced and the ssd would run like new, with the data intact.

Now a ssd isn't a normal drive, you can open it without worrying about dust getting on a platter and killing sectors.


@ Ryder
Would you guys actually take a rma like that and fix it if thread op put them back together, for a small cost of course?
You guys could probably fix those drives and make them like new again... (that is if the pcb didn't need replacing, removing and putting the bga chips on a new pcb from what I've heard is not easy or worth the effort)

[jasonelmore]

That is a really great piece of advice.

If I am correct, Computurd has had these drives for over a year running, testing, benching, just fine. The first power up on a new PSU and snap, crackle, pop. I will also say that he is the very first person to ever present me with a fried SSD, out of 200,000+ Indilinx drives, nobody has ever come to OCZ with a fried one, I am 100% sure the SSD did not cause this and so were my superiors.

So what's a couple SSD drives to you guys? come on now. Keep your customer's happy and it will pay back 10 fold by the people who read these boards. You also might could get those drives fixed, or looked at for next to nothing.

This thread was your chance to look like the good company who cares......
You wont even offer to look at the drives yet you 100% guarantee it's not your fault.

Wont be buying OCZ anytime soon, i'll stick with Corsair

[josh1980]

Companies care, as long as they wouldn't put themselves out of business caring . Can you imagine what would happen if OCZ said "yeah.. we'll send you a new drive" to someone that smoked their drive? Every time someone smoked their drive they'd expect the same treatment.

If a company started honoring warranties on parts that weren't manufacturer's defects where would you draw the new line? It likely wasn't manufacturer's failure since it was used for a year+.

[jasonelmore]

Companies care, as long as they wouldn't put themselves out of business caring . Can you imagine what would happen if OCZ said "yeah.. we'll send you a new drive" to someone that smoked their drive? Every time someone smoked their drive they'd expect the same treatment.

If a company started honoring warranties on parts that weren't manufacturer's defects where would you draw the new line? It likely wasn't manufacturer's failure since it was used for a year+.

they wont even look at it, they are just assuming. That's where the "care" needs to come in.

[[XC] Oj101]

How do you think your customer must feel with everyone stepping back and saying "not my problem"? I guarantee you whether you have a look and find, "oh, would you look at that! A one in ten billion defect that we didn't think was possible" and replace his drive or "see here? This blown diode had voltage put through it the wrong way. I lot more voltage than it was designed to handle" and don't replace his drive he'll feel much better about you as a company. Hell, even if you have a look and tell him "we're really not sure, we'll swap one out". It'd show you care. At the moment the message being portrayed is "we have your money now you can eff right off"

[dctokyo]

At least Silverstone actually had a look at his PSU instead of saying "don't care do not want don't bother sending it in". There's always a first time

+1

[josh1980]

How do you think your customer must feel with everyone stepping back and saying "not my problem"? I guarantee you whether you have a look and find, "oh, would you look at that! A one in ten billion defect that we didn't think was possible" and replace his drive or "see here? This blown diode had voltage put through it the wrong way. I lot more voltage than it was designed to handle" and don't replace his drive he'll feel much better about you as a company. Hell, even if you have a look and tell him "we're really not sure, we'll swap one out". It'd show you care. At the moment the message being portrayed is "we have your money now you can eff right off"

I don't think you realize how little information they would retrieve from a single failed hard drive even if it hadn't been disassembled. I can pretty much guarantee their results would be "inconclusive".

If they were expected to deal with the one in ten billion they'd never sell the product. There are always random issues that aren't often duplicated with other people. I think it's asking too much to expect a company to replace their drives.

Think about it like this. What if I had bought those 3 drives from you on ebay. I got it home, plugged it in and all three of your drives smoked. I call you up and want a refund. Would you give me your money back if those drives were all out of warranty? Or would you instead tell me no because they worked when you used them last and I must have done something wrong? I'd bet money VERY few people in this forum would willingly refund their money to a stranger that said "hey, this stuff was faulty!". I'd bet most of us as the seller would say "you did something and it's your own darn fault".

