Anatomy of a OCZ SSD smoked drive.

[RyderOCZ]

I will start out by saying that I hear you guys, I am listening not just brushing this off, but let me give my thought process here relating this.

I will say right now that the following comes from my understanding of electricity and circuits. I don't have a PhD, but I have been around.

If 1 of the SSD's was at fault, then there should be no way for the other 2 to burn up.

Example: Let's say I have a toaster, blender, and a lamp on a circuit. The lamp is on and I put some toast in the toaster and some concoction in the blender. The blender suddenly becomes jammed which increases the load on the motor and causes more current draw, the motor then overheats and burns the brushes/coils/whatever before the circuit breaker blows.
The lamp and the toaster are fine because their internal resistance is still in line with operating specs, just because the blender drew a bunch of current, doesn't mean the other devices saw all that current.

All 3 SSD's would have had to have a failure at the same time for all of them to burn, which pushes the odds even farther into the very unlikely.

Now, if I were to go to the circuit box and reverse the ground and the hot wires on that circuit, all 3 devices would fry instantly when they were powered on because current is going to the wrong place.

That is what most likely happened here given that all 3 devices on the string all fried. If 1 of them went bad, the other 2 would not go with it.

Someone is going to come along and shoot holes in this because I used an AC example instead of DC, but just replace my devices with 3 DC devices, it still applies. The other devices in the circuit are not going to see the increased current, which is what does the damage. Revers the + / - and you would have all sorts of smoked equipment.

[Computurd]

All 3 SSD's would have had to have a failure at the same time for all of them to burn, which pushes the odds even farther into the very unlikely.

not all three burnt. only one did. one smoked/burned. the others will not power on after the incident, but no smoke or anything. if there was an issue with the device on the last string of the cord, then you would imagine it would cause an issue on the whole cord.

it was the last device on a three sata power cord. why didnt the two BEFORE it on the cord burn up? is the reason they were not powering up because of an increase of voltage 'tripped' them in effect? or some sort of grounding issue, all three are powered and grounded on the same wire. did an overvoltage (caused by a short?) on a bad device, the last device in the series as a matter of fact, cause the power fluctuation on the whole line, as silverstone says it did?

All 3 SSD's would have had to have a failure at the same time for all of them to burn,

thing is, until now, you wern't even aware of which device shorted, or its position on the string. you also thought all three smoked. they didnt.

no one cared enough to ask.

[RyderOCZ]

Overvoltage? the SSD doesn't produce voltage, it uses it. How can an SSD overvolt anything?

[Computurd]

if there is a big catastrophic burning, causing some type of overvoltage on the line, or unstable power fluctuation. dunno. i am no electrical engineer, but i do know this.
electricity comes in one end, but the ground also goes out that same end, correct? so, if you have a bunch of crazy power going on at the end of the line, cannot that power theoretically go back up ground lines, etc? as you can see from the melting and the burning pretty much everywhere, it is entirely possible that power went back up the ground. looks like power went everywhere.

A: power and ground supplied by line from PSU
B: device unresponsive after event
C: device unresponsive after event
Oh Dear GOD: SSD fried here. now, in order for that drive to be grounded, that ground has to go back over the other two devices' grounds first, right? I could be wrong, but don't think so. so if voltage is going everywhere, maybe it went back up the line. it doesn't look like at that point the end device could regulate power in any fashion, or control where it went. hell look at the sata port!

so how does it burn up the last device, without burning the other two, they are just unresponsive. if i put voltage to the ground of those other two devices, would you say that they would be unresponsive after that?? dont know myself, just trying to figure it out.

the ssd may not create voltage, but when it is frying and burning due to electrical current, i would say that it could create some type of fluctuation.

[Computurd]

okay no genius here. but here is my retard rendition of how power is served to the devices: power goes up one side, ground comes back down the other. someone with more brains than i with this type of thing, tell me, is this close?
would they not all share the same ground wire?

[xeon_1]

What probably happend .
SSD that fried eg the last one on the cord developed a short circuit in the range of 40 to 90A proably (not a problem for the psu as it can push several hunderd amps on the 12V rail).
But however the wire and the connector can't take that much current and start to heat up and melt.
1.5mm² wire iirc can take arround 25A before heating up and every obstruction is going to cause more heat (cable clamps of the connectors)
That is why the first two the connectors are fried .
This eiter caused a short on the connector of the ssd's thus breaking them or a solderpoint came lose near the connector preventing the ssd from powering since it is not getting any power.
And the last one is extreme hard to tell visualy with SMB components since you don't have solder pins to inspect..

This is my take on the whole thing.

[Razmatazz]

Just to lighten this thread up a bit...

Since I work at a nuclear power plant, we routinely send off components that fail catastrophically to see what the heck happened.

