All blanket statements are wrong.
Maybe. Depends on the fan and how it is controlled. If it is already spinning full speed at 70% load, it is not going to spin faster (thus louder) at 90% load. And of course, there are passively cooled PSUs too.
Maybe. 90% percent efficiency is 90% efficiency. Period. If your computer demands 300W, and your supply is rated at 90%, that supply will pull from the wall 333W and waste the same amount of energy (in the form of heat) regardless if the supply is a 500W supply @ 2/3 load or 1000W supply at 1/3 load. It is the exact same 33W of wasted energy.
Maybe. 80 PLUS white, for example is rated at 80% efficiency whether running at 20%, 50% or 100% loads.
As for greater power bills, you are talking cents per year! A Gold rated supply is rated at 90% efficiency with 50% load and 87% efficiency at 100% load. That's just 3% difference!
That means if your computer demands 300W, the supplies will pull from the wall 345W for the 87% supply vs 333W for the 90% supply. 12W difference and that is only when the computer is up and running and demanding that much power. Most computers sit near idle (or even off!) more hours per day so it would take years to "pay for itself" in energy savings.
But don't take my word for it. Use this
Electricity Bill Calculator yourself. Check your power company to see what they charge. For me,
OPPD charges 10.06 cents/kWh. I think 10 hours per day, 365 days per year is more than fair to make the point. That works out to a budget busting $4.41 cents per year savings. Now is that Gold supply really going to pay for itself over a quality Bronze any time soon? Nope.
Oh? Says what? With two PSUs running at identical efficiencies (or even 5% different efficiencies) what says the one running at 80% loads will be less stable and produce more noise than the PSU running at 60%? I would like to see your source for that.
Even a PSU running at 100% of its rated load "should" be totally stable with total noise and ripple values properly suppressed to within specified tolerances. The problem there is, of course, there is no wiggle room left. So I'll give you that with 100% loads. But that's not what you said.
No they are not and your postulations are easily disproved.
Ok, I didn't think we needed to add "all other things being equal", "not including really crappy stuff" and "in general" to every post but, , that was the intent... I have never seen a PSU spin
at 100% at across all loads... Fans tyically step up in segments, so they might be oiff at up to 50% load, step \up at 60% agaun at 70% but not 70 thruy 100% ... at least not on any IO have used. I guess its possible but I doubt you can find me an instance where two high quality models of the model series spin the same rpm
over the full range in question... so until ya can support the statement over the full range, Im saying ya haven't made ya case.
2. Bollocks. Wow ...didn't think this would stump anybody or that I would have to explain what "% of rated load means". % of Rated Load = PSU Output / Rated Load ... its significance is in determining efficiency under the 80 Plus standard. No, .... A PSU rated at 90% at 50% load
absolutely does not have the same efficiency at 20%, 100%, 1/3 and 2/3 loads.
First of, I am assuming you weren't talking about a 300 watt Titanium PSU at full load. So let's use Gold as it represents a more typical scenario .... first problem is the 300 watt Gold PSU is not 90% but 87% efficient at 100% load. A Gold rated PSU needs to have 87% at 20% load, will peak at 90% efficiency at 50% load and return to 87% at 100% load. Therefore it would draw 333 watts
only if it was a 600 watt PSU.
Lets do the math ... For calcs with *, the % is iterpolated.
300 watt Gold PSU = 87.0% efficiency at 100% load ... To put out 300 watts, it pulls 344.83 watts from the wall = 44.83 watts wasted as heat
500 watt Gold PSU = 89.4% efficiency at 60% load ... To put out 300 watts, it pulls 335.57 watts from the wall = 35.57 watts wasted as heat *
600 watt Gold PSU = 90.0% efficiency at 50% load ... To put out 300 watts, it pulls 333.33 watts from the wall = 33.33 watts wasted as heat
1000 watt Gold PSU = 88.0% efficiency at 30% load ... To put out 300 watts, it pulls 340.91 watts from the wall = 40.91 watts wasted as heat *
1500 watt Gold PSU = 87.0% efficiency at 20% load ... To put out 300 watts, it pulls 344.83 watts from the wall = 44.83 watts wasted as heat
So here's another "No".
