Article on PSU size & overkill; irresponsible journalism

PSUs: The source of DC power for all components in the PC & often a big noise source.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee, Devonavar

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Sat Jun 23, 2007 3:50 pm

mr. poopyhead wrote:i think the more accurate experiment would be to take PSUs from the same line (like the liberty series) and see if the temps differe significantly between low and high power models.
i already have ea380 and ea500, same model, different power ratings. those are the two i am going to use.

since i don't have real equipment, i'll be using three systems - low power pentium-m, a socket 939 and the one i've already used. by running them overclocked and in stock, i can measure power draw from 60-ish to 300+ watt.

probably might have to use artificial heat load at some point to drop the efficiency as well because i cannot load the 380w psu above 80%.

we'll see how efficiency spreads out over the loads, hopefully.

mr. poopyhead wrote: a poll in the forums would do nicely. though you'd have to carefully select the criteria for it.
a poll sounds like a good idea; can't think of the questions though... any suggestions?




btw, if you look through my original post on ocforums:

http://www.ocforums.com/showthread.php?t=511180

you'd be surprised how many people were supporting the idea.

for it is very much in line with the main doctrine of overclocking - get more for less :P




one more thing - does any one knows of a cheaper thermometer or thermocouple-like doohickey with data-logging capabilities and a usb interface to a pc?

i've got me a basic thermocouple:

http://www.performance-pcs.com/catalog/ ... s_id=21106

but it would be nice to make some graphs as well.

mr. poopyhead
Patron of SPCR
Posts: 376
Joined: Thu Jun 29, 2006 8:37 pm
Location: Mississauga, ON
Contact:

Post by mr. poopyhead » Sun Jun 24, 2007 11:53 am

mcoleg wrote: btw, if you look through my original post on ocforums:

http://www.ocforums.com/showthread.php?t=511180

you'd be surprised how many people were supporting the idea.
i couldn't even finish reading the first page... it boggles the mind that senior members of the forum continue perpetuating the myth that low powered PSUs aren't designed to run at their rated output for a long period of time....

VanWaGuy
Posts: 299
Joined: Mon Sep 11, 2006 1:01 am
Location: Vancouver Wa USA

power supply testing

Post by VanWaGuy » Sun Jun 24, 2007 12:17 pm

Just something to add for the folks wanting data. A while back, Tom's hardware did a stess test on a large number of power supplies. I think that they presented a lot of interesting and useful information.

ck8-04
Posts: 142
Joined: Sun Jan 21, 2007 9:58 pm
Location: Malden, MA

Post by ck8-04 » Sun Jun 24, 2007 5:10 pm

I had a little argument with Jarred Walton at Anandtech in the comments section of this article. He says his XPS 720 H2C measured close to 800 watts from the outlet with a Kill-A-Watt device, but I seriously think he pulled that number out of his ass. What do you guys think?

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Sun Jun 24, 2007 7:47 pm

mr. poopyhead - it was the initial reaction, it gets much better farther in.

VanWaGuy - thanks for the info. do you mean this one:

http://www.tomshardware.com/2005/07/11/stress_test/ ?


ck8-04 - to begin with, i personally think that people who build with crossfire, sli, and quad cores, especially if they overclock them, should look at large psu's. here's whys and whats:

-large psu's that often have better voltage regulation (high quality stuff, of course); overclocking depends on it, especially in this case when rails are loaded so out of balance.

-there's a pretty good chance that lots of the heat from those crossfire, sli, and quad cores will be evacuated through psu thus heating it up. just because someone has the money to buy all that stuff, doesn't mean they have the expertise to make it work proper.
the de-rating curves will kick in as soon as the psu works at a temperatures above rated (2% less power for each 1C rise; in some cases it's even up to 4% for each 1C).

-there are two basic reasons to build a system like that - first is competitive benchmarking, in which case there's going to be a lot more stuff powered from that psu than just overclocked components and it's not going to be a 24/7 system anyway. i say - whatever helps them to achieve their goals; they can put a 2kW psu in it if they need to; it's a different animal altogether from the systems we are talking about.
second reason is to build a room-heater. i think people can do better with pre-made units though :P


all that being said, i would have hoped that some one who does reviews for anandtech would pay more attention - the bit-tech review does specify the draw from the wall, not the psu output.

