Originally posted by: bryanlMindless1: I also noticed some cables spliced into a single wire. This looks improper, but does it really matter, considering that each single wire is only an inch long?
Not necessarily, but considering the overall construciton quality, it would be expected this spliced connection is no better, the connections themselves are more likely a
potential problem, one that shouldn't have existed.
I did see flux all over inside some of the Ultras, and oddly the solder side was more likely to be clean than the component side. And in some power supplies the flux was even on top of the transformers (it was definitely not transformer varnish). The circuit board seemed to be of the same material and thickness as those in quality supplies, but I didn't examine the thickness of the copper (couldn't).
Fiberboard (paper based, FR2 or ??) can be different density while same thickness, even if it is too much to expect FR4 or similar fiberglass you would (hopefully) insist on in a > $140 supply. You can see their workaround for the thin copper by noting where they added jumpers on top and massive amounts of solder flooded on the top to merely aid conduction- no parts account for this beefing up of the traces.
Fan quality seems to be a problem, according to user reports, and one of my Ultra V-series had a fan that would not start reliably.
The fans started on the ones I had out, though I did note one's fan leads looked like they could've gotten into the blade area and kept it from spinning. My more significant gripe with these sleeve bearing fans is they loose lubricant within the life of the PSU, especially horizontally mounted.
You mentioned an Ultra with a +12V output rated for 20A that was connected to a STPR1620 dual diode rated for just 16A. The datasheet for this component gives 2 current ratings: 20A RMS and 16A average (delta = 0.5, temp = 120C). Which is more relevant here?
16A, further derated for temp (made worse by TO220 size package, ~silpad, and thin heatsink) but actually it may not apply to this PSU, I don't recall if the 500W has same diodes in same position or not. I would hope not but the way they're rating for peak output is still a concern, reveals some of their methodology for the rating system in general. Another factor for a good ~ 500W PSU is you'd expect a TO247 sized part for the most loaded rail, as it not only handles more current by virtue of the larger die inside, but the larger heatsink tab speads that heat better. Plus it wasn't a Schottky, but again I don't recall on the 500W.
I'm plan to use this 500W Ultra to replace an unusually light 350W one included in a very cheap case. It will power a computer that has never drawn more than 280W from the AC outlet.
That doesn't tell us what the current is per rail, but a crude guesstimation could look like 70% efficiency, 196W out, you might have 15A or so peaks on 12V rail? Too many variables to know this, I'm just throwing around some numbers and peak may not matter as much as durations- any system can peak high but some systems have a fairly high sustained load for gaming, etc. Sustained load will determine how much loss there is in total, the amount of derating necessary from heat- so as always keeping it cool(er) through good chassis airflow will help.
I think it'll run that system for at least awhile, if there aren't any overt defects but it's not what I'd choose either, unless you're just overclocking the heck out of a Prescott to get a little more life out of it. It's quite likely better
capacity than the came-with-case 350W PSU (a brand you didn't mention), but there's the remaining variable of whether the construction issues end up causing any problems.