AshenSugar
New Member
- Joined
- Sep 20, 2006
- Messages
- 1,998 (0.31/day)
- Location
- ashentech.com
Processor | Athlon64 3500+(2.2gz)@2.94gz(3.03gz) |
---|---|
Motherboard | Biostar Tforce550 (RMA) (m2n-sli delux) |
Cooling | PIB cooler |
Memory | 2gb ocz 533 +1gb samsung 533 4-4-4-12 |
Video Card(s) | x1900xtx 512mb+zalman vf900 cooler(kicks stock coolers arse) |
Storage | 80gb,200gb,250gb,160gb |
Display(s) | 20.1 in dell 2001fp + KDS visual sensations 19" |
Case | Codegen briza seirse |
Audio Device(s) | ADI SoundMax HD audio onboard,using Ket's driver pack |
Power Supply | FSP 400watt SAGA seirse w/noise killer |
Software | Windows 2003 ent server as workstation(kills xp in perf and stab) |
ok to start this is all going to be quoted from pcper forums post a buddy of mine over there did, its VERY informitive and gives a good list of who makes what psu's!!!
note tho that antech has started buying lesser psu's from other OEM's and as such should be avoided for now since you duno if u are gonna get a good one or a squeeler.(bad caps)
Thanks Undersea and others for keeping me motivated. Everthing here is true as far as I know. If anything is wrong, missing or unclear I'll fix it as best I can.
This is not to replace the ongoing PSU thread, but to consolidate what we learn there. Recommended (don't like the word) psu's from that thread can be listed here, and not lost in the length. Lets try to keep posts here about: criticism of the guide, and resources. Continue to praise and bash Psu's in the other thread.
Buying a PSU for your VnF3/4
To save you from reading all the stuff below, just read this first section.
There are about 6 companies that make PSU's. They sell PSU's under their name, other names, and to other companies either off the shelf, or to special order. They also sell these special order psu's under there own names, others etc. Most brands have more than one quality level. They all meet ATX specs.The wattage specificaions we have access to are so distorted by the testing conditions that they do not show the differences between quality PSU's and cheap ones. Quality PSU's are either consevatively rated under good conditions or realistically rated under adverse conditions. Cheap (not junk) psu's are realistically rated under ideal conditions. Parts of the psu that perform reliably at low voltages, and high temperatures cost money. Putting them together properly costs money. Quality control costs money. Circuit protection that works costs money. Features cost money. Warranty's cost money. In a nutshell, quality costs money. Choose an 12vATXv2.x PSU from the top line (the line, not the wattage) of a brand that has an excellent reputation and you will get what you pay for. IF you are on a budget, buy a lower rated (wattage) model. It will outperform a cheaper, higher rated psu in all conditions. If you plan to upgrade and/or overclock and/or future protection (use in your next and even next computer) get a higher rated model. Spending as much or more for a PSU as you do your motherboard is not outrageous (considering the price of a VnF) if you consider all the money you have invested (cpu, drives, memory, GPU), not to mention your data. The stuff below can help,
Buying a psu is much more difficult today than a just few years ago. The power requirements have changed and your requirements have changed. I hope in this guide to to help you identify what your psu priorities are and how to find a psu that meets those priorities. Fortunately or unfortunately, the VnF series boards attract buyers over the whole spectrum of users, from budget builders to crazy overclockers (phase change on a VnF4, are you kidding me) :biggrin: , so everything must be covered.
Lets get the Basics Out of the Way.
We have been recommending 18 amps of 12v power for a basic computer (512MB, 1 hard drive, one CDR or DVD, no 12v connection to Vid card) since the VnF3 came out. Why, because it works. This is more 12v power than most other boards, even other nF3/4 based boards. The reason I think is the capacitors on our boards.
