I didn't read any other comments yet.. but I noticed the author didn't realize what the connector on the back of the Antec Truecontrol was for.. it is there so you can stick the leads of a digital multimeter into them to properly adjust the voltage levels without having to open up the case. As to Its black connectors.. in the world of connectors, black connector housings denote a higher-temperature plastic and will stand up against melting better than regular nylon connector housings. Melting is always a non-issue in the PC world, as the curent levels are low enough.. so basically Antec is just using a higher quality connector than everybody else. It seems the industry has chosen black connectors for SATA, but for purely aesthetic reasons.

huge PSU reviews base at

http://www.amdboard.com/psu.html

#85: That's what I needed to know! Now I can update my spreadsheet. Thanks! #84

Thanks XSasber for #80 From #61: I've built a fancy spreadsheet to calculate a system's power demands, but I don't know which rail to assign the CPU power to. AMD's pdf sez it should be 12V. Anandtech's questionable article keeps implying high 3.3V current benefits AMD users. Do you happen to know which rail is used for the on-board switching CPU power supply is on the ECS mobo?

I would like to have seen a PC Power & Cooling unit included the test. I've often wondered whether it's difficult to obtain one of these for testing, or whether reviewers simply forget about these psu's as they rarely appear in psu roundups. As much as PCP&C brags about its psu's in relation to others, it seems they'd jump at a chance to have them compared. But maybe not.

I have a PC Power & Cooling Turbocool 450 ATX, but use it as a standby these days, backing up an Antec TP 430. I prefer the Antec for two reasons: it works with my UPS and the 450 ATX does not (PCP&C's active pfc circuitry gets in the way of the UPS--I got directions from the company for an internal solder mod I can make to correct this--but I'm too lazy); and the single fan on the Turbocool is easily louder than both fans on the TP430. Cosmetically, the TC450 is as spartan as they come--doesn't even offer a sheathed ATX connector for the $189 price I paid--but does come with a 5-year warranty. Still, I could buy two TP 430's for what I paid for a single TC450 from PCP&C.

But...one thing I've noticed missing from most psu reviews, when they quote the manufacturer's specs relating to rail wattage and amperage, is the operating temperature the manufacturer lists as the basis for the maximum theoretical specs. As far as I know, PCP&C is the only manufacturer to list its maximum specs at normal operating temps inside the psu. Everybody else either lists them at 0 degrees C, or ~25C, which is at best about (25C) ~1/2 the actual operating temp of the psu in normal use. The actual wattages per rail can easily be 33% less than the manufacturer's spec at normal operating temps, therefore. This is almost never commented on in psu reviews.

Basically, right now I'll own either an Antec TP/TC or else a PCP&C psu, and nothing else, for the simple reason that these are the only two psu's I know of which completely separate the voltage rails. As this review rightly points out, but fails to emphasize sufficiently IMO, voltage sharing between psu rails is one of the chief causes of general system instability.

A quick comment about noise levels. Decibel measurement tells only part of the story, of course. What also has to be considered are the frequencies and character of the sounds being generated. For instance, which would be easiest on the ears--listening to a waterfall/fountain at 50 decibels or chalk screeching on a blackboard at 50 decibels? Some fans "whine"--others don't, etc.

Last, I'm not a big fan of heavy venting in the rear of two-fan psus designed as currently sold. The bottom intake fan is actually the "vent" and since it intakes air and blows it up into the guts of the psu where the the exhaust fan removes it from the case, if the back of the psu is heavily vented then some of that hotter air from the intake fan will inevitably be recycled back into the case through the rear venting. IE, a psu with a heavily perforated rear vent panel makes a lot of sense with the standard single-fan ATX design, but can actually be counterproductive with the kinds of two-fan designs now available, IMO, at least as far as case temps go.