OCZ is going to obtain very little data on a situation that doesn't appear to happen hardly ever(if ever). I'll agree that one person would feel better, but how many other people would then expect free replacements when they really DO abuse the drive.

Honestly, many companies won't care to particularly much because one guy's drive burned up. It is more than likely that the average melted drive was because the user did something they shouldn't have. Companies don't want customers that try to do things to their products that they aren't supposed to do. Companies prefer that you use it exactly as designed and nothing more. While I wouldn't expect one of us to be tinkering with a drive, that's not to say others wouldn't.

It's a sad loss, but OCZ is acting just like I'd expect any other company to act. They need to stay in business and keep their prices low. I wouldn't want to pay more just because a bunch of people out there try to "overclock" their SSDs, then when they smoke try to get an exchange.

I want to add that for the record I don't think the OP was trying to do anything that would have caused the drives to fail, but I DO read about people that do experiment with SSDs on that level. I wouldn't want to pay more for my drive because they are replacing drives for those that do experiment. OCZ doesn't know you from Adam. But a guy calling in with a smoking SSD isn't exactly news to jump on. If 1000s of people start calling in, I'd totally expect OCZ to jump, and I'd bet that they would.

I'm pretty sure I won't convince you of my side, and you likely won't convince me of your side. Let's just agree to disagree

[Computurd]

@ryder

I am 100% sure the SSD did not cause this and so were my superiors.

sorry man, i dont buy that. you cannot be 100 percent sure of anything if you havent even looked at it. i have used OCZ, i post frequently on your forums, have done tons of posts/testing about you guys and your products on SEVERAL sites, for a very long time. just word of mouth is big, and i have given tons of it for your products, benching them on tons of raid cards, motherboards, blah blah blah. plastering results all over several enthusiast forums!...and have always been very happy. after this though....dunno man, leaves a bad taste in my mouth. if there had been some attempt at looking at the SSD, or anything of that nature i would feel better about it. its not all about getting them replaced or whatever. i have 496 posts on your support forum, my 9260-8i thread over there has 71,317 views, and 1,118 replies, started it over a year ago, and i still follow it. http://www.ocztechnologyforum.com/fo...hlight=9260-8i

im not some off the street who couldnt be trusted and trying to rip you off for free gear.

if you have a picture of the connector that plugs in to the power supply I would very much like to see it.

i have a few, i didn't even look at them after i took them, i snapped a few before i sent the psu off, and tried to get close ups to show some slight charring on the orange wire. now that i go back and look at them they are terrible though. crap!! however, i will attach them to bottom of this post

u forgot to mention that i saw the second testing with a different psu with OCz burning for second time. Live streaming for some flaming SSD's.

yes tilt watched me test the SSD with a different PSU (and a fire extinguisher handy) on video chat. hes my witness damnit! oh yeah, and the video i have of it it isnt nearly as dramatic as when it blew up though the first time

Did the damage occur the second you turned on the PSU for the first time? I assume you had other devices plugged in on separate power rails and no damage occurred on the other rails?

Did the PSU trip out when it occurred?

damage occurred when psu was turned on. all other devices fine, only the three drives on that one cable. had a total of ten SSD connected, and all my other gear of course. the whole rig! a areca raid card (1880) two gpu ram cpu, pump, fans, four HDD, cd, blah blah blah, really it could have been MUCH worse, if the psu was at fault. it just seems too much of a coincidence that only those devices got blown up, and none on the same rail with the other seven SSD!!
the PSU tripped after about one second of frying action. i jumped to unplug the PSU immediately, but it kicked back on after a one second lag, and then i got it unplugged. the second time on it began frying again btw, not sure if it kicked off itself, or i unplugged it in time. i do have the reflexes of a cat though.....