Your assertion that "It is the exact same 33W of wasted energy. " is blatantly false. Statement 2 holds as true
3. You understood from above issue that we were talking about PSU sizes, but for reasons I don't quite understand, for power cost you went off tangent and mixed in 80 plus ratings Let's call this subject 3.A. The argument that "
Now is that Gold supply really going to pay for itself over a quality Bronze any time soon ?" is non responsive. The cost difference between Bronze and Gold is not relevant when we are talking about Gold PSUs of different sizes and % load ? But let's call\ this issue 3.B
3.a. First I thing we have to define the original statement
"The higher the % of rated, the less efficiency and the greater your power bill. The extra cost to go bigger ***can*** pay for itself with power savings. "
600 watt Gold PSU = 90% efficiency at 50% load ... To put out 300 watts, it pulls 333.33 watts from the wall = 33.33 watts unnecessarily purchased
300 watt Gold PSU = 87% efficiency at 100% load ... To put out 300 watts, it pulls 344.83 watts from the wall = 44.83 watts unnecessarily purchased
Note that I did not say
"***will*** pay for itself with power savings. (and this is true between Gold and Bronze as well as % of rated load). Many factors contribute to whether it will or it won't. These include:
a) What country you live in (Europe costs are as much as twice that of highest US costs
b) Where you live with respect to urban, suburban or rural areas
c) Proximity to inexpensive power (hydro) versus fossil fuels
d) Distance from plant to delivery point (utilities "lease" power lines from transmission line owners (can be $0.11 per kw)
cost of electricity (regionally dependent)
e) Hours per day ... miners have multiple high power systems running 24/7.
f) Power draw of system
g) How long you keep the system, didn't you say cheap crappy PSUs can last 10 years ?
h) Sales, rebates and promotions
For you positions to be valid, both the "% of load one" (3.A) and the "80 Plus rating one" (3.B) there most be no circumstance where it can't be true. This as we will see below is easily done. However, the 3% difference you alleged for Bronze versus Gold is an obvious misrepresentation. The appropriate difference is 5%.
I spent a good part of my life running a municipal power utility which included supervising the power plant operations, distribution system, and customer billing. The math is simple so no need for a calculator. I'll use your chosen example with regard to usage .... but, as the word ***can*** includes various possibilities, and avoiding extremes such as including 24/7 usage at max power miners , I'll use some more realistic numbers where appropriate:
a) System Size: 300 watts seems a bit weak for an enthusiast system, I'll use my own .... typically it pulls 688 watts from the wall and has a 1250 watt PSU
b) I'll use your 10 hours per day
c) You used 10.6 cents per kw, I pay $0.24 ... again, that's why I said "can", not "will"
d) You said
"I have seen budget, no-name generic PSUs last 10 years" , so we'll use your 10 years
So let's do the math ...
- My box pulls 688 watts from the wall .... which @ 90% ~ 619 watts of output .... which in turn @ 87% ~ 712 watts from wall. 712 - 688 = 24 watts
-My 24 watts x your 10 hours x 365 / 1000 ... That's 87.6 Kw-hrs per year
-My 24 cents per kw hour 87.6 kw-hrs = $21 a year
All costs per newegg
Seasonic Prime Gold1300 = $139 ($20 rebate)
Seasonic Prime Gold 650 = $107
That $32 is still a lot less than $84 I pay at 40% of your usage. Even at your low cost for power and tiny load, you'd be making a profit after 7.3 years. However it clearly **can ** pay for itself because your personal situation does not represent the entirety ofof the civilized world or even the US. Everyone does not have your low power costs, I pay 24 cents per kw/ hr, those in other countries pay almost twice that ... and not everyone uses a 300 watt systems, safe to say the majority of this audience does not. With an investment of $32 and a return of $21 per year that's a payback period of just 1.5 years which certainly qualifies as "soon".
Your assertion that it can not be done has therefore been disproved. Statement 3.A holds as true
3. B Now lets try the Bronze / Gold claim....
- Box pulls 688 watts from the wall .... Gold @ 90% ~ 619 watts of output close enuff to 50% ~ 625 watts .... which in turn @ 85% ~ 728 watts from wall ... 728 - 688 = 40 watts ... do the math and that's $350 out of my pocket
We can't do Bronze to Gold as no Bronze on this line ... but we can go Gold to Titanium .... 4% difference instead of the 5% from Bronze to Gold
- Box pulls 688 watts from the wall .... Gold @ 90% ~ 619 watts of output close enuff to 50% ~ 625 watts .... which in turn @ 94% ~ 659 watts from wall ... 688 - 659 = 29 watts ... do the math and that's $254 with your usage at my costs .
