Just bought a 1000W power supply... - by bikerdude
dvrabel on 22/7/2009 at 15:56
Quote Posted by Bikerdude
Er... your missing several important points Adam!
Every PSU reaches max efficiency at 50-75% load, so for 600W this only leaves 1000W or above PSU's.
[citation needed]
Do you have a links to efficiency vs load graphs for specific PC switched-mode power supplies? I can only find good datasheets for reference designs which show efficiency is flat or rising slightly from ~20% load to full load.
David
Judith on 22/7/2009 at 15:58
Quote Posted by Bikerdude
* 1. Voltage stabilty
* 2. Power efficiency
1. Because its a single powerful 12v(80A) rail its incredably stable, so much so that my previously failed overclocking attempts(currently E8400@3.6Ghz@1.3v@400FSB@+0.15v) have now been stable for 2 days, including several manual reboot cycles - that in itself is worth the outlay
2. As I'm only using 550-600w under full load I am hitting 85-90% efficiency.
That is the point.
Hm, is it better to buy a high speed CPU and pay for it once or to buy a slower CPU, overclock it and pay for it every month? ;) Since I switched my 300W PSU to 600W I can see the significant difference in the electricity bills, and I didn't change the rest of the hardware... It might be that damn X360 too, that's why I'm not keeping the two on standby any longer.
Hiatus on 22/7/2009 at 16:14
Quote Posted by Judith
Since I switched my 300W PSU to 600W I can see the significant difference in the electricity bills, and I didn't change the rest of the hardware...
you mean the bills are higher or lower (and why)?
bikerdude on 22/7/2009 at 18:54
Quote Posted by jay pettitt
Most likely your PC spends very nearly all of it's time mumbling along using a good deal less than 100W and your fancy-pants PSU is less efficient than a smaller one would be very nearly all the time.
Er, I'm not being frivolous like... As people know I put a lot of research and time into my choices for hardware for myself and others. Fyi my system is overclocked, so I'm using more than 100w at tickover and under full load I need a very stable psu so that my overclocked cpu/fsb dont see any voltage fluctuations.
Quote Posted by dvrabel
Do you have a links to efficiency vs load graphs for specific PC switched-mode power supplies?
Ive never looked, but I imagine a good place to start would be (
http://www.80plus.org/manu/psu/psu_join.aspx)
Quote Posted by Judith
Hm, is it better to buy a high speed CPU and pay for it once or to buy a slower CPU, overclock it and pay for it every month? ;) Since I switched my 300W PSU to 600W I can see the significant difference in the electricity bills,
That entirely depends on your technical skill and the make/model of Cpu & motherboard, not all cpu's/motherboards are equal and your overclocking results will vary. Also you have to pick the right motherboard, generlay this means the mid-to-high_end motherboards. Sounds like your old 300W psu was suffering from capacitor aging using far more than 300W... ist also lickly that the new PSU supports better standby mode also.
jay pettitt on 22/7/2009 at 23:12
Quote:
Er, I'm not being frivolous like... As people know I put a lot of research and time into my choices for hardware for myself and others.
That doesn't alter the fact that, by concentrating only on the big maxed-out numbers, you're only 'hitting' best efficiency at a theoretical maximum load that almost never (if ever - holy cow, 600W is ridiculous - you can (
http://en.wikipedia.org/wiki/Microwave_oven) cook dinner with that) happens. Which means that the rest of the time (i.e. whenever you're not running the most demanding benchmarking tests) you're missing it. By quite a lot. Doh!
If you want to argue numbers it's important to try and use the right ones.
Judith on 23/7/2009 at 07:26
Quote Posted by Hiatus
you mean the bills are higher or lower (and why)?
I mean they're higher. I'm wondering is it because of both PSU and X360, or just the X360, which is awfully energy-consuming, as I heard.
bikerdude on 23/7/2009 at 09:01
Quote Posted by jay pettitt
That doesn't alter the fact that, by concentrating only on the big maxed-out numbers, you're only 'hitting' best efficiency at a theoretical maximum. If you want to argue numbers it's important to try and use the right ones.
I get what your saying, but lets look at your argument for a sec and apply it to the situation...
If we bought a psu for my system that had max efficiency that was halfway between idle and load (200-300W) like my old 600W psu. Thie means under full load the psu is not only at its its lowest efficiency but more importantly its lowest voltage stability.... Under max load the Vdroop on the 12V line would be severe enough to effect the overclock on the Cpu, Fsb and the gfx card, rendering an overclock on either unatainable.
So what are you and I saying...
* if you dont overclock and spend most of your time in 2D then get a PSU that fits the max load profile of your pc config, in my case a 600W.
* if you overclock and spend 50% of your time playing games, then get a psu that is at least 300W over your max load profile in my case 900-1000W
Quote Posted by Judith
I mean they're higher. I'm wondering is it because of both PSU and X360, or just the X360, which is awfully energy-consuming, as I heard.