Last last, there's no reason or requirement for voltages to be "dead on" in a psu. As long as the readouts are within the spec tolerance, the hardware is designed to function optimally in the tolerance zones--as optimally as it would function with "dead on" voltages, IMO. It should never be forgotten, as well, that motherboard voltage regulation circuitry is no less important than the psu's own internal voltage regulation. If this needs adjustment it can often be done right through the bios of most current motherboards (ram & AGP voltage, etc.) without any need for adjusting the psu itself. I've seen many problems relating to older motherboards with inadequate voltage regulation for newer hardware in which people mistakenly believe that replacing the psu is the solution. Sometimes it may be but often the best solution there is replacing the older motherboard (or both)...;)

Didn't mean to ramble on here, but I enjoyed this psu review and agree with its conclusions, and just wanted to make a couple of suggestions...

That would depend on your definition of what is considered reliable is. By your choice of motherboard I would assume that you are trying to keep the cost down to a minimum so you might be looking at something in the 1500+ to 1900+ range for a CPU. If you are using the memory from the motherboard you are replacing, such as PC133 a faster CPU won't make a big difference because your memory bus is the bottleneck. The type of video card you use can be a deciding factor as well, the newer ones require much more current on the 12v and 5v rails. Since you work on a lot of different systems you may have to do some experimenting with faster CPUs to find out for yourself.

I'm still looking for a power supply article that addresses typical system power requirements and stability. Anybody (with a alot of $$) can buy a powerful supply and install any processor. But I work on lots of systems that needs a cpu/mobo upgrade and don't want to add in an expensive power supply. I need to know what the max freq processor can be installed for a given supply wattage or current/rail. Example: I've got a 300W Compusa power supply with 14A of 3.3V, 20A of 5.0V power and 12A of 12V. What's the fastest AMD processor I can install on a ECS K7S5A Pro with good reliability. Assume a basic system configuration since we are talking about upgrades on a shoestring.

To the anon complaining about the amps not being listed - open your eyes and look at the table on the first page.

also just use P/V=I if you must use the wattage - however this will not hold up quite correctly for the enermax 3.3/5 rails as they are shared.

Who needs big 5v anyway, surely it is the 12v rail that gets hammered now adays with the fans/hdds and processor demands? The enermax would have got my vote since the 12v rail supply outstrips the others. It would be nice to load up a few resistors to the supplies and see the ripple produced then.

Chill on the Grills -

Counting grill openings on the power supply is pretty meaningless - much of it depends on the skill of the designer in handling airflow. It may actually be better at cooling if there was but one intake fan and one outflow fan with no grills at all - as long as the internal heatsinks were designed for that air flow pattern.

Stress them Puppies -

I agree that there should have been more proper stress tests - measure the response to dynamic loading on the DC lines and measure the response to AC mains spikes, surges and dips. It would have required more sophisticated testing equipment than the average enthusiast has on their workbench but the results would have been more meaningful.

Freakin' over Tweakin' -

While it's pretty cool that we can buy DVM's with 3.5 digits of resolution for under $20, frankly any DC suppy that's regulated to within 5% is going to be good enough. I'm a huge fan of PC Power & Cooling supplies because they're built like tanks but the 1% regulation isn't much more than a marketing bullet point. Chips, drives and modules are all specified to DC tolerances within 10% so having a PSU that's regulated to tighter specs doesn't do much for reliability...

All in all, a good review with well thought out criteria...

About the only thing good about that review was that you guys measured the voltage output on each rail. But even at that, it's still suspect since no specs were given as to the type of digital scope you used.

I'm sorry... but seriously, if you're going to test power supplies, there's gonna have to be more work done by you guys than what you did. AG #39 and AG #60 pretty much said it all.

I wanna see how the PS handles power sags... I wanna see a constant load test and a progressive linear load test done.

And please let your readers know what's the most important number they need to look at... AMPS!! Screw watts... I wanna know how many amps the 5V+ and 12V+ rails can handle. Watts mean jack if the PS can convert effiecently and handle loads.

Another vote for "please include a cheap/generic PSU in the next review".

Being able to dial in the voltages on the Antec TrueControl, I was very surprised to see that the unit wasn't tested with its voltages appropriately adjusted. If it was, this fact wasn't made clear in the review.

Thanks for the effort, I enjoyed the review.