522w becomes 417w and 486w becomes 389w after we convert them to what psu supplies to the system. his dell's 700w becomes 560w, btw. (i bet the de-rating curve i mentioned kicked in here; he needs to take that dell apart and start cooling it proper.)

and seriously, dell? man, i better stop reading anandtech; if those people buy pre-overclocked dells instead of building something themselves, i am not sure they are serious about the whole overclocking thing...

p.s. i just remembered, his dell has tec cooling. that's about 120w right there. so, that makes it 440w at full load for components. ck8-04, you probably should mention that next time you talk to him...

ck8-04
Posts: 142
Joined: Sun Jan 21, 2007 9:58 pm
Location: Malden, MA

Post by ck8-04 » Wed Jun 27, 2007 12:11 pm

Wow, thanks a lot for clearing that up. I knew his numbers were fishy.

VanWaGuy
Posts: 299
Joined: Mon Sep 11, 2006 1:01 am
Location: Vancouver Wa USA

thanks mcoleg

Post by VanWaGuy » Wed Jun 27, 2007 6:26 pm

Thanks for pasting the link mcoleg, that was the review I was refering to. I was not where I could stop and look for it.

jonnyGURU
Posts: 29
Joined: Wed Oct 11, 2006 3:02 am
Location: California
Contact:

Post by jonnyGURU » Wed Jun 27, 2007 8:02 pm

What? You guys didn't think I was a member here too? :p

Look... There is no one clear cut "equation" that is going to suit everyone. I see this from both sides. I never did mean to imply a "200% rule" for anyone buying a PSU.

First off, not all PSU's are the same or have the same characteristics. There is no one PSU that's best for everyone. If there was, we wouldn't have so many choices.

Someone said that a PSU performs best at 60% to 85% of their capability. This is not always true, but it is true that most power supplies perform at their worst from 0 to 20% of their capability. On the other hand, there are many PSU's that are at least 80% efficient across the board. An example of how one rule can not apply to all people.

Also, as has been pointed out, you have PSU's rated at several temperatures. 25C, 40C and 50C. Naturally, a PSU that is rated at 50C is going to last longer than anything rated at a lower temperature, but this higher temperature rating doesn't equate with better efficiency, better voltage regulation, etc. Some PSU's can maintain performance at high ambient temperatures do to excellent air flow. But sometimes excellent air flow comes with the penalty of excess noise (read: PC Power & Cooling) Again, an example of how not every PSU is suitable for all people.

I do think there are far too many people buying PSU's that are way too much for their build. Anyone that says I've said otherwise is quoting me out of context. It only takes 800W for me to run a quad FX (dual FX72 CPU's) with a pair of 8800GTX cards. It only takes 600W for me to power a QX6800 with a pair of 8800 Ultra's all cooled with CoolIT coolers using 120W of power in themselves. Why would anyone need a 1000W? A 1200W?

I will say that sometimes when you buy a bigger power supply you do get a better build. Take the Thermaltake Toughpower's for example. You get a better quality build at ANY DC output than you do with any other Toughpower. You get better voltage regulation and you actually get better efficiency than several power supplies on the market regardless of their capability. Does this have anything to do with their output capability? No. But if you have a 600W PSU with 1% voltage regulation and a 1200W PSU with 1% regulation, just knowing that alone, which is better? The 1200W. Why? If the 1200W was only as good as the 600W but capable of putting out twice as much power, than the regulation would be 2%, not 1%. Again, it depends on the individual model of PSU. No one PSU is right for everyone. No one PSU is the "ultimate solution" on the market.

What I think is irresponsible is going to the other side of the spectrum and telling people that they need a 430W power supply just because THEIR machine only pulled 200W from the wall when running 3DMark. 3DMark is hardly taxing to multiple components simultaneously and doesn't fully utilize every aspect of a computer that uses power. And when such statements are made after only running said power supply for a week or a month doesn't take into account such usage for a year or two. Someone here said, "well, continuous means continuous so if it doesn't keep the computer up and running continuously for five years under load then the manufacturer is lying." Sorry, but that's B.S.

Continuous simply means "not peak." Peak is for only a second or two. Continuous can be a couple minutes, an hour.. maybe even 48 hours. But not a year and not five years. They don't figure out warranty's on cars by putting bricks on the gas pedal and having them redline for five years straight waiting for the engine to blow, why is a PSU really any different? When they say a PSU can continuously put out X number of watts, it's for whatever period of time the manufacturer determines is "normal" for the user of that particular power supply. It's very much like MTBF. MTBF is not based on actual usage of the product at full load for X number of years. It's an equation based on the components and contruction of said product under normal use. And normal use is not 85% of the power supply's capacity for extended periods of time.

And then we go back to the variables of operating temperature, etc.

