Basically, you proved my point. You have the OPINION that the Google study is Valid. I choose not
to believe it. I tell you to believe what you want.... instead of trying to argue with an OPINION.
But, what I get is Personal Attacks and Insults merely because I choose not to believe in
what you believe.
I have no problem when you disagree with me. I only require that you came with some proper facts or at least some reasoning why Google should be wrong. There is no need to start insulting me like I'm some idiot who believes everything on the internet. I'm not someone who goes around introducing people to the concept of 'death trains'.
I run my own company and in that capacity I oversee computer hardware with in total around 30 hard disks. Most of them running 24/7. With a fair share of harddisk failures over the years and several high availability applications running, I have a higher than average interest in the subject. I monitor the temperature of most hard disks continuously with emails going out if a disk overheats.
Actually none of that is really important to the debate, it has nothing to do with my opinion either. Even if I had agreed on your conclusion I would have posted the same. In fact I don't think we really differ that much on the conclusion, although you keep misinterpreting my opinion. All we differ on is that I'm taking Google's take that cooling disks below 37 degrees Celcius is worse and not better for your disks.
What I did do was tosimply explain that your study is flawed and that a 8 disk test set is not big enough to draw any useful conclusions from. Anyone with basic education in probability calculations could tell you that your findings are not significant. No other experience or credentials are needed for that. if you make false claims then you can be sure that people will correct you. BTW I'm talking about your suggestion that your home run test has any weight and not on whether disks perform better at room temperature.
Sure you can have your own opinion, but if you start trying to persuade people that your opinion is right, you better have something more to back this opinion up. What have you offered besides insults and assumptions that Google could conceivably be lying. Of which again you have not been able to give us even a shred of evidence.
Besides there is plenty to talk about iof you want to discuss opinion without resorting to insulting me. For instance, you say that I should "hit the books" if I'm stupid enough not to think heat has an impact on the workings of a hard disk. I'll let it slide that I never claimed that temperature has no effect on hard disks, but let me bounce that argument back at you. With the tolerances and engineering of hard disks, there will obviously be an optimal temperature where they function best. Make it too hot and things start to run out of tolerances, but this also happens if you make it too cold. If the disk is running at the optimal temperature the life will be longest.
Now how does this optimal temperature get so high? I suggest that it is because disk manufacturers measure hard disks performance in a 42 degree oven (in an effort to speed up it's aging). Makes sense that that this testing behavior actually steered hard disk tolerances to conform to this test methodology and thus that hard disks perform better around this 42 degree test temperature? Besides, leave a harddisk spinning in a standard case and it will reach a temperature of around 40 degrees on it's own (plus or minus 5 degrees or so). So from that observation it makes perfect sense that disks work best in that temperature range.
So we have:
- Google ran statistics on data on their 100,000 harddisks over years and found that the optimal operating temperature for a hard disk is around 40 degrees Celsius (on average)
- Google shows that cold disks actually break more often than slightly hot disks
- Google has the only reliable real life data/research on this subject either of us could find
- Google is a renowned organization and the people who performed the tests are well known scholars. Neither has a financial interest in telling people to stop cooling their disks.
- Both the test temperature at which reliablity tests are performed and the natural temperature that a hard disk will reach in a standard PC case is around this optimal temperature.
- Maybe most importantly, their findings simply make perfect sense
On the other hand we have ... well nothing actually:
- You had some disks fail that were searing hot. This complies with Googles findings and common sense that you should not run a disk outside of the manufacturer reported maximum temperature of around 55 to 60 degrees Celsius.
- You have had 8 disks running for 3 years with no disk failing. There is a calculated 52% probability of this happening so that's hardly a significant finding. Besides the test set is too small anyway.
- You have done no tests on different operating temperatures within hard disk operating limits. Let alone a test with hundreds if not thousands of disks to be able to get any significant findings. So you have no findings proving that Google's research is incorrect.
Again, temperature IS an issue with hard disks, but in a normal application there should not be a problem. On the other hand, if you modify your hardware in some ways, like for instance decoupling the disk to get rid of the noise, one should really keep an eye on the temperature since it might overheat the disk. I warned of this a few times already.