It’s seems a bit weird that the damage is on the SATA connector not the power terminal.

abso-frickin-lutly (did i spell that right ) very odd that. why all the damage there, yet no damage to my beautiful raid card? thats what it was connected to.

could you please measure the thickness of the metal plate in pic 3 computurd?

looks to be very thin bout 1/32? heck im not sure. thats an estimate...



@josh1980...you make valid points, but so do others, and so do i
and tbh it is more about the cavalier method in which it was handled mroe than anything to me i guess. irritating.
cmon man, its 3 30gb vertex, outta ten i had hooked up. yea it sucks to lose em, but they are easily replaceable. at least it wasnt the whole rig.
just odd that all other devices on that rail are perfectly fine.
odd that there is so much burnt all over the sata connector.
odd that it was just waved off so easily.
odd that i am wasting my time with this post

[Computurd]

terrible terrible pics, i should be shot for such terrible photogrpahmanship (but not with a camera! )

[lowfat]

Don't suppose you have a pic of where the cable plugs in to the power supply? That is where the issue would likely be.

[DooRules]

That is a really great piece of advice.

If I am correct, Computurd has had these drives for over a year running, testing, benching, just fine. The first power up on a new PSU and snap, crackle, pop. I will also say that he is the very first person to ever present me with a fried SSD, out of 200,000+ Indilinx drives, nobody has ever come to OCZ with a fried one, I am 100% sure the SSD did not cause this and so were my superiors.

well i guess you did not see the 3 ocz 30 gig turbos that i sent in for rma last feb...
in a very similar situation last feb i purchased 2 new 30 gig turbos to go with the 4 i already had, also picked up a new thermaltake 1200 w psu
upon first powerup there was a sizzling sound and then that god awefull burning smell, after shutdown much to my dismay i was out 3 turbo ocz drives and one psu
when i asked for rma i was up front about what happened and I got rma for all 3, when i sent in the 3 drives the burn smell was unmistakable, no doubt what had happened
also got psu replaced
this all happened while connected to my lsi 9260-8i, no other components outside of 3 drives and psu were affected, very lucky there

[Tiltevros]

did they accept ur drives DooRules? this is very critical

did they accept the rma?

[NEOAethyr]

If it's not user error then why wouldn't they just replace for the shipping cost?

I mean, if you just got a psu and hooked it up, fried a drive or 2 by accident.
Sent psu back, say they fixed it and didnt' tell you about it, tell you that the psu they are sending back is ok with it's qualification papers.
Now you goto send back the ssd's you fried and ocz says they don't want anything to do with it?

If it's still under warranty I don't see why they shouldn't replace those drives...
:\

I can understand that the drive's "could" of been in an env that is not the norm.
Ie getting fried by a psu maybe.
But like I said, if there's still a warranty, then what the heck...

I guess maybe in the end f@ck it and just not get them replaced...
I can't help but think that psu makers should warranty the equipment that hooks into them.
Stuff like this happens and the user is in the middle of a dispute between 2 diff companies and is told to fit the bill for themself.

Imagine if that were your new $300 cpu ^^ ...
Or your ram that could be impossible to replace at one time or another.
Or hell, some top of the line vga card that cost you $500+.

And to be onest.
I think it's policy not to replace stuff damaged like this (fried stuff) :\.
Might be anyways.

Sorry dude about your losses.
I think if I were in your position, I wouldn't of took it apart if I were to send it back.

Edit:
@DooRules
Simluler?

That sucks in 2 ways.
For 1, you lucked out and the thread op didn't.
for 2, maybe there is a prob with the drives ;| (or maybe not though I dn whatever lol).

[yngndrw]

did they accept ur drives DooRules? this is very critical

did they accept the rma?

and I got rma for all 3

(3char)

[Ao1]

Comp, why not get the SSD’s independently tested? I don’t know how much it would cost, but if it’s reasonable at least you will know for sure if it was the PSU or the SSD’s.