I installed an HX520W in a friends machine. He was running a QX6700 and a pair of X1950XTX cards. We've all seen HX520W's run more than that, right? Well, it wasn't my first choice for PSU, but it's what I had handy and the build worked fine so I let it go like that. God damn Thermaltake Soprano case didn't have much airflow or something because the PSU overheated and shut off and I couldn't get her to come back on again no matter what. Did I blame the PSU? No. It's cram packed in that case and there's only one intake and one exhaust fan and it's really kind of tight in there when you have two video cards, two hard drives and two opticals. I replaced the HX520W with a Seasonic M12-700 and I haven't heard back from him (this was 8 months ago) so I assume everything is fine. Does his build need 700W? No way. I would say that open air, his PC could probably run forever on a good 500W power supply. But his build wasn't open air and his Soprano case was tucked under a desk so I would expect temperatures to get a little hotter than 40 or 50C around the power supply.

All I'm saying is there is no clear define line for what power suply is right for everyone. YES, I will say it again, there is little need for anyone to run out and buy a 1000W power supply. And I can understand wanting to write an article that dispells the recommendation for 1000W power supplies. But it needs to be understood that "this particular PC, in this room that is this particular temperature is only using this many watts when running this particular program. The temperature of the air going into the PSU is this and the exhaust temperature is that and that is because I have this many cold air intake fans, this many exhaust fans in this particular case."

Variables. There's too many variables to tell someone they only need a 430W power supply for their machine. That is what's irresponsible. If it wasn't the author's intent to imply that, I apologize. But the way I've been seeing this interpreted is "this is all you need to run your PC" with only knowing what kind of CPU and video cards a person has and that's wrong. If a person wants to buy a 1200W power supply "just because" then it kind of makes sense that you might want to tell the guy to save his money and buy something smaller. But when someone says I want the 1000W Thermaltake because it's a better build quality and has a better voltage regulation and better efficiency at 200W DC than the 500W you're suggesting and I measured my PC's AC draw to only be about 350W AC," why are you still going to try to talk the guy out of his choice? It's all about eduating, not just telling someone what they should get because someone else made the point that too much money is being spent on oversized power supplies.

If someone is looking at a 1200W Thermaltake and you tell them to just get the 1000W because it's the same build quality, voltage regulation, efficiency, number of connectors, etc. for $50 less, THAT sounds like a sensible recommendation. But if you tell someone to get a 500W Super Silent Whatever because they only really need 400 to 500W even though the Super Silent doesn't have as good of voltage regulation up to 500W than the 1000W, is just as efficient, is rated at the same temperature but would run closer to it's rated temperature because it would be running closer to it's maximum output and might not last longer simply because it's not built as bullet proof is, plain and simple, irresponsible.

ATWindsor
Posts: 285
Joined: Mon Apr 18, 2005 11:53 pm

Post by ATWindsor » Thu Jun 28, 2007 12:21 am

jonnyGURU wrote: Continuous simply means "not peak." Peak is for only a second or two. Continuous can be a couple minutes, an hour.. maybe even 48 hours. But not a year and not five years. They don't figure out warranty's on cars by putting bricks on the gas pedal and having them redline for five years straight waiting for the engine to blow, why is a PSU really any different?.
But people are not recommending buying a PSU wich runs on 100% load all the time. They are recommending getting a PSU that runs on 80-85%when the power-draw is the highest they can measure. In reality it will run on much less most of the time for almost every user. Is it really irresponsible to give people advise that prevents them from using more money than they could, if they get the same result? Even if a higher rated model is slightly better specced, what difference does it make if the lower-rated one is good enough? I hardly think its more responsible to give people the impression that they need to purchase a more expensive product than nescessary.

AtW

nutball
*Lifetime Patron*
Posts: 1304
Joined: Thu Apr 10, 2003 7:16 am
Location: en.gb.uk

Post by nutball » Thu Jun 28, 2007 1:59 am

jonnyGURU wrote:Look... There is no one clear cut "equation" that is going to suit everyone. I see this from both sides. I never did mean to imply a "200% rule" for anyone buying a PSU.
I've read this statement before. "Clearly those of us who know what we're doing won't do such a silly thing, but it's best to tell the silly n00bs this so they don't hurt themselves".
Variables. There's too many variables to tell someone they only need a 430W power supply for their machine.
The logical extension of this is that wattage doesn't matter at all. If there are so many variables that a recommendation of a 430W PSU is irresponsible, those variables make a recommendation of a 1000W PSU equally spurious. Oh, unless some variables are more important than others, for example the build quality of a PSU. Is the presumption then that high wattage PSUs automatically have better build quality than lower wattage PSUs? When you recommend a 700W PSU, are you doing so because you're really recommending 700W, or are you doing so because you're recommending a certain level of build quality? Which is it? Capacity? Quality? Is the former a proxy for the latter for silly n00bs?


EDIT: I'll tell you why this whole issue riles me so much.