I think Geoff did a good job on the review. The SilentX looks like a good PS for those that want a quite PC, but you will need an additional case fan to cool your enclosure. I have used several Antecs and they are good as well, though I did have one motherboard that had boot problems with the Antecs but had no problems with PC Power and Cooling supplies. I am not saying that the Antecs were out of spec but some chipsets are sensitive to the timing issues, such as how long it takes for the voltage rails to turn on, in what order they come on, and how quickly they ramp. The Zalman and the Vantec are a bit of a mystery to me with their low 12v current rating. With the newer motherboards, CPUs, the increased amount of memory, and the way manufacturers of motherboards use 12v it is a critical element along with the 5v.

I'd also like to see power supplies that arent name brand,
i wonder how many of the reviewed power supplies are actually made by them.. and not rebranded... and without facts, probably most
Antec, and Enermax are the only 2 that i know are made by them.
I'll stick to my Sparkle's and cheap Powmax's
I dont need no stinking fan controls on my PS.

I couldn't get my Antec to work with Ati AIW 9700pro and Nforce2 ultra 400 board.

Service quy tried two cards and two boards from different makers, but no go. Different Antec models didn't help either (it was tried with 380W and 480W TruePowers, I think).

Fans would just spin a little when powering on, and then the system would stop.

I had to change to a cheaper power supply (which kind of invalidates the point of the article ;)

The system had no problems with a regular 9700pro nor other cards I threw at it.

Clearly the newer 9800pro works, since it is used in one of the test systems.. And AIW seems to work in KT400 based board.. Sigh..

It's too bad you did not include Seasonic's Super Silencer 400W supply in this roundup. For many quiet pc enthusiasts, this is the power supply of choice these days short of using a totally fanless solution.

This power supply boasts close to 80% efficieny (!!!), a 0.99 power factor (with active PFC), and smart fan control. This means cooler and quieter operation and saving $$$ on your next electricity bill.

It would have been interesting to see how this innovative little power supply fares against the more noisy, power hungry competition.

There's a review at: http://www.silentpcreview.com/modules.php?op=modload&name=Sections&...

Am amazed at how poor US petrol seems to be - even regular pump stuff here is 95 octane, and Shell's Optimax, which I use, is more like 98.

Would you dare put cheap gas in a Porsche? Of course not...

http://www.usatoday.com/money/autos/2003-07-30-premiumgas_x.htm

All Porsche engines are designed for premium, too, but it's not available everywhere. "Our cars must be able to drive all over the world, and so we are able to run on regular," says Jakob Neusser, director of powertrain development at Porsche's research and development center in Weissach, Germany. "You don't have to feel that a mechanical problem or anything else will happen" using regular gas, even in the highest-performance, regular-production Porsches.

Nevertheless, I believe in using a quality power supply for the PC. This is an even better review than the last one. It might have been interesting to load up the power supplies with straight resistive loads and see how close they came to their actual power ratings. It is likely that some of the units would sag or have more ripple when heavily loaded.

The decibel ratings seem way off to me. I have a tornado fan in one of my systems that's rated at 54 Decibels and I highly doubt your PSU's are anywhere near that volume. It's so far away from quiet it's not funny. If it didn't make such a difference in tempature I would get rid of it. But with the fans running in my room in the summer it's not so bad.

"...I could be the only one who considers dedicated output circuitry sexy..." Well, you and Bender.

Interesting that most deviations sag to negative. I guess because no temporary (millisecond scale) storage like one would get with a battery or a large transformer. I'd suggest that a rsd measurement to go with each voltage measurement would be informative. Nice read, I liked it.

You need to simulate load at or close to mfr's claimed ratings before you have an ide of how good a supply is. This is easily done using high wattage resistors of the correct value (which is calculated with Ohm's law). Also as another poster said, how does the power supply cope with transients?

All your article showed is performance at some (unknown) rating. This is meaningful only if I have a similar system to you - but even if I did there is no reason to think that a 550W supply is any better than a 400W supply based on your article. If you had shown that a 550W supply achieved low ripple and cool operating temperatures at 550W load while the 400W supplies got hot and couldn't produce clean power at 400W load then that would have been something useful. If you had loaded up a system with components to see if you could make the weaker supplies sag and then compared the performance of the higher-rated supplies that might have been useful too.