There are plenty of labs that will be able to get to the bottom of it. Here is an example in the UK:

http://www.lpdlabservices.co.uk/prob...stigations.php

[josh1980]

Comp, why not get the SSD’s independently tested? I don’t know how much it would cost, but if it’s reasonable at least you will know for sure if it was the PSU or the SSD’s.

There are plenty of labs that will be able to get to the bottom of it. Here is an example in the UK:

http://www.lpdlabservices.co.uk/prob...stigations.php

The link says that the primary point of contact is Dr. Howard Coulson. That should be a hint as to the price you could expect. Honestly, from the stuff I see the company I work at pay for investigations, you'd be talking at an absolute minimum probably around $50,000. Although I wouldn't be surprised if the price was over $100,000. That would be for just one drive. The equipment to do the investigation and the expertise and skills to do an investigation is much more expensive than you'd ever want to know. There's ALOT of documentation, pictures, electron microscopy, etc involved with an investigation. A good investigator could obtain enough information on the product being investigated that they could provide the engineering designs to duplicate the hardware without needing the product manufacturer.

That's why OCZ isn't too interested in this situation. They don't want to drop that kind of money to find out what is wrong with "some dude on the street that smoked his drive"(no offense to the OP). If people started calling in every day with this issue on one particular model, they'd drop the cash if they felt their name was going to be tarnished by the number of cases. We're not talking small potatoes by any stretch of the imagination. OCZ doesn't want to drop that kind of dough so they can tell someone that there had to have been a power surge. There would be little or no data for OCZ to use to improve their products in the future. It's pretty obvious there was a power surge of some kind just from the pictures. mA current flows don't smoke a drive like the pictures show. With no valuable data for OCZ to use on a large issue that's affecting large volume products they don't have any interest in the drives.

Edit: For the record, Silverstone is one of a handful of companies that develops and designs their PSU, but has very little in terms of manufacturing facilities. Most or all of their PSUs are outsourced for production. I am not trying to imply that makes their design inferior. Just posting an interesting fact.

[chuchnit]

Well after reading this thread, OCZ can blow me. It's not like he was even demanding replacements. Its not like OCZ sales products that are unique anyways. Corsair, Mushkin, Gskill, Supertalent and a host of other companies sale the EXACT same SF drives that you do!! The only reason to pick one brand over the other is the customer support. Really SSD's are all you guys got going for you anyways. Your PSU's have a bad reputation, and memory offerings are terrible. SSD's and the reputation of the brand is all you guys had going for you in the educated buyer's market.

[yngndrw]

mA current flows don't smoke a drive like the pictures show.

That's actually quite a bold blanket-statement. A 0402 surface mount resistor is rated to 1/32 Watts, which is 6.25mA @ 5V. You could quite easily get it hot enough to smoke while keeping in the mA range.

Additionally there's no way of knowing that the SSD was just drawing current in the mA range - Who's to know there wasn't a defect of some sorts which caused it to draw much more ?

Also you mention professional testing, but what about OCZ just doing some in-house testing and investigation with the equipment and knowledge that they have ? The OP isn't asking them to send them off to have some professional analysis, he simply wanted them to take an interest and look at them.

Look at the testing which Silverstone did, it was hardly an exhaustive test but it was a huge amount more than OCZ did.

[Ao1]

..... Honestly, from the stuff I see the company I work at pay for investigations, you'd be talking at an absolute minimum probably around $50,000. Although I wouldn't be surprised if the price was over $100,000.

We are not talking mega sophistic equipment that might be required to investigate a failure in a nuclear power plant. Move the decimal place over three places and I suspect you will get a closer ball park figure. Either way it doesn’t hurt to ask for a quote and chances are anyone with expertise in the field will give you a good idea of what the problem was just by looking at the photos.

Silverstone only used basic equipment to check the power supply and stimulate a voltage overload to see if the PSU tripped out. One of those bits of kit they used is a Fluke 289 Multimeter, which costs around £400. I’m certainly not knocking Silverstone, at least they made the key checks, but the cost side of this is not a big issue, in fact it’s a complete non-issue, because anyone manufacturing electrical goods will have access to test equipment that they use for their QA. The issue is piece of mind to know what caused the problem.