I think that I'm glad that I don't live next to you. Who's to know just what might happen with one of those catastrophic failures?


Regarding the last several posts, I am interested to hear what the replies will be by those who do know more about this kind of thing.

[Computurd]

yeah its getting interesting for sure...nuts and bolts

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

hello mr doorules. sorry i missed your post earlier thx for sounding off, i know you from OCZ forums as well, and i appreciate your insight
i was upfront too, but seems the dice didnt roll my way... well at least we get an interesting mystery outta the whole deal well, first the mystery of what happened to the drives,
and the second mystery, why they granted you an rma, and not me, when we have almost the exact same circumstances..

I can't help but think that psu makers should warranty the equipment that hooks into them.

much like you can buy a surge protector that has 250K of insurance for connected devices, there IS insurance for equipment hooked to a PSU, at least silverstones. they did say they would pay for it, if it was their fault.
they just genuinely (i suppose) do not feel it is their fault. they tested the damn thing and it was fine is what it boils down to. hell I tested the thing!
but when i tested the drive, of course it smokes still...and the other two still dont respond.

[johnw]

Was there any visible damage on the circuit board or components of the two SSDs that failed but did not smoke? I saw your picture of one of them, and I did not see any damage, but it is hard to see much in that picture.

If there is no visible damage, maybe OCZ can be convinced to do a failure analysis on at least those two. If you know how those two died, it would be a big piece of the puzzle in determining whether it was the power supply or the third SSD that caused all the trouble.

Computurd, the way I think about the devices on the power supply is like resistors (in reality, the load may not be totally resistive, but a resistor is usually a decent first approximation). For each voltage, you have the voltage wire and the ground wire which form the posts of a ladder, and the 3 SSDs are like resistors forming the rungs of a ladder. If there was an overvoltage on one of the wires, it should affect all 3 SSDs and I would expect a large amount of current to then flow through all 3 SSD "resistors". On the other hand, if one of the SSDs developed a short circuit (its resistance went way down), then a large amount of current would flow through that SSD, but not through the other 2 SSDs. In that case, the question becomes what happened to the other two SSDs? Which is why I was asking about visible damage.

[josh1980]

There's several issues to deal with, and I've put serious thought into how the whole string of drives could have all failed, with one smoking. I've been rather vocal in this thread(more than most of my comments) because this is a very complex issue.

The easiest explanation would be a failed regulator caused overvoltage on the whole string and the overvoltage protection failed. This cause one to burn and the others to just burn out. It is entirely possible to have one smoke and others not even though the are all identical. It just has to do with the fact that each drive is slightly different. Some might have different type of solder, might have more solder, might have less, etc.

The second easiest explanation would be that the installer plugged in the modular cable backwards. This would put the wrong voltages, grounds, etc in the wrong places. All devices in the string would likely be damaged beyond repair.

Aside from those two scenarios, I can't really come up with any other ideas that are really possible. If the 12v rail had somehow been shorted to the 5v rail(which COULD happen) then the overvoltage on the 5v rail would have to fail, the overcurrent on the 12v rail would fail, and the power supply would take the full brunt. Being that this power supply isn't a small one, it could probably handle that type of current. Of course, the location of the short would be obvious(smoked component/fire) but no other SSD drives would likely have any damage. I have seen this happen from a friend . He plugged in a modular cable backwards. (Who'd have thunk that if you have to force the modular cable into the PSU it might be wrong!? LOL)

The fact that all 3 drives have failed leads to one conclusion. All three drives were subjected to the same "problem", likely excessive voltage. Whether it be user error from improper installation of the modular jack or the PSU itself, excessive voltage wiped out all of the drives. One of the drives failed in a most visible way, while the others not so much.

Undervoltage wouldn't cause the drive to smoke, the drive would simply not start.

Also, there is no way that power would go "up the ground" unless the ground was not attached. If the 12v and 5v were connected, but the ground were not, and they shared a common ground, you would get power from 12v towards 5v. That would be very bad. It would probably damage any components on the 5v power supply, but again, this would be user error with installing the modular cable.

An analysis of the failed SSD drives would only prove an overvoltage condition. It would not be possible to determine the cause of the overvoltage, only that one existed.

I hope I answered everyones questions from the last few comments.. it's getting late.

I think that I'm glad that I don't live next to you. Who's to know just what might happen with one of those catastrophic failures?

Actually, as part of the engineering of the plant, we have to assume components fail all over the place. Any one component failure will not result in a "bad scenario".

[[XC] Oj101]

I will start out by saying that I hear you guys, I am listening not just brushing this off, but let me give my thought process here relating this.

I will say right now that the following comes from my understanding of electricity and circuits. I don't have a PhD, but I have been around.

If 1 of the SSD's was at fault, then there should be no way for the other 2 to burn up.