The VnF4 ultra http://www.newegg.com/Product/ShowI...IA nForce4 Ultra ATX AMD Motherboard - Retail
vs the Asuys A8N
http://www.newegg.com/Product/ShowI...IA nForce4 Ultra ATX AMD Motherboard - Retail
These capacitors contibute to cpu voltage stability (enhancing overclocking), but put a serious strain on the psu during start up. 18 amps is more 12v power than many excellent but older psu's were designed to put out. You are welcome to try and use an older lower rated 12v psu, but if the board won't start or you have stability problems that develope down the road as the psu deteriorates from the stress, you have a good idea where to start looking.
The new ATX 12v 2.0-2.2 PSU spec cpu's are designed for lot's of 12v power and a budget 2.0 spec psu will start and run a basic VnF just fine. Since the VnF4 is capable of running on a 20 pin ATX12v connector (the VnF3 is 20 pin) an ATX12v psu that meets the 12v requirement may not have to be replaced, depending on the specific PSU quality and overall output capability. Be aware that 12v vs 5V balance is different and this causes stress and may cause voltages to go out of spec. I do not recommend spending money on a pre 2.x spec psu. (note 1.3 spec psu's will not have an imbalance problem but are usually less efficient and cost as much or more than a 2.x)
The ATX v2.2 (board) spec http://www.formfactors.org/developer\specs\atx2_2.pdf
The 12vATX2.2 PSU spec http://www.formfactors.org/developer\specs\ATX12V_PSDG_2_2_public_br2.pdf
The main differences the 2.x standard and the 12vATX are higher 12v as a percentage of the total output, connectors, and mandates at least 70% efficiency. There are 3 2.x versions, The only differences are the elimination of the -5v line, and whether or not pc-e connectors are part of the spec.
More than you will ever want to know about psu's http://www.xbitlabs.com/articles/other/display/psu-methodology.html
Picking your PSU Wattage (X2 processor, jump one grade) Always 12vATX v2.x
Base this on where you think you will be when you want to buy a new PSU. This PSU if powerful enough will be suitable for your next system. Like a good 12vATX is suitable for a v2.0 system, This psu willl be good for BTX (24 pin + 4-pin already adapted)
Of course this is general, use your judgement. I emphisise the future because of the myriad of upgrade possibilities, dual cores, 6800gt-7800 gpu's (ati new one as well), and reports of 3 amp 12v harddrives in the future.
Standard User, 2 sticks memory, 1 HD, 1 CD/DVD 350-400 watts
Power User, Add 2 devices or CD/DVD RW 400-450 mostly extra 3.3-5V
Gamer, PU with pci-e connector Vid card 450-500 incl small O/C
OVerClocking- SLI Buy the best you possibly can
12v requirements (Amperes) note: There is a general trend to more power, and concentrating this increase to the 12v line, which should cause these values to increase with time
,.,.,.,.,.,.,., Load (+CG),.,.Boot.(+CG)
cpu ea.,.,.,.,. 5.,.,., 2,.,.,. 3*,.,.,0.5
gpu ea,.,.,.,., 3.,.,.,,.,.,.,.,1
HDs ea.,.,.,.,.,1**,.,.,.,.,.,2.8**
Fans ea,.,.,.,.,0.2,.,.,.,.,.,.0.5***
CD.,.,.,.,.,.,.,.1,.,.,.,.,.,.,.,2
*calculated from bios temp when boot is paused, draw while boot is being processed is undoubtably much higher
** (+0.5 amps for 12v only harddrive)
*** large or powerful fans can draw up to 1 amp starting
Loads estimated from info in below links by interprolating my load and idle temps. Same sources have 3.3 and 5v data
http://www.xbitlabs.com/articles/video/display/ati-vs-nv-power.html
http://www.xbitlabs.com/articles/cpu/display/athlon64-venice_6.html
http://www.seagate.com/support/disc/specs/sata/st3120026as.html
http://www.plextor.com/english/support/manuals/man_320.zip
New, Common board power usage This is a psu review, the beginning is a power usage anaysis. And another,
More boards tested Note that the great configuration differences make this study confusing. This study would have been much better if some attempt had been made to match components on the boards. The individual devices and conclusions are quite useful.