Seasonic Prime Ultra Titanium 650 = $114 w/ $20 rebate card and 10% off
Seasonic Prime Ultra Gold 650 = $107
That's a whopping $7 ...as you can see jumping up two 80 plus levels at $7 is not really a big deal.
So a double "no" here as upping up to
double the wattage or
2 steps of 80 Plus rating is clearly doable with a positive return on the investment. It won't work every time . But yes it **can** and quite often does.
But you also used Gold efficiency numbers but then said it won't pay from Bronze to Gold....that's not 3% but 5% .... Jumping from Gold to Titanium, just 4% still has payback for even at your very low wattage and less than average electric costs. Let's do the math for your example, th0 forgive me for using 325 watts instead of 300 to make the math easy
-650 watt Gold @ 50% load = 90%, w/ 325 watt load = 361.11 watts at wall
-650 watt Titanium @ 50% load = 94% w/ 325 watt load = 345.74
345.74 - 361.11 = 15.37 watts x your 10 hours per day x 365 days x $ 0.106 / 1000 = $ 5.95 per year for you
345.74 - 361.11 = 15.37 watts x your 10 hours per day x 365 days x $ 0.240 / 1000 = $13.46 per year for me
"Now is that Gold supply really going to pay for itself over a quality Bronze any time soon? Nope. "
Before going further, wanted to mention, I went to Seasonic because it's a line that I am most familiar. I didn't use Bronze to Gold as the numbers you used for the baseline were Gold not Bronze. In additoon, you misrepresented the efficiency improvement from Bronze to Gold as 3% when it was 5%. So I used Gold to Titanium to go along with the baseline % efficiency you used. Looking at Seasonic because it's the line with which I am most familiar, there is no line that exists with both bronze and gold models at same wattage, at least none on pcpartpicker
S12 has Bronze no Gold
M12 has Bronze no Gold
EVO has no Gold
https://www.newegg.com/Product/Productcompare.aspx?CompareItemList=-1|9SIADZJ7EW9545^9SIADZJ7EW9545,17-151-198^17-151-198
$159.99 - 10% promo = $143.99 - $30 rebate = $113.99 for Titanium and $106.73 for Gold = $7.26
Even at your low electricity costs, a very low 325 wattage your suggested usage you'd be making a profit with Titanium over Gold after just 1.22 years ... with a 12 year warranty, I think that easily qualifies as "soon". At my costs, the payback period is 6.47 months.
So your assertion that one can not "soon" get a return on investment for improving efficiency rating fails, and miserably so.
5. You are aware that the ATX standard of +/- 5% of rated voltage is a minimum. When overclocking, voltage control circuitry on all of your components has to work harder with varying voltage and increased noise. For office builds, the ATX standard of 5% is fine.... for moderate overclocking, I want < 2.5%, for extreme overclocking, I want < 1%. It's not so much what it is at the stable load as what it is during gaming where load varies. I was troubleshooting a a new build whereby the user had upgraded GFX cards and was having issues .... I measured the wattage at the wall and it was still less than is should be appropriate for PSU rating ... so was perplexed. In a job meeting at the power plant, we were having issues as various load areas were energized. I asked the design consultant about my PSU problem and he recommended that power supplies for control systems always be sized for at least 1.25 to 1.5 times the max theoretical load so that you have enough headroom foir those momentary spikes. I called vendor and asked them to send a larger PSU explaininmg that they could charge my CC and would return the smaller PSU if it solved the problem and the larger if it didn't ... whichever it was they could credit my account accordingly. Wattage "at the wall" was still less than the rating of the original (both were Corsairs VX450 and 550) . Happened again several years later with a poair of Antecs.
In recent years I have noticed that vendors tweak their voltages cause most reviewers report data at rated loads ... less usefull I think as most folks follow the 1.25 to 1.5 rule and we now often see more voltage variation at these loads. Again, it's not the steady load situation that has shown itself to be problematic, it's when load is bouncing over wide ranges, voltage regulation gets iffy. I remember an Antec TrueControl PSU that had voltage adjustments for each rail so you could calibrate your system to your own loadings. Would like to see that feature more prevalent.
As for the noise, well that's just basic stuff ... jonnyguru work for ya as a valid source ? ...using the VX450 referenced abive
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story2&reid=64
"Typically, we see ripple and noise increase their waveform as the load increases. " Im with jonny.
My back hurts just thinking about lugging a monster EATX case outside for cleaning. I have found a nice mid tower case and standard ATX board still supports more drives and more expansion cards than I will ever need while still providing lots of cooling options and still be fairly easy to lift and carry outside. 