Example: Let's say I have a toaster, blender, and a lamp on a circuit. The lamp is on and I put some toast in the toaster and some concoction in the blender. The blender suddenly becomes jammed which increases the load on the motor and causes more current draw, the motor then overheats and burns the brushes/coils/whatever before the circuit breaker blows.
The lamp and the toaster are fine because their internal resistance is still in line with operating specs, just because the blender drew a bunch of current, doesn't mean the other devices saw all that current.

All 3 SSD's would have had to have a failure at the same time for all of them to burn, which pushes the odds even farther into the very unlikely.

Now, if I were to go to the circuit box and reverse the ground and the hot wires on that circuit, all 3 devices would fry instantly when they were powered on because current is going to the wrong place.

That is what most likely happened here given that all 3 devices on the string all fried. If 1 of them went bad, the other 2 would not go with it.

Someone is going to come along and shoot holes in this because I used an AC example instead of DC, but just replace my devices with 3 DC devices, it still applies. The other devices in the circuit are not going to see the increased current, which is what does the damage. Revers the + / - and you would have all sorts of smoked equipment.

I'm not saying you're wrong or a liar, just mentioning the way I see it. For three years I worked as a computer technician, and I have seen all to often that a malfunctioning device on a multiadapter has killed the PSU of a computer on the same adapter

[zeroibis]

What is strange is you would think the PSU would have cut the power. Years ago I had the clip on my motherboard that held the northbridge heatsink on come off while the computer was on. The metal block then fell on the GPU however the PSU cut the power and everything still worked afterword. What I still do not get is regardless of the cause why did the PSU not cut off?

[Ao1]

I will start out by saying that I hear you guys, I am listening not just brushing this off, but let me give my thought process here relating this.......

So if Comp got the SSD’s independently tested and it transpired it was a fault on one of the SSD’s would OCZ reimburse the cost of the tests and RMA all of the drives?

[Ao1]

I’m not an electrician but if Comp had his SSD’s wired up in this configuration why would the first SSD not have received the worst damage if the PSU was at fault?

From what Comp has described the worst damage was on the last one on the power line and the damage was then less extensive as you went back towards the PSU.

If something went wrong on the last SSD that caused a voltage surge would this not work its way back up the power line, (via the other two SSD’s) before it could trip out the PSU? Would that not explain why the PSU tripped out, but only after damage occured on all three SSD’s connected by the same power line?

[yngndrw]

What is strange is you would think the PSU would have cut the power. Years ago I had the clip on my motherboard that held the northbridge heatsink on come off while the computer was on. The metal block then fell on the GPU however the PSU cut the power and everything still worked afterword. What I still do not get is regardless of the cause why did the PSU not cut off?

I once had a Tagan PSU where I accidentally shorted the 12V rail to ground. The OCP didn't kick in and the whole PSU died. (The retailer replaced it no questions asked mind, Scan International.)

But it makes you wonder how many PSUs have a faulty OCP system.

I’m not an electrician but if Comp had his SSD’s wired up in this configuration why would the first SSD not have received the worst damage if the PSU was at fault?

From what Comp has described the worst damage was on the last one on the power line and the damage was then less extensive as you went back towards the PSU.

If something went wrong on the last SSD that caused a voltage surge would this not work its way back up the power line, (via the other two SSD’s) before it could trip out the PSU? Would that not explain why the PSU tripped out, but only after damage occured on all three SSD’s connected by the same power line?

It's resistance not distance - While technically speaking the last SSD should have the highest resistance (From the PSU's point of view.) due to the extra cable length and connectors in the way, the internal components of the last SSD must have had a lower resistance.


josh1980: You've gotten me thinking with the unconnected ground thing - What if the ground path ended up being through the SATA data cable of the last SSD - That would also explain the charring on the data connector.

[josh1980]

I’m not an electrician but if Comp had his SSD’s wired up in this configuration why would the first SSD not have received the worst damage if the PSU was at fault?

From what Comp has described the worst damage was on the last one on the power line and the damage was then less extensive as you went back towards the PSU.

If something went wrong on the last SSD that caused a voltage surge would this not work its way back up the power line, (via the other two SSD’s) before it could trip out the PSU? Would that not explain why the PSU tripped out, but only after damage occured on all three SSD’s connected by the same power line?

If you had 100s of miles of cable between the drives, and the drives were all exactly identical(which isn't really possible in the real world), then you could have a situation where the first drive would fry from overvoltage, but the other 2 drives wouldn't. THe length of the wire is so short(feet, not miles) that the same voltage is applied to all 3 devices equally. Regardless of which device on the string was "first" or "last" all devices would see the same voltage.

Theoretically, the second device would see thousandths of a volt less then the first device, and the 3rd device thousandths of a volt less than the second device.

Being that all devices are in parallel(and not series) they would all be exposed to the same voltage.