A modern decked out computer has a hard time drawing 250w DC at full bore, so why so high? First of all, power regulation is generally best when power draw is 1/2 to 2/3 the rating. Also things like poor AC quality and higher temperatures lower* output. Since I cannot know your circumstances and configuration, you will blame me if your psu let's you down so there is also some fudge factor. I make very good fudge. Your upgrades won't stress the psu. (Wattage Calculators on the Web use too much sugar)
*PCPower&Cooling is almost immune to these factors, because the rest of them (even the best) are so coy with their #'s you cannot be positive.
Marketing
BrandX 450w (and various letters and #'s) PSU. This is a marketing/model name. Only a starting point.
ATX 2.x.,., has modern connectors, is not 12vATX ver2.x compliant
sli ready.,., has 2 pci-e connectors but is not nVidia certified
silent,quiet, by what standard? at what temperature. Fans that are shown to ramp up in a straight line cannot be silent. See SPCR http://www.silentpcreview.com/article28-page1.html for reccomended quiet psu's
p4 A64 ready. has 4-pin connector, 12vATX, not necessarily good enough, not 2.x
Active Power Fractor Correction (APFC) Does not reduce power(watts) usage (can reduce commercial acct bills that pay for VA's) Seems to be stadard ON 2.0 PSU'S, Identified by no 115-230V switch on back. Apfc returns clean power to the grid, so general usage will reduce strain on the system. (Maybe eliminate the need for a new power plant in the US)
Efficient Get a #.
Modular Wiring Causes resistance in wiring (psu works harder to compensate), and long term reliability questionable, for now. For those who like to display their case. I save the money and tuck the extra wires out of the way.
Reading and Understanding Specs- A key to getting the right psu
Straight into the snake oil. Buying a new PSU is like walking onto a used car lot. They can't straight out lie, but they can misdirect and deceive with impunity. Further down we will analyze the actual specifications. There are no standards. If you can't find the ones you need, do not buy that psu. This is a partial list, as it is probably endless.
NewEgg, on the details page has a photo of the spec tag (not always legible though) and a direct link to the manufacturers page for the product. I highly recommend this resource for your research
Specs
General Rant on PSU Specs: There is no standard for the listing of the specifications, except those of the ATX standards. This leaves a great deal of leaway for the specs to be distorted by the measuring conditions used by the manufacturer or just not listed. For example continuous, peak, and temperature should be listed clearly by every maker and brand, and they are hard or impossible to find, even on excellent psu's. One reason for this state of affairs is that a psu designed for adverse conditions (low voltage, high temps, high transient responce, tighter regulation etc) will by definition have lower output/$ when compared to a psu speced to ideal conditions. This is the main thing you buy when you buy a quality psu, the ability to power your computer through the rough times. Note : things have gotten worse at manufacturers web sites lately. I hope it is because of all the new 2.x models, and they will get caught up, and correct and complete them.
On Case specs Version changes can cause disagreement with manufacturers web page. Sometimes they are just different, don't know which is accurate.
No listed 3.3-5v limit Most psu's have a combined 3.3-5v limit that is lower than each individual rail. If you are not sure the psu has 3 individual rails,(very few do) pass on this brand. If the model# reflects the individual totals pass on this psu, and the brand. Confirm at the website. You have to multiply the listed amps by the voltage (AxV=Watts) to get the watts. Bad psu's depend on our failing to do this. (2.0 spec cpu's seem to not have this problem)
Are the values labeled Continuous, or is a higher peak value listed? If not, these could peak values and are good for 1 minute, not continuous (If you are still interested check the manf. web page for specs.)
Subtract max 12v (and estimate that you will actually be using @ 18 amps during boot) wattage from listed max total. Is the difference enough to run your memory and other 5v devices (50 watts, mostly 5v)?
Specs Page- If specs different from retailers page, retailer maybe selling older version.
Go here before you buy, this is where you will find the truth if it is available. You will probably have to look hard for the important ones. If you can't find it, assume the worst.