I was involved in a thread over at an NVIDIA fan-site a year or so ago. One guy posted up saying that he was going to upgrade his graphics card, and wanting to know whether his existing 300W PSU would cope with the additional load. His choice of upgrade? A GeForce 6200. Yes, 6200, and this was at the time that the 7-series were in full swing, it wasn't even a current-generation product at that time. The guy was very clearly and obviously on a tight budget.

So I posted up and said "yeah, your PSU should cope". I was descended upon by the unwashed hive mind telling me that I was being irresponsible, and that there's no way that the hive mind would ever consider running a rig "on less than 500W". He should buy a 500W, or 600W (the highest recommendation was an 800W PSU). The extra headroom, it was explained, was required "in case of sudden spikes". What, I asked, would cause a 0.5kW spike? I didn't receive an answer from the hive mind.

Anyway after a little bit of a flamewar, said chap went out and spent $150 on a 500W PSU, $150 he didn't need to spend on a PSU he didn't need. Then, then he posted back saying that "his PC was running very smooth, thanks guys", at which point I was told that this was clear proof that I was talking rubbish, and they were right.

Let's just go through that slowly again:

- there are two proposals, a 300W PSU and a 500W PSU
- you test a 500W PSU, and it works
- you conclude based on that test that the recommendation for a 300W PSU was clearly erroneous

See? This is the level of thinking we're dealing with.

The hive have read your mantra and they just pick it up and run with it with their sense of cognitive reasoning set to "off". There may well be subtleties and nuances which you as an accredited Guru can see, but those are being ignored by the unwashed hive mind who read your pages. And they're causing other people to spend (a lot) of money that they don't really have. That is irresponsible journalism.

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Thu Jun 28, 2007 3:29 am

First of all, kudos to Jonny for responding, always good to hear the other side of the argument. However, a lot of it does seem like handwaving, for example the case where it has no airflow so you recommend a 700W PSU even though he doesn't need it, wouldn't it be a better idea just to improve the case airflow?
when someone says I want the 1000W Thermaltake because it's a better build quality and has a better voltage regulation and better efficiency at 200W DC than the 500W you're suggesting
I doubt any of the 500W PSUs on the recommended list would be less efficient at 200W than the 1000W TT. Also, very tight voltage regulation is a low priority on SPCR, because we don't get many mad overclockers.

jonnyGURU
Posts: 29
Joined: Wed Oct 11, 2006 3:02 am
Location: California
Contact:

Post by jonnyGURU » Thu Jun 28, 2007 3:43 am

jaganath wrote:First of all, kudos to Jonny for responding, always good to hear the other side of the argument. However, a lot of it does seem like handwaving, for example the case where it has no airflow so you recommend a 700W PSU even though he doesn't need it, wouldn't it be a better idea just to improve the case airflow?
Yes, and no.

The friend insisted on the case because it was white and it fit in his desk. As soon as I was handed the case I wanted to shove it up his ass.

He didn't "need" a 700W PSU. And I never said that. He might have been find with a more solidly built 600W rated at a higher operating temperature. I was working with what I had and what I had was a Corsair 520W that clearly wasn't up to the task (but was more than happy on my load tester prior to the install) and an M12 700W that I had just wrapped up reviewing.
jaganath wrote:...we don't get many mad overclockers.
I wouldn't necessarily say that. Many people come to SPCR for the thorough PSU testing methodology and pretty much ignore any comments about how a PSU is "not recommended" because it's "too loud." Not that noise isn't important to everyone, but you have to admit that it's to the extreme at SPCR. I mean, SPCR has shot down PSU's for noise that my H2O pump easily drowns out. :D Fact of the matter is, such journalism is not irresponsible when used in context. Like when people quote me as saying such things as "get twice what you actually think you need" out of context. But when it's given such wide open exposure and people come to it from all over the web despite the niche nature of the website, then it becomes irresponsible. Irresponsible is such a harsh word, but without such a caveat it's easily misinterpreted. As nutball said, "cognative reasoning is set to 'off.""

I've got a plane to catch. I'll drop in later.

jonnyGURU
Posts: 29
Joined: Wed Oct 11, 2006 3:02 am
Location: California
Contact:

Post by jonnyGURU » Thu Jun 28, 2007 3:54 am

ATWindsor wrote: But people are not recommending buying a PSU wich runs on 100% load all the time.
You'd be surprised.

The article that I see getting quoted the most is the DailyTech pre-launch 8800GTX article where they run 3DMark on a GTX and say, "oh look, it only pulls 375W from the wall." (paraphrasing here.)

All of the sudden, I see people suggesting 400W PSU's for people with 8800GTX cards because (as usual, the press figure is quoted out of context) "DailyTech shows that your PC only uses 375W from the wall at full load."

