Well, either it works or it doesn't. How did you determine that it was the cause of your motherboard and graphics card failures? Maybe it was the graphics card failing which fried the motherboard and caused the PSU to temporarily not work (it has built-in protections). Alternately maybe the motherboard died, causing the graphics card to die (since it supplies power to the card through the PCIe slot) and causing the protection circuitry in the PSU to temporarily pop.
This is true. It would be very hard to determine a definite answer as to why a series of failures happened. At the very least, I do not believe the PSU is in perfect working order, but that belief is all I can have in this situation; and I doubt any "free" failure analysis could be in depth enough to determine if it is in "perfect" working order such that the PSU could even be returned to the client. I don't expect corsair to throw millions of dollars at analysing a single PSU to even come close to a definitive answer.
I didn't give any history. When I first got the PSU, it wouldn't turn on. Whether in the system using the motherboard or just using the paperclip. I didn't feel like doing the RMA at that time, so I just left it alone for a week or two. When I had a little more patience, I decided to go ahead and start the RMA process, but I needed to test it again. It turned on with the paperclip, so I popped it in a system and it seemed to be working fine. This was probably the biggest mistake. I should have simply returned it anyways letting them know it would not turn on when I first got it but seemed to be working now.
I had one or two instances over the next 18 months where the computer turned off and wouldn't power on for a little while, and testing with the paperclip, the psu would not turn on. One of my backup PSUs would boot the system in each of these cases. I certainly take responsibility for being lazy and not RMA'ing the PSU.
In the final failure, the system turned off while I was busy with something else, so it probably sat there for an hour or so. When I went to turn it on, sparks came from what appeared to be the ATX motherboard cable, and the system didn't post (should have posted during the 6 seconds it took me to power the system off).
Now obviously my observation on where the sparks came from is subjective, but I'll explain why I say it
seemed like they came from the motherboard power cable. Judging by the trajectories of the sparks, the origin was out at least several inches from the motherboard. The origin also seemed to be far too high for it to have come from the GPU (the only other thing sticking out that far that could have sparks). Do I know for a fact that it came from the cable? Absolutely not. Even if it the sparks did come from the cable it doesn't mean the PSU was the cause. I inspected the cable to look for burns or tears in the shielding (a black mesh), and nothing I saw couldn't be explained by simple wear. The origin of the sparks is completely subjective, so I cannot say for certain.
What I do know is that after it happened, the PSU didn't turn on again with the paperclip test, but did a little while later (could fall into your PSU protection scenario). Aside from laziness, the only other thing was the knowledge that if I sent it back, it would probably work when they get it, they'd send it back, and I'd be stuck with the shipping charges. Granted, 15-20 dollars shipping + a new power supply in the same quality range would probably be less than the motherboard and GPU.
The only case where your suggestion that the PSUs in-built protection should not have been the cause is the first one. Each of the other times it could have been that, but it also could have been a number of other possibilities.
We can easily poke holes in assumptions in cases where all we can do is assume. You can take any failure situation and say it could have been anything, and company X's product was not necessarily at fault. Let's say the PSU was found to have failed, and in the way that it failed it certainly was possible that it damaged other parts connected to it. We could then say that the power source was at fault -- that the surge protector or UPS was the culprit. Even in these cases where the cause seems to be obvious at face value, it certainly isn't. Without providing a clone of this little environment in some sort of mini-universe, and analyzing the event as it happened, it would be impossible to even begin finding a real cause.
I completely support your skepticism on the assumption that the PSU is the cause on the grounds that it is practically impossible to determine a cause with any conclusive certainty.
So of course I don't know for certain, but I certainly would not put the PSU in another system if they returned it. I should have never put it in, and after the first time it turned off while in the system, that should have been confirmation that it wasn't a fluke and something was wrong with it. In my opinion, it wasn't "either it works or it doesn't", it was "it works most of the time, and I can only hope when Corsair gets it, it doesn't work".
