Last year I went to Interop in Vegas to take a look at what was considered the newest and greatest. I came back horribly disappointed at the fact most companies were in the business of selling service plans and warranties galore while neglecting the hardware they pushed, sometimes neglecting the hardware entirely relying on a 'magical cloud' ideal. This ideal has failed before and it will likely fail again, time will tell. Many booths sent their sales engineers alone, who could off the top of their head list every single certification and test that their product "passed" but not tell me what their product did, how it worked, or what it was built with/made of. On the semi-rare occasion I'd get someone who would know a little, they could at most tell me who they sourced the power supplies from, eg sparklepower, but not tell me what they had ordered the supply built to or with, after all that was proprietary information! I ran into about 5 or 6 companies that I could deal with out there and that knew their stuff, and knew that reliability was something that people were in the market for, rather than a "we'll get it fixed when it breaks" attitude. This leads me to the strange place that I usually don't buy a piece of equipment before I tear it down and look at it's individual components. I want the best chance of something lasting it's full service life, disregarding power failures, fluctuations, and natural phenomena. Every year that passes it seems like what I ask is unreasonable in some way, even though I know that I'm not the only one thinking in the same lines. People think I am crazy if I tear down an eval unit to see what it's made of, and think I am crazier if I send a unit back if it were not built like it was promised to be, but I think that folks should get the use out of say a $15,000 router, and that it should still be usable long after it's service life had passed. In more cases than not, with the economy the way it is, they'll likely be using that same piece of equipment long after they were going to replace it.
This leads me to testing equipment and such, and I find it's amazing how many failures you can eliminate by paying attention at the beginning. Some failure modes only show with time though, and only experience will show that. No environmental chamber or accelerated testing procedure seems to fully replicate just what aging and constant usage at it's real world capacities with real world dust, real world lazy people, and strange real world environmental swings can do to a unit, nor does anyone ever have the time to test something in such a way. One of the most frustrating things that happens to me is when I get something that fails in a predictable manor 2 or 3 years from the purchase date, because usually by then the issue is terminal in whatever other identical units I own. I learned this the hard way running a folding@home farm with power supplies and cheap motherboards. It's one thing to be blindsided by one PSU or component, but it feels like your are a deer trapped in the middle of a railroad track with constant scheduled freight deliveries forced to re spawn and experience death several times over when these failures mount.
I've lurked at this site for a few years, and keep rediscovering it every so often, and I have to say, I just love it when you tear the PSU's apart, seeing the engineering, or lack there of, that goes into a product before you buy it is great, knowing the pitfalls of one product over another is great, and knowing what I should expect when I peer through a fan grill is also wonderful. Thank you for all you do, the only thing you are missing is a time machine to simulate what aging does to the power supplies