On Sat, 2015-04-04 at 13:11 -0600, Pete Travis wrote: > After viewing lots of graphs from really serious PSU testing (ie > jonnyguru.com), I decided that "PSUs are most efficient at 50-60% > capacity" was a good enough generalization. If you work out the > maximum draw of your system, add around 40%, and find a quality! power > supply in that range, you're set. Yes, it's very hard to estimate how much power you will need, especially as you don't have proper specs, or any specs, on your hardware. None of my hardware ever came with any, and I place dubious faith with any technical details that you get from things like the data that dmidecode retrieves from your hardware. You really need to know peak and average power requirements, and from which supply. If you happened to have something that said it needs 12 watts, you need to know whether that's 12 watts from the 12 volt or the 5 volt supply, or both, and in what proportions from each. Without that level of details, it's all guesswork. Also, it's somewhat hit and miss to guess at where a switchmode power supply is operating efficiently and reliably. They're not efficient when underloaded (when your supply can provide much more than you're actually using), and not reliably when run too close to the limit (having no overheads to cope with extra demands). > > Emphasis on quality, though. The dramatically cheaper ones lie, or > state consumption rather than output, use some esoteric metric for > wattage specs, or otherwise aren't actually capable of what they > claim. Kind of like the fanciful PMPO stereo specifictions, where we get tiny computer speakers claiming that they're 100 watts, when they can barely manage 1 watt. The theory behind PMPO (peak music power output) is a fanciful idea based on something like if a device can supply 1 watt for 10 seconds, it (allegedly) can supply 10 watts for 1 second, or (allegedly) 100 watts for a 1/10 of a second, because overly simplistic maths has been used, with no regard for what the circuit could actually do. Of course, manufacturers add other random bogus factors to the formula. I have as much faith in computer supply claims, regarding power, as I do with the shonky amplifier power claims. -- tim@localhost ~]$ uname -rsvp Linux 3.18.9-100.fc20.i686 #1 SMP Mon Mar 9 17:04:05 UTC 2015 i686 All mail to my mailbox is automatically deleted, there is no point trying to privately email me, I will only read messages posted to the public lists. George Orwell's '1984' was supposed to be a warning against tyranny, not a set of instructions for supposedly democratic governments. -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Fedora Code of Conduct: http://fedoraproject.org/code-of-conduct Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines Have a question? Ask away: http://ask.fedoraproject.org