I'd like some clarification on how the sound level tests were performed. If the db meter was in fact placed 3 inches away from the back of the PSU then the sound pressure level was NOT measured. You measured turbulence instead.. the air blowing on the db meter will throw everything off. If however the db meter was positioned in such a way that no air was blowing on it, then the results should be accurate.

The reasone why I bring this up is I've measured a loud sparkle PSU before and was barely able to make the meter twitch over 60db with the meter less than one inch away from the fan.. but not in the path of the blowing air. When the PSU fan was blowing on the meter then I had results similar to these...

Dan

#42 (Collective,) yes, the PC Power & Cooling supply was bought ~3yrs ago when the Powerman fans went bad.

#43, perhaps you would be interested in some basic training:
http://www.largeprintreviews.com/manners.html

What a crap review. First, why even WASTE putting together systems to test load? Folks, this is REALLY easy:

You need 12v and 5v volt load source. You can keep wiring in parallel to keep the voltage the same but increase the wattage and use a decent multimeter (Fluke) to take DC AMP/VOLT readings. Next go for a measurement of what is called AC Ripple (AC on a DC Circuit). One potential source to load a PSU with: Light bulbs. And before any of you laugh, this is a simple exersize in electrical engineering, ask any 1st year student. Remember OHM's Law.

Hmmm, Well, I have used mostly the 300W but am currently running my server off of one of the 250's. How long ago was this? Is it the 3 years you speak of having the Power and PC Cooling? If so then InWin may have changed their quality recently. I have only been using them for about 1 1/2 years.

Powerman ratings were both 250W.

The first came with my InWin mini-tower.
The second was a free refurbished replacement
for the first.

Glad you've had success with your Powermans.
I was not. Both died with fan problems.

[quote]After two cheapo Powerman supplies went bad, I did some research and spent some money. Bought the PC Power & Cooling Silencer 275.[/quote]
I have built countless systems around PowerMan powersupplies and had zero problems. They are In-Win's house brand and have always provided me with flawless performance in either Athlon or Pentium 4 systems. What rating were yours and where did you get them from?

Another thanks to Brian Wilcox for all that info.
Especially liked your Directron link.

After two cheapo Powerman supplies went bad, I did some research and spent some money. Bought the PC Power & Cooling Silencer 275.

Pros:
Quieter than PC Power's own high-pitched K1 CPU-Cooler.

The best thing: Lots and lots of high-quality connectors, both long and short.

Outstanding support. Had a question about my order and also wanted to know proper placement of the 110 Alert Heat Alarm. Both times, the woman who answered my call answered questions quickly, courteously, from memory, without reading a script.

Now 3 years old and still running strong.

Cons:
Not totally silent, try finding one that is.
After two years and no cleaning, fans become
slightly noisier when system is pushed.
Shipping took a little over a week.

Thanks for the info!

Dodger,

Their older ones were. As far as these current offerings go, I'm not sure. What I do know is this:

-Sparkle Power Supplies are of excellent quality and definitely on par with PC Power and Cooling's offerings.

-PC Power and Cooling DOES NOT build their own power supplies, but buys and slightly modifies pre-existing designs.
As far as the 475W and 600W power supplies, I'm not sure who makes them. There is the possiblity that they actually build them, but I doubt it.
See this link:
http://www.tdl.com/~netex/cases/ps.html

-PC Power and Cooling sells excellent power supplies. They are overpriced for what you get considering most of their offerings are just slightly modified Sparkles. (But at least they don't sell junk.)

In fact, they build basically nothing. Most of their case products are built by California PC Products.

http://www.calpc.com/

Take a look at that site. CalPC's offerings will look very familiar if you normally buy from PC Power and Cooling.

PC Power and Cooling's cheaper case products are slightly modified Taiwanese offerings.

-Brian Wilcox

dbwilcox, if PC Power and Cooling power supplies are just Sparkle Power Supplies with better fans, how come...

a) PC Power and Cooling offers 475W and 600W power supplies while Sparkle doesn't.
b) None of the specs line up at all on models that offer the same wattage (different line line tolerances and amps on each rail)
c) PC Power and Cooling has a much better reputation in the business.

I also think it would be a good idea to try to burn two CD's at once, just to be cruel.

I also think that the oscilloscope stuff would be useful. However, someone would have to explain EVERYTHING to me.

I think that hooking up lots of hard drives would be a good way to increase the load. Heck, eventually, some PSU's may fail and others may succeed.

I think the overclocking thing is a good idea. Find a system with exorbitant power requirements, and quality components. That would probably give you the best "spread" in overclocking results, if any.

Oh, and if I'm not mistaken, Athlon's are hard on 5V, and P4's on 12V, right?

ER

Ag#16 The individual specs for each 'rail' often add up to more than the PSU's maximum output capacity. The numbers reflect that you can get at most that much load on each rail and then you factor in that the total load has to be below the maximum capacity. There is a whole thread discussing this over here:
http://forums.2cpu.com/showthread.php?s=&threadid=25358

Nice review. Somebody has to work this stuff out and its good to see it at Tech-Report. I agree with Indego and AG's #12 #25 that it would be very instructive to see how clean the waveform is on each rail under load. But I disagree that one needs a 'test' load other than a PC. I don't know much about power supplies (although I know someone who probably does) but its the reactance )= combined impedance, resistance, and capacitance) of the load that stresses the PSU. A fast heavily loaded PC is as probably a better load than a straight resistance or other unvarying combination.