Specs tend to get scattered (the better to confuse you). Some times they are on the page for the whole line, sometimes on the product page, the detailed spec often linked from there. Very often you will have to check all 3 to find all the specs you want. All to often they are no where to find
Be sure you are getting the correct specs for the psu you are considering. Pitfalls include, the model at the retailer may be discontinued and the specs at the site are for a revised model, retailer linked you to a similar but higher quality psu, or you clicked the wrong link. Check the part# carefully.
Note: Temperature is the most important variable. The higher the rated temperature the better the psu, the higher the real output, the longer the life of the psu. Remember that the ambiant temp you are concerned with is the temp inside your case. 25C is out of the case. The higher the operating temperature inside the psu, the lower the output. All things being equal, a high efficiency psu will have a lower actual operating temperature for the same rating and at the same output.
Rated Operating Temperature
Most commonly
0-25C The lowest rated temperature, universal in cheap psu's and creeps into good ones. The psu will never be able to output the listed wattages in your case. I have a VERY cool case temp, (more than 1 sqft ventalation, 3 120mm fans (incl psu) and an 80mm blow hole) 26 to 30c. Common case temps approach 40C. For every degree C over 25 output goes down. If the rate of decline is not specified, assume it is very high. PC power says it all in this graphic. http://www.pcpowercooling.com/pdf/Turbo-Cool_510_vs.pdf
0-40C A quality PSU will output rated power in almost all situations. At temps below 40C output is actually higher than rated. This is a hard spec to find online. My tagan (50C) and seasonic (40C -2w/C to 50C) include the info in the manual, but that's post purchase.
0-50C Server Class, unfortunately around $200. (exception http://www.pcpowercooling.com/products/viewproduct.php?show=S47ATX&view=techspecs, not 2.0 but very good. (I don't include the Tagen, because not posted at site.)
Efficiency Rating
The amount of DC power output per 100 watts of AC input. The balance is heat, ie electricity you paid for and got nothing from. An 80% psu will require 200 watts of AC to supply 160 watts to the board. A 65% psu will require 246 watts of AC, adding the heat of an 80 watt light bulb to your psu, 46 more than an 80% efficient psu. This extra heat requires more fan noise and/or speeds the deterioration of the psu components.
60-65% typical 12vATX
70%+ 12vATX v2.x required
80%+ new green standard and best available
Output Ratings (Thought I forgot about them didn't you) There is no standard, and often continuous is not identified. You can be sure only if identified or the presence of a peak voltage spec.
Remember that before we were concerned with amps per rail, and that you had to multiply Volts times Amps to get Watts and the 3.3v and 5v combined rating. With the dual 12v rails common on 2.0 spec psu's there is also, now, a combined 12v rating, typically two 18 amp rails maximum 30 amps. this is a result of there being one 12v (good for 30 amps) rail with 2 18 amp circuits coming from it. (There is at least one brand that has 2 separate 12v rails)
Combined Output. This will be the Maximum power available at any one time, and is the real output of the psu, if listed as continuous. (see peak power next) It accounts for the lost power of combined rails and any limits maximum 12v output puts on the 3.3-5v combined output and vice versa. This should be very near the marketing/model # (or advertised ) rating. (400 watt psu at 385 combined etc) If not look else where. Of course this is artificially high if the psu is rated for 25C, but is not a deal breaker if treated honestly. The more of these limitations that are listed, the better the psu.
Peak power. A built in reserve that keeps the psu from shutting down immeadiatly if stressed above the continuous rating. (I'm pretty sure this is listed only for 40C and above psu's, and the circuit breaker will not go until temps reach over 40C) This is your assurance that the Combined Output is Continuous and so are getting what you are paying for.
Regulation
Ripple/Noise ATX spec Amount of A/C that makes it through the filters Tighter equals better
Range, ATX spec, 5% all (+) rails, not to worry unless determined O/Cer Tighter equals better
Transient Response How much the power load can change and maintain regulation, as % of load, higher the better (don't know if this defines condition of Voltage regulation rating or is a seperate test. I have seen different loads and capacitors.