The location of the burned device on the wire is inconsequential. The position of the short on the PSU wire and the number of drives that smoked is inconsequential. The important fact is that all 3 drives failed simultaneously, which makes it practically impossible to claim manufacturer's defect. The odds of 3 working drives all failing simultaneously when powered on is extremely remote. If an SSD were to fail randomly 1 out of 1000 times, the odds of all 3 failing would be 1 in 1 billion. The fact of the matter is that with no moving parts, SSDs have the same chance of failing when powered up as leaving on 24x7. That changes the odds from 1 out of 1000 to perhaps 1 in billions(because every instant in time is a "chance" for an SSD to fail). The statistical chances of 3 SSDs failing simultanously and it not being SSD manufacturers defect then exceeds someone's lifetime chances.

@yngndrw

I had contemplated the same thought, but I dismissed it because the ground for the data cables shouldn't be electrically connected to the same ground as the power cord inside the SSD. Obviously the PSU would still be ground for both cables, but you wouldn't want the ground in the data cable to actually be a path for power current flow. The ground in the data cable is to help attenuate any signals from the environment and attenuate the signals on the data wires so they don't affect other wires on the data cable.

Now, if there had been a short on the SSD from the power cable ground to the data cable ground, then perhaps you would see what we saw. But yet again, this would be the users fault for spilling water on the drive/getting debris in the drive/etc.

The only 2 possibilities that I see with the situation are user error(not necessarily deliberate), the user using the equipment beyond it's intent and causing the failure, or a failure of the power supply. Since nobody here is going to sit and try to claim Silverstone lied on its reports and it found and fixed a defect(which we could probably never prove anyway) the only other possibilities involve the user.

[Movieman]

Been reading thru this and the one point that was mentioned BUT wasn't really given too much weight in the thread was that these SSD's ran for a substantial timeframe prior to this on another PSU.
I think that is really relevant and needs to be a major consideration.
I also think that it would be wise of OCZ to take a look at these drives and not necessarily for replacement but for knowledge.
Maybe there is an issue,maybe not but if it was me I'd want to know.

[josh1980]

Been reading thru this and the one point that was mentioned BUT wasn't really given too much weight in the thread was that these SSD's ran for a substantial timeframe prior to this on another PSU.

I think that is really relevant and needs to be a major consideration.

Exactly! That's one of the 2 most important facts. The 2 most important facts in this discussion have been that they all have worked fine for months before this incident and they were all attached to the same string on a new PSU and have all failed simultaneously.

[Computurd]

What I still do not get is regardless of the cause why did the PSU not cut off?

it did. once for sure, but it cycled back on. the second time im not sure if i got it in time, or if it tripped again.

The second easiest explanation would be that the installer plugged in the modular cable backwards. This would put the wrong voltages, grounds, etc in the wrong places. All devices in the string would likely be damaged beyond repair.

impossible. they are 'keyed' to where you cant do that

The important fact is that all 3 drives failed simultaneously, which makes it practically impossible to claim manufacturer's defect

sorry dont buy that. if all three devices failed simultaneously and were not on the same wire, then yeah, you would have a point. but they were on the same wire. if silverstone can say that the burning device caused an overvoltatge and/or power fluctuation on the line, then i would think that it is possible, they do design/manufacture/test PSU every day. they know what they are talking about.

I had contemplated the same thought, but I dismissed it because the ground for the data cables shouldn't be electrically connected to the same ground as the power cord inside the SSD

you act as if you personally dismissing that actually MEANS anything. you do not have the device in front of you, you have no idea what happened. you are speculating, you cant "dismiss" anything. you are hardly an authority on the issue, with all due respect, of course
one thing that is not shown on the pictures is that the PCB is warped and bubbled out around the failure 'site' i have tried to take a picture that can quantify this for people viewing this, and the other threads, about this issue. the extent of the damage to the PCB is hard to say that it could control voltage in any fashion.

The position of the short on the PSU wire and the number of drives that smoked is inconsequential.

are you sure of this? what type of electrical engineering training/documentation do you have to prove that? you say you work at a nuke plant and have failures all the time, but that they send the parts off for testing/analysis? i assume you are not the one doing the testing/analysis, correct?

The ground in the data cable is to help attenuate any signals from the environment and attenuate the signals on the data wires so they don't affect other wires on the data cable.

so then, why is the sata port burnt up? short answer: you dont know. this PCB is so damaged, burnt to the point of bubbling/warping, that there is no way in the world that there is any rhyme or reason as to where the voltage did or COULD have went during and after the failure.

It would not be possible to determine the cause of the overvoltage, only that one existed

then you could also not disprove that the SSD was the cause.