[josh1980]

We are not talking mega sophistic equipment that might be required to investigate a failure in a nuclear power plant. Move the decimal place over three places and I suspect you will get a closer ball park figure. Either way it doesn’t hurt to ask for a quote and chances are anyone with expertise in the field will give you a good idea of what the problem was just by looking at the photos.

Silverstone only used basic equipment to check the power supply and stimulate a voltage overload to see if the PSU tripped out. One of those bits of kit they used is a Fluke 289 Multimeter, which costs around £400. I’m certainly not knocking Silverstone, at least they made the key checks, but the cost side of this is not a big issue, in fact it’s a complete non-issue, because anyone manufacturing electrical goods will have access to test equipment that they use for their QA. The issue is piece of mind to know what caused the problem.

Actually, about 90% of the stuff we buy is used in other industries(and in some cases your home, vehicle, etc. Many parts you could buy yourself on ebay or at Home Depot. You'd be amazed. I know I was. I thought nuclear power would be bleeding edge and I'd get to see some very cool and very nifty toys. Not as sophisticated as one might think. Alot of the sophistication is in the computer systems. Everything else that is sophisticated is used solely by the nuclear industry.

I've always found it interesting that a pressure switch that cost less than $1000 would fail, but we'd spend 50x that to find out why it failed. If a water line broke in our house we'd just replace it and say it was old or something. In other industries they'd cut out that section of piping and send it off for analysis to see what the failure mechanism is.

That's actually quite a bold blanket-statement. A 0402 surface mount resistor is rated to 1/32 Watts, which is 6.25mA @ 5V. You could quite easily get it hot enough to smoke while keeping in the mA range.

Additionally there's no way of knowing that the SSD was just drawing current in the mA range - Who's to know there wasn't a defect of some sorts which caused it to draw much more ?

Also you mention professional testing, but what about OCZ just doing some in-house testing and investigation with the equipment and knowledge that they have ? The OP isn't asking them to send them off to have some professional analysis, he simply wanted them to take an interest and look at them.

Look at the testing which Silverstone did, it was hardly an exhaustive test but it was a huge amount more than OCZ did.

Yeah, but a 1/32 watt resistor wouldn't leave that much charring on the board. That's why I said "mA doesn't smoke a drive like the pictures show". I've seen larger resistors turn to a chunk of carbon and we didn't even smell it. You will smell something burning before you see smoke. The fact that he saw the "magic smoke capsule" break means he did alot more than blow a 1/32 watt resistor. If you call up a company and say "my component smoked when I turned it on" then you did ALOT of damage. There's going to be more than 1 very tiny resistor burned. You likely had a low resistance connection to ground or high voltage condition that burned out alot of components.

It's quite possible that OCZ doesn't own the equipment to perform the kind of testing necessary to determine the cause for failure. Like I said before, the equipment and personnel with the skills necessary are few and far between. Most companies outsource those kinds of work because it's cheaper to pay someone to do an analysis on the rare occasion you'd need it than to try to maintain the equipment and people in-house. There's a few companies(AMD comes to mind) that do not have the equipment to examine a CPU if it were to arrive a charred mess. If I had smoked 3 drives I wouldn't have likely tried to call them. I'd expect that if I had called them on it they would have said I had to have abused the drives or done something wrong to burn up 3 drives. I would however, have put little faith in the PSU I had just bought and would have decided to start testing it(and calling Silverstone) and see what was wrong.

Comparing the testing Silverstone did was much cheaper to perform than testing on electronics such as an SSD. If the power supply had been bad they would have said something like "bad caps" or "bad regulator" or something. Easy when the caps are very large and when they fail its typically very obvious. For electronics it's a whole different ballgame. Comparing the amount of work Silverstone would have to perform, they had the easy part of the job.

Just my $0.02.