To which I almost always have to respond, "if you think 3DMark is 'full load' than you must only be using your PC for XLS." :( "Sell me your video card."

Mr Evil
Posts: 566
Joined: Fri Jan 20, 2006 10:12 am
Location: UK
Contact:

Post by Mr Evil » Thu Jun 28, 2007 4:54 am

jonnyGURU wrote:...Someone here said, "well, continuous means continuous so if it doesn't keep the computer up and running continuously for five years under load then the manufacturer is lying." Sorry, but that's B.S.

Continuous simply means "not peak." Peak is for only a second or two. Continuous can be a couple minutes, an hour.. maybe even 48 hours. But not a year and not five years...
I said that. If it can't reach the specified MTBF at full load continuously then the manufacturer is lying because the ATX spec says that's what the PSU must be able to do. Maybe no PSUs do live that long under full load, but in that case there are no true ATX PSUs available.

As long as the maximum power is not exceeded, it doesn't matter how much power is drawn; only the temperature matters.

A176
Posts: 21
Joined: Thu Feb 08, 2007 10:09 am

Post by A176 » Thu Jun 28, 2007 5:29 am

jonnyGURU wrote:
ATWindsor wrote: But people are not recommending buying a PSU wich runs on 100% load all the time.
You'd be surprised.

The article that I see getting quoted the most is the DailyTech pre-launch 8800GTX article where they run 3DMark on a GTX and say, "oh look, it only pulls 375W from the wall." (paraphrasing here.)

All of the sudden, I see people suggesting 400W PSU's for people with 8800GTX cards because (as usual, the press figure is quoted out of context) "DailyTech shows that your PC only uses 375W from the wall at full load."

To which I almost always have to respond, "if you think 3DMark is 'full load' than you must only be using your PC for XLS." :( "Sell me your video card."
is it too much to ask that people buy a watt-meter device themselves to measure their load? or to rephrase, what is the value of a $30 device which could end up saving $1-200 or more?

mr. poopyhead
Patron of SPCR
Posts: 376
Joined: Thu Jun 29, 2006 8:37 pm
Location: Mississauga, ON
Contact:

Post by mr. poopyhead » Thu Jun 28, 2007 10:38 am

nutball wrote: EDIT: I'll tell you why this whole issue riles me so much.
i totally understand your pain... i was involved in some similar nonsense over at the sapphire tech forums. those guys recommend a PSU upgrade for EVERYTHING, performance, noise, heat, viruses... i couldn't talk sense into ANYONE over there... they all thought i was lying about my setup (seen in my sig) running on a 400W PSU.

lol@ "unwashed hive"... harsh... but funny.

C'DaleRider
Posts: 4
Joined: Thu Apr 12, 2007 11:11 pm

Post by C'DaleRider » Thu Jun 28, 2007 6:01 pm

Mr Evil wrote: If it can't reach the specified MTBF at full load continuously then the manufacturer is lying because the ATX spec says that's what the PSU must be able to do. Maybe no PSUs do live that long under full load, but in that case there are no true ATX PSUs available.

As long as the maximum power is not exceeded, it doesn't matter how much power is drawn; only the temperature matters.

Where in any ATX spec guide is there a MTBF requirement? I cannot find it, despite perusing the Power Supply Design Guide for Desktop Platform Form Factors, Revision 1.1, March 2007, produced by Intel.

The ONLY mention of reliability in that guide, which is the Bible for designing ATX power supplies, is this (found on page 47 of the guide):

9 Reliability

9.1 Reliability - RECOMMENDED


The de-rating process promotes quality and high reliability. All electronic components should be designed with conservative device de-ratings for use in commercial and industrial environments.

Electrolytic capacitor and fan lifetime and reliability should be considered in the design as well.


Notice the big RECOMMENDED in this section? When Intel puts recommended in their guide, it's not a requirement for getting the designation as a qualifying ATX power supply, but is instead "the status given to items within this design guide, which are not required to meet design guide, however, are required by many system applications. May be a required item in a future design guide."



And all this MTBF, MTBF, MTBF....like each and every power supply or other component that lists a MTBF is supposed to last that long.

MTBFs are usually a bit of a con. They don't go around testing hundreds of units over a long periods of time since they can't afford the costs and don't have time to spare.

Basically, what happens is - they know or determine the reliability index for the weakest point (or points) in the product and approximate the reliability projections for the product itself based on that.

Let's suppose that the "normal" usage lifespan for a component is judged to be three years, and the MTBF is quoted as 1,000,000 hours. What that means is that if you take 1,000 of the components and run them all for three years, you would expect roughly 26 of them to fail within that three year period.

