How to make accurate failure check

1. Freeze Event, 2 in a row
Didnt have an issues for several months, and then recently i had a freeze in the middle of the desktop, all i was doing is downloading something to a non system drive. Now, when i restarted another freeze right at the desktop.

2. BIOS
Then i was going into BIOS and checked if there is any unusual health data, temps or what else.. there is nothing critical.

3. Safe Mode:
I was going into safe mode and checked any kind of event logs if there is any message about a failure, but the only stuff logged was when the shutdown was unexpected and the lack to start important system tasks right after my freeze. There was aswell a entry about a device not working, but i have no clue what device. The GPU doesnt behave like that, in fact only the most critical stuff may cause a instant freeze, such as the OS drive or RAM (but the RAM never had issues and its VERY unlikely that they turn bad all of a sudden) The software is 99% clean, i did a full virus check with the newest Kaspersky a few hours before the event took place and i always browse with highest security settings, so its close to impossible to have bad datas.

4. Another system start, now all went well.. such as nothing else ever did happen. Super weird. Instability somewhere else...

5. I checked my GPU driver list on system drive, now the icons for the drivers did load very very slowly... and it took a full min in order to REACT (not responding). I wanted to check information of that file... i had to wait a whole min.

Aswell made few other tests:
CPU Intel Burn at Normal mode: 100% load, max. value 75° (totaly green for a SFF)
GPU Witcher 2: 100% load, no issue found

6. Logical issue seems to be the system drive, that drive is a M4 128 GB and not even half a year of age. But its the only issue i found and right at that moment it works proper... but i know those drives, they slowly will burn down in many cases.

Now i think it must be the drive... the second SSD causing failures it seems, and both was a system SSD, which is twice that worse. However, best i can do is to check that drive for failures, i just dont know how to execute... anyway, kinda sad that those SSDs have a longlife endurance of close to none because of either bad firmware or a dying controller. But even, when i try to get warranty, what else should i say.. it may work when they are being investigated. So i have to hunt down the failure.

I would check to see if the SSD has a firmware update, or perhaps your system's SATA controller card might need tweaked. Do you have an external SATA card to rule out the onboard SATA?

It would also be interesting to plop the drive in another (Known good) machine and see how it fairs, if you get the same freezes then you know its the drive. Also I would imagine an ATTO benchmark or two would make the problem glaringly obvious if it was the drive.

I have no SATA controller card and in fact my internal controler is still non native so it will top out at 300 MB/s however, in usual conditions a SSD wont deliver much more than that since not all data are served in a way in order to let them reach the 500 MB/s value, so i can live with.

And i dont think i can attach a additional SATA card because all my slots are either taken or unable to access to it (i got 2 PCIE ports which are both taken). The only slot i have left is a internal Mini-PCI-E@PCI-E x1, but this is probably not fast enough for a SSD. Its usualy used in order to upgrade that SFF with a wireless LAN for example (it already got teaming LAN, so natively its not included because those are usualy made for performance).

Well, dunno if any PCIEx1 for mini does even exist, but gonna check out. In theory i do need a adapter including a controller chip which can convert a SATA SSD over a PCIEx1, but thats like asking for rocket tech... why should that even exist.

Finally why cant the stuff just work proper... we kinda invest endless ressources into mobile tech but bigger PCs are slowly degenerating.

Also i currently have no second PC to test it out with. Maybe in 1-2 years but thats to late. My old PC died because of burned capacitors... and the even older one is that old that it does lack SATA 2 (so its useless).

ATTO bench run (and aswell Crystal Disc) didnt reveal any sort of issues.. and results are pretty normal for a non native controller. Although did reach close to 300 MB/s several months ago and now it dropped to 280 MB/s but that could be possible since the drive can have some wear. I can try out the windows tool but that need non desktop mode for OS drive.
http://imageshack.us/photo/my-images/831/atton.png/

Windows tool didnt deliver any failure detect aswell. Its just truly weird.

Ofc when i got the machine i didnt know how much of a issue freshly developed stuff can be. The SATA 3 is only a little bit faster than a SATA 2 but finally they are all controller limited and the controller development was slower than the SSD development. The "native" label was only here in order to lower confusion but in fact, a controller unable to deliver SATA 3 is simply not a SATA 3 controller... easy as that. But even the USB 3.0 Asmedia port never truly worked. A few times it seemed to work but its heavely unstable... which is probably a driver issue but i lack to find a working driver.

Well yeah... i never ever buy half baked stuff ever again. Have to realize the fact that nowadays portable systems are the big runner and supported much better, however, even the IPHONE 4 GS had many issues, we simply dont build with quality anymore... its all a matter of harsh economic rules.

Finally i just can add, if there is instability at all its extremely hard to detect nowadays, its just a mind buster because there can be endless amount of sources and no tool can detect it.


But i kinda found out that some weeks ago i turned my BIOS to standart settings because the AHCI is kinda having some issues when there is a USB drive attached. Its mixing stuff and can have issues at booting.

So thats why i did reset to standart, and it worked again. However, i forgot to disable spread spectrum and C1 state, those stuff can indeed lower the stability, but its no issue as long as there is no sensitive device inside such a build. A SSD simply is one of those sensitive devices and its totaly possible that it can cause the SSD to get unstable.

So i turned it off now and so far no issue anymore, but i truly hope it is solved yet. The issue surely is caused by SSD, however, its not as easy as that, because the SSD (unlike the previous one which TRULY was faulty) is not necessarely faulty, its just extremely sensitive to unstable frequency-matters...

In term a SSD is faulty.. the OS might show a lot of bluescreens, if it does never show any, it can be the case because of a instability which will be affecting the SSD. However, its a instability the rest of the system can handle, however, not the SSD. It aswell does describe why the load is no matter at all, because the issue is caused by a "ground bounce" which is always working in the background, no matter the load.

Anyway, the entire matter is advanced knowledge by experts, there is no tool able to detect, thats the biggest problem. So its hard to say how to counter.

But it does remember me of my second SSD, i made a backup there. Now recently the scanner had a issue scanning it, because it apparently was corrupted. So how comes that a SSD got corrupted files at all? The answer simply is, because there was some sort of instability somewhere, and i guess i found the issue. Its aswell the biggest SSD issue, because a HDD can handle it, while a SSD cant. A SSD is simply much more sensitive to "how the data is transfered". Flashing and reading a SSD is a very sensitive process.

So everyone make sure every single spec is 100% stable, else the SDD can become the biggest foe. And generally for a portable drive i recommend a 1.8" HDD, they are extremely small and much more reliable in the long term when used with endless amount of different (and maybe not fully stable) devices!