If something went wrong on the last SSD that caused a voltage surge would this not work its way back up the power line, (via the other two SSD’s) before it could trip out the PSU? Would that not explain why the PSU tripped out, but only after damage occured on all three SSD’s connected by the same power line?

seems to be my thinking. not sure if it is possible or not.
we have yet to have a qualified person weigh in on the issue, other than SilverStone, who would be biased a bit though

a big issue with me is this, this has happened before. and the RMA was approved for another user under the exact same circumstances. you note that they arent coming in here denying the rma was granted, right? they were claiming prior to that that it was an once in a million never ever happened before in history totally impossible scenario. yet then a user, a member of their own support forum, comes forward with a very similar scenario. are they hiding something? doubt it, but it does show that maybe there isnt a very good QA system over there. . maybe if they had a system they could have looked it up and said "hey there was a past failure that was similar to this, maybe we need to look at this"
i dont think that Ryder knew of the other RMA, seriously i follow my gut on some things, and i dont think he would hide it. hes a stand up guy, and my issue is with his company, not with him. but the point is this: he wasnt aware of it, and he should have been. it is possible that this has happened fifty times, tbh, but with other devices, maybe HDD's burnt up on some rails with SSD's. maybe it isnt getting the attention because it simply isnt being tracked. or someone files RMA with the SSD, but not other devices on the rail. or maybe they have filed it with multiple SSD on the same rail, WHO KNOWS? they should, thats for damn sure. what are the odds that these things being constructed in some shack in china are now going to start failing, en masse? well, the odds of that just went up twice as much, since we now have two cases i dont think that will happen, but im ranting so it.

[Computurd]

[yngndrw]

sorry dont buy that. if all three devices failed simultaneously and were not on the same wire, then yeah, you would have a point. but they were on the same wire. if silverstone can say that the burning device caused an overvoltatge and/or power fluctuation on the line, then i would think that it is possible, they do design/manufacture/test PSU every day. they know what they are talking about.

You should ask them to go into more detail about their conclusion and if possible get them to setup an example situation where they can show what they think happened and it's effect.

[Computurd]

here is what they originally sent me after testing:

Hi Paul,

Here is what we find out on our side when testing the PSU and cable

This issue was caused by a large over current which could be cause by the PSU or HDD.

We have checked for any damage components inside the ST1500 and found nothing wrong, the power supply and cable are perfectly fine.

The cables with the burned mark on the +3.3v wire was also tested and work fine it didn’t damage the hard drives we connect to it
The scorching was only on the surface of the wire not inside the connector.

And we tested “Short Circuit Protection” of this ST1500 PSU and it works fine.

So we confirm that this short circuit was not caused by the PSU side.

We don’t know what caused this problem but we can tell you that we are very sure it was not this ST1500 power supply.

Please see the attached Report

Thank you


Joel
Silverstone Tech Inc,

i added the underline

[josh1980]

impossible. they are 'keyed' to where you cant do that

They may be keyed so that you can't do it, but I guarantee you that if you force it a little, it'll go in. And the PSU will still power up and you WOULD fry your stuff. I have been curious of this ever since modular PSUs hit the market. They are very easy to plug in backwards. I have tested this myself because I was curious. I wanted to see if it was possible to accidentally install the wire backwards into the PSU. It's much easier than I thought and I have since made the decision that I will ALWAYS make a visual verification that the cables are properly connected to the PSU before I use them. The little teeny piece of plastic that makes a few of the holes circular and square is VERY easy to force in. I realize they are keyed to only work one way, but I guarantee you that you can still force them in the other way. (Of course, if you only have 5 wires and they are 2 rows (3+2) then reversing them wouldn't really be possible. All of my PSUs have been either 1 row of 5 pins or 6 pins in 2 rows of 3.) Since I did do my homework and looked at pictures of your power supply online , they are all 2 rows of 6 or 8 jacks(minus the 23 pin atx plug of course). This means that I can assure you that you might accidentally install them backwards without knowing it until it was too late.

sorry dont buy that. if all three devices failed simultaneously and were not on the same wire, then yeah, you would have a point. but they were on the same wire. if silverstone can say that the burning device caused an overvoltatge and/or power fluctuation on the line, then i would think that it is possible, they do design/manufacture/test PSU every day. they know what they are talking about.

You need to understand the difference between same wire and same circuit. If they share the same wire any voltage applied to any wire will be applied to all devices on that wire. Some PSUs have a separate regulator and protection for each 12v bus. If they separate the circuits by modular plug then all of the devices that are powered by that plug would have the same voltage applied.