How do I work that out? If you run 1,000 of them for three years, the total number of component-hours is 3 * 365 * 24 * 1,000 (roughly 26 million). The MTBF is 1,000,000 component-hours in normal usage, i.e. in devices within the first three years of their life, hence about 26 of your 1000 components would be expected to fail during the three-year trial.

What an MTBF of 1,000,000 hours most emphatically does not mean is that if you just take just one of the components and run it continuously it would typically run for 1,000,000 hours before failing.

The way the number is calculated depends heavily on what the manufacturer judges to be the "normal" component lifespan. In the above example, if they decided that people typically only used the component for one year, then an MTBF of 1,000,000 hours would mean that roughly 9 out of every 1000 components would be expected to fail in the first year. But, crucially, it could well be that a hell of a lot more than 9 would be expected to fail in the second year.

C'DaleRider
Posts: 4
Joined: Thu Apr 12, 2007 11:11 pm

Post by C'DaleRider » Thu Jun 28, 2007 6:32 pm

mr. poopyhead wrote:
i totally understand your pain... i was involved in some similar nonsense over at the sapphire tech forums. those guys recommend a PSU upgrade for EVERYTHING, performance, noise, heat, viruses... i couldn't talk sense into ANYONE over there... they all thought i was lying about my setup (seen in my sig) running on a 400W PSU.

lol@ "unwashed hive"... harsh... but funny.
Why would they think you were lying about your power supply? Were they stupid or something?

I upgraded my wife's and my computer earlier this year. Her computer, typically what I was running before I upgrade----she gets the "hand-me-downs"---LOL!......was an Athlon 64 system. It had an A64 3800 X2 cpu, an Asus A8N-SLI Deluxe motherboard, 2GB RAM, one HD, two opticals, and an ATI X800 Pro video card. She had just picked out a new case, an Antec SLK-3800 which included an Antec SmartPower 350W unit in it, and that's what ran the hardware above for a month before I completed the upgrades to both our computers. (I got the first and she got the second.)

While it did power the system quite well, it did run a bit warm....warmer than I liked, but it never refused to run or ever gave problems. She now has a C2D system, e4300 on a Gigabyte DS3 motherboard, 2GB RAM, HD, two opticals, now powered with an Enhance ENP-5150GH power supply. Probably overkill, but runs very quietly, never has a problem, and runs cooler, too.

I have complete overkill in my system....an Ultra X3 power supply. It's running a C2D system......an e6400c pu on an Intel BX2 motherboard, 4GB RAM, one HD, two opticals, an ATI X1950 Pro video card, and a watercooling system.

So why the X3 when I know it is completely unnecessary and overkill? Well, I had a Corsair HX620W in it originally. But the opportunity came up to buy the X3, brand new retail and never opened or touched, for less than $150.

Now, I hate to pass up a bargain like that. So, I sold the Corsair and ended up only spending about $35 or so more to obtain the X3.

So, in my mind, I got a power supply that is, admittedly, not being used anywhere near its full potential, but gives me, hopefully, years of expandibility in my system without ever worrying about not having enough power.....and I got a heck of a deal, too.

Otherwise, I'd never have purchased it and wouldn't recommend anyone else buying it unless you just had the money to burn and just had to have it. I didn't have to have it, but couldn't pass the bargain by.

:)

Mr Evil
Posts: 566
Joined: Fri Jan 20, 2006 10:12 am
Location: UK
Contact:

Post by Mr Evil » Thu Jun 28, 2007 7:24 pm

C'DaleRider wrote:Where in any ATX spec guide is there a MTBF requirement?
Strangely they seem to have removed the detailed reliability specifications from the more recent documents, but it definitely used to be there, as evidenced by the one I linked to earlier. In fact I'm pretty sure they used to have a similar section in all their specifications, not just the ATX one. I don't know why they changed it, it's not in the change list. The EPS12V specification still contains the MTBF part, but that's for servers (some consumer ATX power supplies claim to be EPS compliant though).

Even with it missing from the spec now, many manufacturers still have the a similar full-load MTBF specification quoted for their power supplies, e.g. this one from Silverstone.

And yes, I know MTBF does not correlate well with lifespan, as I wrote about earlier, but if a power supply is specified to be able to run at 100% load continuously and it can't, then the Trade Descriptions Act is probably going to have something to say about that (in the UK at least).

But in the absence of hard evidence (not anecdotes), we're all just going to be arguing around in circles.

ATWindsor
Posts: 285
Joined: Mon Apr 18, 2005 11:53 pm

Post by ATWindsor » Thu Jun 28, 2007 9:25 pm

jonnyGURU wrote:
ATWindsor wrote: But people are not recommending buying a PSU wich runs on 100% load all the time.
You'd be surprised.