Originally Posted by wumpus
31, don\'t be an ass. Name one site that has done an interesting power supply comparison with better data. Provide URLs please.

Ok, this is officially the crappiest report TR has ever done.

Four units reviewed? [i]Four?[/i] No power curves. Poor table explanations. "Facts" and figures that don't add up. The list goes on and on...

...and just what, exactly, does Diss's scary-looking girlfriend have to do with power supplies?

Directron had an interesting article that basically came out with the conclusion that power supply quality is directly proportional to weight:

http://store.yahoo.com/directron/psu.html

To Dodger, PC Power & Cooling power supplies are basically nothing more than Sparkle power supplies with more efficient fans installed.

Originally Posted by pHaestus
http://search.ebay.com/search/search.dll?cgiurl=http%3A%2F%2Fcgi.eb...com%2Fws%2F&krd=1&from=R8&MfcISAPICommand=GetRes...[om]

Originally Posted by pHaestus
A friend that works at a PC shop has had a disproportionate number of Antec TruePower PSUs returned (they never had problems with the old Antec 400W units tho): take that as you will because it is secondhand info.

I have used a ton of PSUs, and the only one I regularly recommend is the Sparkle 400W. The fan is loud as hell, so I usually replace with an Enermax speed adjustable fan. Rock solid for over a year with overvolted and overclocked Tbirds, XPs, and now a TBred.

I concur about the need for boosting VCore to stress the 5V lines. Enermax leaps to mind as a PSU that performed acceptably until you got serious about overclocking. There are a few motherboards out there that will let you use 2.1V on a SocketA chip; combine that with high end cooling and actually strain those PSUs.

Related to concepts of 5V stability under stress is whether a PSU has potentiometers inside for tweaking of the 5V line. This is a major selling point for me as I won\'t buy a PSU for my systems without pots.

Have you seen this article regarding testing PSUs:

http://www.hansworkshop.com/article_powerspecs.html ?

Some more quantitative testing methods.

Hmm,

You know I still wonder if the 200W PSU in my Shuttle SS51 is sufficient. I'm OC'ing a Northwood (no voltage tweak needed), have a GF4 in it, a 7200RPM drive, and an optical drive.

I wouldn't think that it would be sufficient, but I've had no problems thus far and Shuttle has said that it's even sufficient for use with a 9700Pro

TDawg

Originally Posted by dissonance
Alright, someone send me a \'scope.

I second AG#12's comment about using a scope. I always put a new PSU on my Tektronix to see what the noise level on each output looks like under load. You'd be amazed at the variation in quality -- from < 20mV "clean" DC to horrible ripple over 200mV that could cause system instability or even damage. I don't think a PSU review is complete without electrical noise measurements.

Originally Posted by dissonance
AG22: the currents are pulled directly from Antec\'s manual, and the stickers found on the Vantec and Thermaltake power supplies.

What kind of warranties do the various PSU's carry? Having been (literally) burned by an Enermax 530W PSU, I would like to know what the various manufacturers will do to replace a dead unit. Oh well, the Antec 400W unit I've been running since that fateful day has served me well :)

Oh goodie, we've got the first table straightened out.

Care to tackle the next? Or do we just scatter numbers into the cells at whim?

[quote]
Over the years, I've seen numerous systems whose laundry list of problems were all remedied by swapping out an underpowered or low-quality generic power supply. Even the latest problems with ATI's Radeon 9700 Pro have been attributed, by some, to poor quality and low-wattage power supplies in users' systems.
[/quote]

I concur with his statement. Ever since the 'thlon came out, skimping on a generic power supply is a no-no.

I like the Antec's so far, never really tried anything else since I've been building...

Please consider testing the generics and "second tier" psu, I know there is a difference, but nothing you wrote convinced me to spend the extra money on a name brand, other then being allowed to say I've got a name brand psu in my computer. And thats just trendy.

Thanks for the gf shot -- it's been missing from too many reviews lately. :P

For those of us who do not have monster full-tower cases, the gigantor-length ATX and Molex leads are a pain in the ass. All they do is add resistance and waste power, not too mention clutter up the inside of the case and murder airflow.

I was forced to return a PSU not too long ago because the leads were all too damn long. Long leads are a big plus for the tall-case crew, but a huge huge downer for the mid-ATX owners. Well, for me, anyway.

Also- every single PSU I have used shows a drop on the +5V line when you increase the CPU's Vcore. The size of the drop is my barometer for the PSU's quality level. Next time try cranking the vcore as high as you can and watch those numbers plummet. I've got 2 PSUs here, a cheapie and an Enermax- both perform identically, until you start raising the vcore. Guess which one poops out!

Good article otherwise- I hope to see more! As well as more in-depth testing. This is a great place to start, and I'm sure the next PSU article will have twice as many graphs :D

Upon further consideration, I have to ask the same question as MadManOriginal. Do those optical drives draw a lot of power when they aren't spinning? I mean isn't it the motors that draw the power? Perhaps burning a CD, copying files to the hard drive from the DVD-ROM, etc would have pulled some more power off of them? Overclocking would have been nice too.

Your output power per rail chart and the amperage per rail chart figures do not add up for the antec PS. For example to get 360 watts on the 3.3v rail you would need well over 100 amps. Why don't these #'s jive?