Some PSUs have 1 very large rail for 12V. 1 huge 72A+ rail for all of your devices is not the "safest" for your components. Quality power supplies that are safe for your components shouldn't have 1 big rail. Of course, companies cut corners and sell PSUs with 1 big 12V rail and call it a feature! Check out this link.
http://www.silverstonetek.com/produc...1200M&area=usa

That shows you 2 designs for the zeus series by Silverstone. Anyone with a clue as to the inner workings of PSUs would NOT purchase a single rail 12V if they want overcurrent protection. PERIOD. Would YOU put 95A(1180W) through one single device on your computer? But what does Silverstone say on their website? "Class-leading six or single +12V rail(s) with 95A". You are NOT class leading if you have 1 single rail.

Anyway, back to your PSU...

From the silverstone website:


1500W continuous and 1600W peak power
100% modular cables with cable bag
Efficiency 85%~88% at 20%~100% loading with 80 PLUS Silver certification
Class-leading eight +12V rails with 110A (Peak up to 120A) (bold is mine)

Do YOU know how to figure out if that is 110A rail or 8 rails that total 110A without reading their bullets and believing what they say?

I'm sure at this point you ARE now checking to see if you have 1 rail or multiple. I'll tell you.. If you look at the specs on silverstone's website it shows 8 rails of 25A(30A peak).Some of the larger PSUs have 1 big fat rail for all of the devices on your computer with no overcurrent protection for that rail. Why no overcurrent protection? Because at 72A+ you'll end up with a small fireball/smoke if you really try to put 72A through any one device. You'll melt wiring in your case. It'll be pretty darn obvious what failed. I'm not talking small fry "smoke" from an SSD either.

The fact of the matter is that since you were on 1 modular wire, you could not have possibly had more than 1 regulator for the 12v and the 5v respectively. If one of the devices had smoked and the others were fine then it would be easy to blame a random failure of the one smoked drive. My power supplies(all the same model OCZ PSU) are 16 AWG. That gives the cables a rating of 22A max design. Since his cables didn't melt off their insulation it's pretty reasonable to say that you did NOT hit the overcurrent protection limit either. So your power surge didn't exceed 22A(or if it did, not by very much). I have no idea where their overcurrent protection is set, but I can tell you it's likely more than 110% of rated load. It's definitely more than 30A(the peak current rating).

Since all 3 devices are now non-functional all of this information leads to a likely voltage issue. Either the PSU provided the wrong voltage or a ground was not attached so power tried to find an alternate path. Silverstone claims the PSU is operating fine. Since everyone across the world is not having the same issue it's safe to say it's not an inherent design issue. However, there could have been a manufacturing issue. Silverstone also claims that the PSU is operating fine, so this again isn't likely, or they lied.

The SSDs may have some components that would might need a higher voltage than 12v(unlikely, but I must consider this possibility for the purpose of this thread) and those components were to be shorted to ground, the associated drive would burnout but the other drives would not be affected. Since all 3 drives failed this is not a possibility. Also, the SSD doesn't have a massive heatsink on it bigger than your PSU's internal heatsinks. So the SSD is not capable of generating a voltage and then applying it to a cable in a high enough quantity to overcome the voltage from the PSU itself and burnout the other components. If one of your SSDs had tried to apply 15VDC to the 12VDC wire, then 3VDC differential would have created a current surge that would have burned out whatever small circuit in your SSD that was trying to output 15VDC. The only way one regulated voltage is overcome by a different regulated voltage is by having the supply from the later MUCH larger than the former.

you act as if you personally dismissing that actually MEANS anything. you do not have the device in front of you, you have no idea what happened. you are speculating, you cant "dismiss" anything. you are hardly an authority on the issue, with all due respect, of course
one thing that is not shown on the pictures is that the PCB is warped and bubbled out around the failure 'site' i have tried to take a picture that can quantify this for people viewing this, and the other threads, about this issue. the extent of the damage to the PCB is hard to say that it could control voltage in any fashion.

When I read this thread I considered everything that could have resulted in your unfortunate mishap. I have tried to provide the most accurate information I could and tried to consider all possibilities as long as we aren't dealing with multiple simultaneous failures. The list of 2 or 3 possible simultaneous failures would be quite large and beyong the scope of being able to discuss in a thread. The board warping would be commong for a PCB that has overheated. Not a particularly big surprise considering we know there was alot of heat at the site. I dismissed the notion of the grounds being connected because you do NOT want regular current flow through the ground of the SATA cable. Why do I know this? Here's a link to the theoretical "faraday cage".http://en.wikipedia.org/wiki/Faraday_cage Anyone with an electrical background would realize that a faraday cage isn't a faraday cage if it was designed to have current flow through it(from my link, it states "Faraday cages cannot block static and slowly varying magnetic fields" and "static" would refer to designed constant current flow). A faraday cage is used to block out any possible signals that might interfere with the signal in the data cable. So theoretically, if the ground on the data cable WERE an intended path for current flow, it would no longer be performing its duty as a faraday cage.

are you sure of this? what type of electrical engineering training/documentation do you have to prove that? you say you work at a nuke plant and have failures all the time, but that they send the parts off for testing/analysis? i assume you are not the one doing the testing/analysis, correct?