Input Voltage
Of special interest to North Americans (110-120v)
If listed as 115v AC, the psu is not guaranteed to maintain rated output in areas with lower AC power levels (normal or brown out). PSu's that support 90v AC and lower are available. Input voltage labeled 100V(rms) is roughly equivelent to good 115AC
MTBF Typically 100,000 hrs, more than 10 yrs. Does not include fans.
From BWM,
Here's what I have up to this point. The material you have posted is great. Other information for you to consider offering in your material:
There is a lot more to a power supply than volts, amps, and watts. Be very leary of test reports/reviews on the internet. The vast majority of websites that do PSU reviews do not have the test equipment or knowlege to perform a proper review. When I say proper, I mean the ability to verify that the unit under test meets its own published ratings and the requirements of the official ATX Power Supply standard that applies to the unit under test. In my opinion, as an engineer, anything less is a disservice to the reader, who may well take a poorly written review as the gospel.
I must emphasize the need to verify each of the specified requirements in a dynamic fashion, which the official ATX standards describe, not the static conditions under which most reviews are performed. Regulation, for example, must remain within the official standard limts at maximum rated load AND minimun specified AC input. Same at minimum load and maximum AC input, and at several points in between. The same principal applies to other specifications as well, noise and ripple in particular. Noise and ripple are important because they generally increase under light load, and again at maximum load. Most power supplies are cleanest at 1/2 to 2/3 of their maximum rated load on each rail.
One very important and almost universally ignored aspect of a power supply is startup timing and housekeeping, also covered in the official standard. Most people including reviewers do not fully understand what happens when they press the front panel power ON button. It starts a sequence of events beginning at the PSU, which the motherboard must respond to correctly and at the proper time, resulting in the assertion of a turn-ON command back to the PSU, when and only when the specified conditions are met. An example of the importance of this is the period of time in which Antec Truepower models were blamed for a rash of system won't power-up or post problems, almost exclusively involving nvidia based motherboards. To put it mildly, the power supplies were not the problem. Although Antec did make an accomodation which successfully addressed the problem without making their product non ATX standard compliant, the problem was the nvidia ACPI/power management implementation.
Measurement notes:
One cannot possibly verifyy that a voltage or any other quantity/property is within allowable tolerances when the measurement instrument itself can contribute enough error to throw it out of whack. For the purpose of PSU voltage measurement, the following should be understood:
1 - the tolerances for the positive rails are generally +/- 5 percent. Thus a reading anywhere between 11.4 and 12.6 is within tolerance. Now, add into this the accuracy of the meter used. Analog meters are usually accurate to within 5%, meaning the displayed reading may now be off as much as 10%. Obviously, this is nowhere near accurate enough, as now one must accept that a displayed reading between 10.8 and 13.2 may well be within the 11.4 to 12.6 tolerance.
2 - a wise rule of thumb to follow is that the meter used should have a tolerance such that its own allowable limits of error do not impact the least significant digit of the quantity being measured. In other words, in the case of 11.6 to 12.4, the error of the meter should be no more than .5%, to ensure that the meter's worst case allowable error has no more than one LSD impact on the reading. There are plenty of cheap digital multimeters on the market that won't meet that rule of thumb, but there are reasonably priced DMMs that do. The DMM should have a basic DC accuracy of plus or minus .1 percent. Why not .5%, you ask? That's a little off topic for this piece, but I'll be happy to explain it to those who may want to know why.
3 - do not take what appears to be an unstable voltage reading in the BIOS or any software monitor utility as being truely noisy/unstable. It probably isn't, and I'll be happy to explain why when/if asked.
4 - regarding what most reviewers post as a load bank, it cannot be a reactive load (ie wirewound power resistor) for a valid noise/ripple measurement. The inductance of this type of resistor is a reactive element (impedance), not a pure resistance. The presence of a reactive element in the load, ie capacitive or inductive, will skew the measurement one way or another. A proper load bank is a carbon pile resistor immersed in transformer or other oil expressly intended for the purpose, in conjunction with the capacitve load stated in the official ATX standard, to simulate the resistive and capacitive type of load that motherboards present to the power supply.
