Power Supplies: Let's talk stress and rated power

[angelkiller]

Ok. This is an issue that's bothered me for a while. I've discussed it with FordGT90Concept before, but we couldn't come to an agreement. So I'm going to look at this objectively now so I can understand.

I have an issue with the statement "It's bad to run a PSU at 90% of it's rated power."

First, that seems to go against the definition of a PSU's rated power. The wattage usually given to a PSU is how much power it can continuously deliver. But that's isn't fully clear and somebody has to ask, "Well, for how long?" I know parts don't last forever, and you can't expect them to, so I think a PSU should be able to continuously deliver its rated power over its expected lifetime. And of course this needs a definition also. Expected lifetime is just that, simply how long the manufacturer and consumer expects the PSU to last. CDs and in compact disks, don't last forever, but consumers expect them to last at least for so long. Same with PSUs. Me personally, I think a PSU should last anywhere from 5-8 years. So, in my opinion, I think a PSU should be able to give 100% of it's power continuously for 5-8 years.

But then that brings us back to the original statement. Apparently i'ts bad to run the PSU at 100% because it stresses the components too much leading to degration and early failure. But that seems to contradict the whole purpose of rated power, which I defined above. So now we have the statement, "The PSU is rated to give X amount of power, but actually using X amount of power is bad for the PSU." Clearly this statement is contradicting. If the parts can only be safely be run at 70-80% of the rating, why is it rated that high?

Discuss. Pick out the flaws in my argument.

Again, I'm trying to be objective and I don't want this to get heated. Thanks.

[Tatty_One]

I think the first (perhaps of many!) issue is what you mean by "rating"?, do you mean "Peak", you see the issue is, some PSU's that are called let's say 750W actually have a peak rating that is higher, then you have some that are designated 750W but that is in fact their peak rating, confusion to start with!

[angelkiller]

Some PSUs are indeed rated by peak power output. But I'm talking about continuous. How much power can a PSU continuously give out for extended periods of time? Most PSUs are rated this way.

[LAN_deRf_HA]

Well there could be a set standard for advertising/rating a psu, but nobody could regulate that other than the companies so you know that won't work. And then how would sunbeam get away with selling you "600watt" powersupply for $15 dollars that only really cost you one house when it eventually catches on fire.

[angelkiller]

Good point. I think I have to make the assumption that the PSU is able to deliver its full rated power.

[Tatty_One]

Originally Posted by angelkiller

Some PSUs are indeed rated by peak power output. But I'm talking about continuous. How much power can a PSU continuously give out for extended periods of time? Most PSUs are rated this way.

Again, thats got to be down to PSU/component quality and probably why we pay more for a better PSU, I don't think there is an across the board standard however at a guess.... and thats all it is I would go with something like a continous 80% of rated draw.