The article that I see getting quoted the most is the DailyTech pre-launch 8800GTX article where they run 3DMark on a GTX and say, "oh look, it only pulls 375W from the wall." (paraphrasing here.)

All of the sudden, I see people suggesting 400W PSU's for people with 8800GTX cards because (as usual, the press figure is quoted out of context) "DailyTech shows that your PC only uses 375W from the wall at full load."

To which I almost always have to respond, "if you think 3DMark is 'full load' than you must only be using your PC for XLS." :( "Sell me your video card."
To be honest, when i logged my PCs usage, I was rarly, if ever using more than 25 watts over the usage in 3dmark in "the real world". Then you are up to 400 wats. If you assume 80% efficency of the PSU, 400 watts from the wall is then 80% load for a 400 Watt PSU.

AtW

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Thu Jun 28, 2007 11:41 pm

jonny, thanks for posting here.

guys, for goodness sakes, we finally got an expert to post; no need to bite his head off. most of the things he is saying are pretty much the same as what i have been talking in my previous posts - build quality, voltage regulation, voltage doesn't kill psu's, people do, etc; and not a single car analogy :P

so, jonny, i do see where you are coming from with this. i do disagree with you on one point though - most people don't buy psu's based on the quality. the ones that frequent your forums do; some few like here go by whatever that certain something they look in psu, be it silence or modular cables or whatever. the majority of people just find the first reference on how large a psu suppose to be and buy it; the smarter ones go and read reviews and then buy; still without full understanding of what they need.

it's easy to sell 1k watters to people who keep repeating that 8800 cards need 30A.

reviewers don't understand what they are doing either; that kinda sux and misleads lots of folks. there's m/b 3-4 places that are competent when it comes to psu's (and yes, it does include jonny's site and spcr, even though they have different priorities).



and to all of you - why 80+% loads and all-continues on-time and 3dmarks and whatever else - have any of you actually read the article? no, really? :P

C'DaleRider
Posts: 4
Joined: Thu Apr 12, 2007 11:11 pm

Post by C'DaleRider » Fri Jun 29, 2007 1:02 am

Not to beat a dead horse, but I really think the exclusion of MTBF requirements from the current ATX standards is almost sensible.

The average consumer does not understand what in the heck it is or how it is calculated. So when his or her power supply expires in less than a year, he or she screams how the ps was rated with a MTBF of 10 years and their ps shouldn't have failed......so this equals a bogus company with shady practices.

Of course, this is so far removed from the truth it's almost funny.

While I do understand that in statistics, the mean does not exactly equate with average, in this case it's close enough.

So, you have an average life span.......MTBF. This, of course, means that half the products produced will fail in less time than stated and half will fail in a time longer than stated. The outliers are the items that do not even turn on after initial assembly and those that seem to last forever.....short of immersion in a bucket of salt water and feeding it 420V as a supply voltage.

So while the mean may be 10 years, one should realize failures will and do happen, and happen well before the stated MTBF.

:)

jaganath
Posts: 5085
Joined: Tue Sep 20, 2005 6:55 am
Location: UK

Post by jaganath » Fri Jun 29, 2007 1:19 am

and not a single car analogy :P
actually jonny did make a car analogy:
They don't figure out warranty's on cars by putting bricks on the gas pedal and having them redline for five years straight waiting for the engine to blow, why is a PSU really any different?
I accept that someone running dual 8800GTXs and quad-core and overclocking everything to within an inch of its life may well need the better build quality and tighter voltage regulation of a 1000W TT Toughpower, but for virtually everyone else with even a vaguely normal PC a 430W Seasonic will be perfectly adequate (or a 520W Corsair if they must have the latest and greatest power-sucking graphics card). Considering the overwhelming majority of SPCR's demographic fall into the latter category, our recommendations are totally appropriate for the demographic we serve. This may just be a situation of different strokes for different folks.

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Fri Jun 29, 2007 4:04 am

jaganath - darn...