5 - an oscilloscope is mandatory under all conditions, to verify that rise time, fall time, overshoot, timing, and housekeeping specifications spelled out in the ATX standards are met.
6 - the transient response of a power supply is very important. When it is not fully compliant, anything connected to it is at unacceptable risk.
In summary, a fair job of writing a credible review can be done with a decent oscilloscope, .1% accurate DMM, a purely resistive load bank, a variac, and a thermometer such as a Fluke thermocouple model from the 51-54 model series. IR thermometers are not appropriate for this or CPU temperature.
There's a link in my siggy which covers relatively simple yet accurate temperature measurement techniques. I've been an engineer and NIST certified metrologist for 25 years. I may have more material for you later. If you have any questions, just ask.
Temp Measurement Link
BWM
Thanks BWM, some very good stuff there. The clarification on the nVidia start up problem is very enlightening, and should be considered in our Recommended List, and should help explain our all to common startup problems.
Manufacturers and the Brands (Stolen from NV4TeHWIN's post at Nvidia Forum
I've edited for space and note that Brands can and do have PSU's from more than one Manufacturer. Top Power for example may make one or more RaidMax but RaidMax brands many a trash PSU.
Power Supply Manufacturer TOP 10 (or TOP 5)
Here's something you might not know. Most companies that sell power supplies don't actually make them. They slap a sticker on someone elses unit and sell it as their own (commonly referred to as "rebadged"). While many of them have their own features added to try and niche out a spot in a crowded market, this usually doesn't hurt it's performance and might even improve it functionally, not just add shiny eye candy. Many though, it's just eye candy. And many do absolutely nothing... they just slap a sticker on the side.
--=--=--=--=--=--
#1a
Topower
(some TTGI - Super-Flower, OCZ, EPower/Tagan, RaidMax, Vantec, ACI)
#1b
FSP - Fortron Source Power
(Fortron, Sparkle, Cooler Master*, Zalman, Aopen)
#1c
CWT - Channel Well Technology
(Antec, Lead Power, Enermax, Xclio, Turbolink)
While these top three encompass more than the re-badges I've named, you can usually bet on a solid performer if you pick one of them. It's highly debatable which are better than others, of course. But all in all, everyone has had pretty good experiences. Not suprisingly, they are priced accordingly. None of them in the 400W range are going to be under $50. If your power supply is one of the above, chances are you can trust both the unit itself and it's claims on wattage per line.
#2a
Sea Sonic
#2b
Wintech
(Ultra X-finity)
#2c
Acbel Polytech
*(Stateside, just the Cooler Master True Power)
Same with these, a bit more esoteric. If you can afford a Sea Sonic these have the best efficiency of any consumer switching PSU for computers. You'll pay for it up front, but your electric bill will be less in the long run.
#3a
Sirtec
(Chieftec, Enlight, ThermalTake, High Power)
#3b
HEC - HeroIchi Electronic Co.
(HEC, CompuCase)
#3c
AMS - American Media Systems
(Mercury)
Here you start finding dissention in the owners. Some claim it's the best in the world, some wouldn't let their dog run Windows 3.11 on one. Some of them have great reputations (TT & HEC) but the problem is a "squeaky wheel" one - it's hard to know just how good/bad things are when only people with problems post.
#4a
Youngyear
(Aspire, Logisys, MGE, Ultra X-Connect, Rosewill)
#4b
ATNG
(Coolink, CoolMax, Rosewill, StarTech)
Now we're into budget territory. Maybe it's just a quality control issue; some people love them but most hate them, and from bad personal experiences. I personally wouldn't touch them. They should work, but no one will be suprised if it smokes itself and takes your motherboard with it. They rarely come close to what their ratings state, very poorly constructed. (but usually colors or chrome and UV sleeved and colored Molex's and LED fans!) These for the most part are toys, although you might get lucky. Not all of them are, but it does take more effort in sifting through them than it's worth.