Honestly, I don't think I have to prove a damn thing to you as to my background. If you want to reject my opinions, feel free to. At the end of the day I won't lose any sleep over your dilemma. I was only trying to help YOU out. But, since I'm not a complete jerk I'll fill you in my background. If I don't know what I'm talking about, I don't comment. Simple as that. I know ALOT about electricity and power and how to use it. I could give you very detailed explanation as to what a "switching" power supply is(heck, I could make switching circuits for you if I had the parts!) but don't ask me about what metals are acceptable for liquid cooling. I couldn't tell you if those pretty blue/pink/green liquids that are $1.25 for 4 oz are better than tap water for corrosion. But.. if you insist.

As a child I would read my mother's books on electrical theory. In 6th grade I wrote a 6 page essay on electricity. In high school I began to tinker with things and experimenting with the different faucets of electricity and owned several of those "build your own circuits". At 18 I enlisted in the United States navy and attended 3 schools totaling 2 years in the Navy Nuclear Training pipeline where I graduated in the top 10% of my class. The Navy Nuclear Training pipeline is considered to be MIT level of difficulty, and successful completion of the pipeline is a very big deal. Failure rates are up to 33% for each of the 3 "steps". Upon completion I received a license to perform my duties as a Navy Nuclear Operator. Training costs are typically $1 million per student.

I then spent 2 years on nuclear submarines as an electrician, responsible for maintaining all electrical equipment on the sub. I was responsible for all electrical generation, distribution, and maintenance on said equipment. Then I spent 2 years on shore fixing the same equipment on submarines. Then I spent 2 years working on some of the Navy's computer systems. Then I spent 2 additional years working on a carrier, responsible for electrical power generation, distribution, and all maintenance associated with the equipment. This was everything from the generator itself on down to the plugs in the wall for your laptop when you are underway.

After getting out of the Navy I worked for a brief time as an IT consultant for the US Navy as a contractor, then worked the last 3 years at my current job. Currently I am responsible to research and write work instructions to fix all instruments in a nuclear power plant. I deal solely with the theoretical aspects of performing the work, minimizing the cost for parts, minimizing the number of extra parts we buy(by determining the most likely cause for failures), and the human performance aspects of the work that needs to be performed.

So to put it short, I have a rather extensive(and expensive) education under my belt. Although I have no degree(yet) I make alot more money than some people with EE degrees because of my extensive background. Companies care about the degree, but they also care about how you've applied your education.

so then, why is the sata port burnt up? short answer: you dont know. this PCB is so damaged, burnt to the point of bubbling/warping, that there is no way in the world that there is any rhyme or reason as to where the voltage did or COULD have went during and after the failure

See my explanation for faraday cage above. Answers to your questions are provided above. You are right, there's no certainty for what caused the SATA port to burn up. Which, I did, in fact explain on my post when I gave possibilities. But thank you for reiterating that I don't "know". I've never proclaimed to "know" the answer, only provide several different scenarios that are the most likely to have caused it.

I do know why your SATA port burned up. High current was applied! (this is a joke for the most part)

[QUOTE=Computurd;4625288]then you could also not disprove that the SSD was the cause.[QUOTE]

I can disprove that the SSD was the cause. I discussed this above with the SSD generating a higher voltage for components. I find it extremely unlikely that the SSD actually needs higher than 12VDC, but I examined the possibility just for you .

a big issue with me is this, this has happened before. and the RMA was approved for another user under the exact same circumstances. you note that they arent coming in here denying the rma was granted, right? they were claiming prior to that that it was an once in a million never ever happened before in history totally impossible scenario. yet then a user, a member of their own support forum, comes forward with a very similar scenario. are they hiding something? doubt it, but it does show that maybe there isnt a very good QA system over there. . maybe if they had a system they could have looked it up and said "hey there was a past failure that was similar to this, maybe we need to look at this"
i dont think that Ryder knew of the other RMA, seriously i follow my gut on some things, and i dont think he would hide it. hes a stand up guy, and my issue is with his company, not with him. but the point is this: he wasnt aware of it, and he should have been. it is possible that this has happened fifty times, tbh, but with other devices, maybe HDD's burnt up on some rails with SSD's. maybe it isnt getting the attention because it simply isnt being tracked. or someone files RMA with the SSD, but not other devices on the rail. or maybe they have filed it with multiple SSD on the same rail, WHO KNOWS? they should, thats for damn sure. what are the odds that these things being constructed in some shack in china are now going to start failing, en masse? well, the odds of that just went up twice as much, since we now have two cases i dont think that will happen, but im ranting so it.