Hmm you new psu should use less power Judith, so other factors have to be involved which has altered the outcome. The only way to be sure is to measure each psu with a Watt meter..
jay pettitt on 28/7/2009 at 08:55
Quote Posted by Bikerdude
I get what your saying...
oh good :)
Quote:
So what are you and I saying...
if you dont overclock and spend most of your time in 2D then get a PSU that fits the max load profile of your pc config, in my case a 600W.
oh dear :(
What I am saying is that, if you want to get the best out of your bits and bobs by matching hardware to your system so that you see real-world improvements (that includes you by the way - there's no guarantee that theoretical improvements are useful), you want real-world data. Your microwave may use 600W or more, your PC does not. Not even when you play games on it. 600W is an entirely theoretical and irrelevant figure with little real-world consequence.
My only slight worry is that the kids will read all this, aware that you spend a lot of time researching hardware, be dazzled by your clever insights into Wattages and stability, misinterpret your valiant defence against criticism as confidence, be dazzled by the big numbers and marketing blurb and, happy to be told that their super gaming rig is special and needs the biggest PSU ever, consider this thread a sound recommendation that 1kw power supply is an ideal match for a over-clocked gaming PC without applying critical thinking. It isn't. Unfortunately your reasoning on this occasion is a) faulty and b) based on the wrong data; people (including you) should have zero confidence in this thread as a recommendation for how to match PSU to their system for best performance.
Quote:
If we bought a psu for my system that had max efficiency that was halfway between idle and load (200-300W) like my old 600W psu. Thie means under full load the psu is not only at its its lowest efficiency but more importantly its lowest voltage stability.... Under max load the Vdroop on the 12V line would be severe enough to effect the overclock on the Cpu, Fsb and the gfx card, rendering an overclock on either unatainable.
This is nonsense. For starters I've got my doubts that voltage stability is going to be affected in any significant way. For seconds you're assuming that your precariously over-clocked processor is only going to be stressed when your PSU is at it's peak efficiency - i.e. when your PC is drawing 600W. Aside from the fact that the only way your ever going to get your PC to get there is by artificial stress testing (even then it's fantastically unlikely) it's also irrelevant. PSU Power consumption is no measure of processor load. Your processor can and will be maxed out at almost any Wattage. You can have zero confidence that you're going to get a good return in terms of stability or efficiency improvements based on only knowing that your PC can theoretically drink up to 600W of juice.
What you need, if you really want to match your components, is a profile of what your PC is actually doing during real world use. That way you can pick a component that is at it's best when it's most useful to be so. Even if you're a heavy gamer you would almost certainly be better off with a smaller PSU.
Quote:
The more powerful a psu is the less work it has to do there for it runs cooler and quieter.
1W is 1W, regardless of the size of PSU. Ye cannae change the laws o' physics cap'n. That you've almost certainly mismatched your PSU with your system means that it will be working at lower efficiency, which means it will be, by definition, generating
more heat and noise.
The only thing a 1kW PSU will do effectively is pedal the false notion that bigger is better and make gamers feel inadequate amongst their social peers and ever encourage them hand over yet more cash for stuff that is of little benefit.
Briareos H on 28/7/2009 at 09:36
Quote Posted by jay pettitt
That you've almost certainly mismatched your PSU with your system means that it will be working at lower efficiency, which means it will be, by definition, generating
more heat and noise.
I won't comment precisely on other things said (although 1kW is ludicrous even in the case of a highly overclocked system with dual graphics cards), but this is wrong. Even below the max efficiency (always around 50% load), for similar model build at similar effective load you'll generate less heat with a bigger PSU. Choosing a ~100W bigger PSU than what you need if you have the money can be a reasonable choice if you want the quietest PC possible.
Anyway Biker, I'm not sure that buying an average 1kW PSU for a theoretical perfect voltage stability up to 500W effective peaks is wiser than buying an excellent 600W PSU.
jay pettitt on 28/7/2009 at 09:55
nope - a smaller PSU working at 80% efficiency to provide 100W of power to your system will require an additional 20W in for generating waste heat and noise. A big PSU working at 75% efficiency doing the same will require an additional 25W generating waste heat and noise. That's more heat and noise made.
If you want less heat and noise then efficiency is what you're looking for, not size of PSU. You might be able to arrive at a rule of thumb that gets most people close to where they want to be by saying +100W, but you can't assume you'll always see real-world improvements by ever increasing the size of the PSU. The relationship between the two things isn't strongly correlated. You might slightly improve the spread - a large PSU might be more efficient across a wider range than a small PSU - but a smaller spread can still be preferable if it's better targeted.
Quote:
Anyway Biker, I'm not sure that buying an average 1kW PSU for a theoretical perfect voltage stability up to 500W effective peaks is wiser than buying an excellent 600W PSU.
That's what I was trying to say, only better :)