C'DaleRider - "So, you have an average life span.......MTBF. This, of course, means that half the products produced will fail in less time than stated and half will fail in a time longer than stated. The outliers are the items that do not even turn on after initial assembly and those that seem to last forever....." - could you please reference a source?

jessekopelman
Posts: 1406
Joined: Tue Feb 13, 2007 7:28 pm
Location: USA

Post by jessekopelman » Fri Jun 29, 2007 3:25 pm

mcoleg wrote: C'DaleRider - "So, you have an average life span.......MTBF. This, of course, means that half the products produced will fail in less time than stated and half will fail in a time longer than stated. The outliers are the items that do not even turn on after initial assembly and those that seem to last forever....." - could you please reference a source?
What C'DaleRider is talking about is the 68-95-99.7 rule. What that rule tells us is that for a normal distribution (any random sample can be assumed to be normally distributed if it is sufficiently large) 68% of the samples will be within a single standard deviation of the mean. The problem with MTBF is that it tells us the mean, but nothing about the standard deviation. A mean of 100,000 hours with a standard deviation of 100,000 hours will tell us nothing useful about TBF. However, one could reasonably assume that, for an assembly line manufactured product, standard deviation will be relatively small and that 68% of the samples will have TBF similar to the MTBF. That said, MTBF and lifespan do not necessarily have a tight correlation. The thing is, PSU manufactures tend not to tell us the life expectancy of their products. Perhaps it is only 2 years and you are just lucky if it last longer? Since we have no way of knowing, we are forced to make an assumption as to life expectancy (based on expected use) and assign it across the board. If we assume that all PSU have the same life expectancy, then MTBF becomes a valid measurement of how the quality of one device stacks up against another (e.g. if two PSU both have a 10 year life expectancy, the one with the greater MTBF should be expected to be less likely to fail before 10 years).

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Fri Jun 29, 2007 5:43 pm

jessekopelman - thank you.

actually, i found that a good psu will last at least 4-5 years. i have some i retired simply because they've outlived their usefulness, not because they broke. i would assume a 5 years lifespan if i was pressed to. still, would be nice to know...



another thing - after all that is written in this thread, i do hope we can get jonny to post more. i do have a question or two about his initial post...



and i do have to say sorry on behalf of the performance crowd to you all - i do see how it could be frustrating to be in a minority on this particular issue.

jessekopelman
Posts: 1406
Joined: Tue Feb 13, 2007 7:28 pm
Location: USA

Post by jessekopelman » Sat Jun 30, 2007 12:45 am

mcoleg wrote:i would assume a 5 years lifespan if i was pressed to. still, would be nice to know...
This is an interesting issue. I've seen many a PSU last > 5 years, but I really don't think their life expectancy is so long. After all, how long is the warranty? Really no need to have a life expectancy much greater than the warranty. I wouldn't be surprised if the life expectancy on budget PSU is < 2 years. It could very well be this factor and not running at too high a utilization that leads to so many early deaths. So, really I think life expectancy is a fundamental question to answer before coming to any grand conclusions about the effect of temperature, utilization, or anything else (other than the obvious conclusion that extremes of anything are bad).

jessekopelman
Posts: 1406
Joined: Tue Feb 13, 2007 7:28 pm
Location: USA

Post by jessekopelman » Sat Jun 30, 2007 12:56 am

mcoleg wrote:and i do have to say sorry on behalf of the performance crowd to you all - i do see how it could be frustrating to be in a minority on this particular issue.
I don't think we are in the minority, other than in the very narrow field of DIY desktop computing. I would be very surprised to hear that people who build things with AC-DC PSU for other purposes overprovision them by 100%. Of course some people have the luxury of hot-swappable redundant PSU, so maybe I'm not comparing apples to apples . . . I don't know . . . As an engineer, I can't think of any situation where I've thought that I'd be in trouble if something had greater than 50% utilization. 80% or higher? Now we're getting into a realm where one worries about margins of safety/error. Maybe what I'm really saying is that instead of 1,000W PSU, what they should really sell are two 500W PSU in a single package with automatic failover!

mcoleg
Posts: 410
Joined: Fri Feb 23, 2007 11:55 pm

Post by mcoleg » Sat Jun 30, 2007 2:01 am

actually, we can interpolate some of the things we need to know about a comparative lifespan - as that experiment i proposed with two psu's of the same series but with different power rating.

i already got me a thermocouple with pc interface; now i need a usb adapter for it and the whole thing should be ready for testing.

gmat
Posts: 43
Joined: Tue Sep 03, 2002 3:50 am
Location: France

Post by gmat » Sat Jun 30, 2007 8:12 am

Since the lifespan of a PSU is directly correlated to its running temperature, it IS somewhat related to its power draw. Still the very fine tests run on this website showed that the maximum efficiency of a PSU was reached over 75% of the maximum rated power draw (and the efficiency curve being somewhat parabolic).
Since the most efficiency mostly translates into heat, we could say that an overpowered PSU that runs the same load at 75% efficiency will be shorter lived than a lower power PSU that runs this load at 85% efficiency.
In short, if the PC draws around 350W (for an extreme gaming system) a 450-500W PSU will live a lot longer than a 1000W one. And it's less surprising when coming from an engineering background (pick the components rated for the task...)

Post Reply