All I can say is that at the end of the day, there's only 2 possibilities. The PSU is faulty or user error. Do I really care which one it was? Not in the slightest. If it were the PSU then it was a fluke(a sucky fluke). If it were user error, then again, I'm not affected. I haven't made the same mistake nor heard of anyone make the same mistake until now.

One realization needs to be made in this thread. There's too many possibilities to negate them all. But, there's a few possibilities that are so incredibly unlikely that money and time shouldn't be spent trying to find out the problem. That's where OCZ is. The fact that all 3 drives have failed at the same time means they all likely share the same common mode of failure. Is it likely that 3 drives all had manufacturing defects that all decided to show themselves all at the same time? Heck no. Is it more likely that it was user error, bad PSU, or something of the like? Yes. They don't want to spend big bucks to tell someone it was one of those choices. Anyone that has a background in this stuff will likely come to very similar results.

I'm sure I'm going to get quote a bunch more times, but I'm not going to continue to read this thread. What I see is a bunch of people that don't have a background in electrical theory discussing how OCZ is at fault. I understand that as the end user it sucks. I'd be trying to figure out what happened if the same had happened to me. I'd also accept the fact that either I screwed up the connections and reversed the modular connector or the PSU was bad. If I had decided that the modular connector wasn't backwards, I'd never use that PSU again.

I really have better things to do with my time that try to argue with people that don't really know what they are talking about. I'd love to sit and explain in great detail all of my thoughts on this topic and instead educate people on all of my thoughts on this, but I think there's a limit to the size of posts. People fall into 2 categories in this thread. Those that have an electrical background and those that don't. If you don't have an understanding of electrical theory you shouldn't try to guesstimate what you think happened. Leave it to the big boys(and OCZ/Silverstone) to tell you. If they don't care that's entirely their perogative. If you care that they don't care, stop buying OCZ. On the other hand I know when I need to eat my losses and accept that I friend a component, and I know when I need to fight the system and demand new hardware and investigations.

-Josh

PS - Feel free to read these myths on PSU... They're very educational for those interested.

http://www.tomshardware.com/us/press...ower-supply-20
http://www.tomshardware.com/us/press...-supply-PSU-23
http://www.tomshardware.com/us/press...ower-Safety-26 <-- This one gives some more explanation as to why single rail power supplies = bad.
http://www.tomshardware.com/us/press/PSU_cooling_fan-28

For those of you that read those myths, you'll probably never look at power supplies the same way again

[yngndrw]

My power supplies(all the same model OCZ PSU) are 16 AWG. That gives the cables a rating of 22A max design. Since his cables didn't melt off their insulation it's pretty reasonable to say that you did NOT hit the overcurrent protection limit either. So your power surge didn't exceed 22A(or if it did, not by very much). I have no idea where their overcurrent protection is set, but I can tell you it's likely more than 110% of rated load. It's definitely more than 30A(the peak current rating).

I'm going to respectfully call you out on that part.

14 AWG cable would be able to take the surge current for a short period (I.e. The length of time it takes for the OCP to kick in.) without melting its insulation. Therefore I don't think that shows that an OC condition didn't happen.

The SSDs may have some components that would might need a higher voltage than 12v(unlikely, but I must consider this possibility for the purpose of this thread)

I believe that might actually be a possibility as from memory I think it requires a higher voltage (~13V) to erase the memory. At least when erasing / programming PIC microcontrollers you need a higher voltage.

[Computurd]

first, read yngndrw post above you (post 74) that in a nutshell negates half the bluster in your post. i have no idea why you are going on long tangents here about single and multiple rails. you are seriously off-topic here. this is a multi rail psu. simply inflating your post to make it look big and important is not doing any good when it gets down to the meat of what you are saying, some of which is VERY flawed.

This means that I can assure you that you might accidentally install them backwards without knowing it until it was too late.

im not stupid would be my first reply to that. if you did your homework as well as you think you did, you would see there is a long prong on one side of the modular connector. that prong CANNOT be installed backwards, it is impossible, as i said. there is not a hole on the other side that would allow it to be plugged in. the prong on the side of the connector would not slide in.
i have made this simple for you, even modifying the pics, so you can get it. OK the plug on the side of connector HAS to go in in order for it to be plugged in. there is a hole on one side, NOT on the other.
Physically Impossible To Plug This In Backwards.
that eliminates your "user error" angle. i really must read the rest of your post before i reply further, but we will start with this:

. I'd also accept the fact that either I screwed up the connections

why, it cant be done

I really have better things to do with my time that try to argue with people that don't really know what they are talking about.

like people who actually have the equipment in front of them, and can tell you cannot plug it in backwards?

On the other hand I know when I need to eat my losses and accept that I friend a component, and I know when I need to fight the system and demand new hardware and investigations.

couldnt possibly be user error. impossible. you are making a string of guesses all based around one very very flawed theory: that its MY fault and i connected it incorrectly.