It's a real shame when kids write hardware reviews without knowing what they're doing. The lemmings rush out and buy TEH WINNAR and mfr's decide not to bother supporting the enthusiast community with review samples because the results are pure bullshit - why should they give out free samples when the results are a crap shoot?

I hardly think the TruePower 550 can be called obsolete just because there is a new model with (literally) knobs on. Don't forget, not everyone looking for a high-spec power supply is doing so in order to support overclocking projects of dubious utility and merit. The original niche for this type of PSU (and the most important one, by any objective measure) is high-end multiprocessor-laden RAID servers, and the most important features for that purpose are the current capacity per rail and the stability of the DC voltages. Standard 3.3V, 5V and 12V (or close enough within standard tolerances) will do just fine, thank you.

I'd be curious to see how this PSU from seasonic stacks up against yours:

http://www.silentpcreview.com/modules.php?op=modload&name=Sections&...

Slight mistake on the first page.... the enermax with 36A on all its rails delivers 480w on thr 5v rail but a mere 432w on the 12v rail, and having a combined 3.3/5v score of 200w? errr...might wanna try 180 instead.

And consequently, it doesnt offer the highest wattage on every rail from the units on test, its actually got the lowest 5v rail.

#53 again...the vantecs -5 and -12v rails seem to be listed in reverse as the two table contradict each other.

To everyone asking about PC Power and Cooling power supplies: I emailed them several times asking if they wanted to participate in the comparison, but didn't get a response to any of my emails.

Discraceful ! shame !
How could you exclude the PC Power and Cooling supplies?
The specs alone blow all the supplies you tested out of the water.
How and you recomend a supply running at +/- 3 to 5 % ?
I am sure they have value, but in my book +/- 1% is better than +/- 5%.

My Turbolink 420W PSU is handling my system well even when I overclocked the system a little bit from 2083 Mhz to 2200 Mhz with a ASUS A7N8X 1.4 Rev.
Which can't handle 200 FSB at rock-hard stability but, I managed to get it to POST,boot-up and benchmark with it
The Turbolink PSU handles my load of 2 Optical drives, 2 ATA100 HDs, 4 Case Fans, and a Radeon 9700 fairly well and manages to be fairly quiet.
I could however go with a brand-name PSU just to be safe but, I the problem to what to do with the Turbolink PSU. My case is only two that are full-towers right now at my places and the rest of computers are mid-towers you know the wiring mess a 420 PSU can do interior of a mid-tower case anyways.
The Turbolink PSU must of cost $40 which isn't bad nor cheapo the case I get cost $80 which can with the PSU

your review was mediocre at best.

It would have been better if you had done some line impairment testing,
looking at how the supplies dealt with dips, brownouts, and sags, lost
cycles, off freq., etc.

It would have been even better to look at how the supplies dealt with
major changes in load like when a P4 wakes up from it's deep slumber
and goes to full active in a matter of a few clock cycles time or what
the +12V line looks like if you spin up an array of 4 IDE disks all at once.

How efficient are these supplies?

How much hash do they put back onto the AC line (conducted emissions)?

Great review, and I agree that if you are an enthusiast, a high end power supply is a great investment. However I've built hundreds of systems using generic power supplies for joe sixpack and I've had fewer power supply related issues than I've had with faulty ram or motherboards in 7 years of system building.

and my BMW runs great on regular unleaded :D

Good review. However, there is one suggestion that I would like to throw out. Next time there is a PSU roundup, maybe you could throw in a cheapo generic PSU just to show the gerbils how much of a difference there is between the crappy and quality PSUs. This review was great, but at the same time there was no real low quality unit to compare it to. It bugs the heck out of me when my friends are suckered into buying a case with a generic PSU that claims extremely high wattages but the unit itself is probably worth $10-$20. Then they wonder why their system is unstable.