#5
CRAP!
(L&C, Deer, Allied, Eagle, CodeGen/Foxconn, EverPower, Maxpower)
If you believe one of these things will power your computer, I have a 7-band underdash amplifier/equalizer that's 250W+250W that I'll sell you for $5. You'll love it. Really, if you have one of these supplies, don't mention it to us. Just silently toss it out and buy one in the #1 category.
--=--=--=--=--=--
Resources
Reviews
Reviews with software voltage readings, or AthlonXP cpu's (usually both ) are worthless. As far as I know, no one does a review that does everything. I guess it takes to much time, equipement and expertice to do a thorough test. Here are the best I know of. If you read some, see if you can spot their weaknesses. IMHO all three together are not complete, but there is a enough info in the side articles to get you up to speed.
XbitLabs use the calander to go forward in time. Unfortunately, there is not a seperate psu catagory. Highlight, Full power and crossload regulation, lots of info in New Methology article
SilentPCReview Highlight, Noise and temps at full power, lots of info in other articles
PCPer A few at bottom of page. Highlight, low input voltage testing on high draw system.
Toms Hardware 19 modern PSU's tested, 6 of which fail. Highlight, Full load 24 hr with regulation and ripple. Because this is a very important test, I am going to point out it's weakness here as well. Promised sound and temps data is mostly not given, and not specified well. An out of case test, so at unusually low temps (fanless Fortron gets big benefit compared to its target life). Many if not all these psu's are 40-50C rated. This is a load test, so there is no interaction with the constantly changing load of a pc. Low 230v doesn't hammer a PSU like low 115v, but that goes with European testing. Still a very important test.
Background Info, Voltage Calculators, Misc Stuff
Hard OC's power supply thread, Very Good, just a little dated (pre 2.0). What I'd like to emulate here, I'm pretty sure the same brands will end up at the top of our list. Maybe being on this list should be the first prerequisite, and other brand will need to show a lot to be included. Depends on how much emphisis we want to give to the past record of quality.
SilentPCReview
again, lot's of good info, biased twards silence.
Decent Voltage Calculator Could nit pick, but the best I've seen.
Tips
Fan style
Fanless, in ther new quest for quiet fanless psu's are being developed and released. In my opinion they are prone to overheating causing short term shut down and long term damage. Can be used with forethought about how to cool the case and psu. There are alternatives in the quiet psu's that will give equal decibel levels, cost much less and deliver superior performance. 'Til technology catches up, not recommended.
1 80mm fan at exhaust Old style, noisiest, especially when not thermo-controlled.
2 80mm fans, quietest type when inner fan controlled by wattage output, outer by temp. This syle psu is longer (to accomadate the extra fan). Best with rear case fan, air flow bypasses cpu area.
1 120mm on bottom Modern darling, quiet, rear case fan can be eliminated with this design in silent type system. Not the panacea marketers suggest. Large fan cramps components, rear vents recirculate hot air, (suggest top case vent), can fight cpu fan if cpu is at top of board, 90d turn restricts flow.
I suggest either of the newer styles, unless your personal requirements require a particular style.
Reset Usually in the manual. Try 1) turn off with switch while plugged in at least 1 sec, repeat and 2) unplug psu and wait 5 min. Try both to reset logic if you suspect psu has turned itself off.
I recommend cleaning the psu as soon as warranty expires and inspect yearly. Change the fan every three. Be careful that controlled fans will start and run at controlled voltages and any replacement fan be powerful enough to cool the psu. Check the specs on the original. It is possible to use a slower (quieter) fan, at your own risk.
this guide is a bit old, prices have droped alot since it was put out but the info is all still valid.The Original Manufacturers and the Brands that buy from them. Brands can buy from more than one manufacturer.
HELP NEEDED This has been tried before, but I have never found a complete list. I think it will be really helpful in seperating the wheat from the chaff.
TopPower
Tagen, E-Power, OCZ,
FSP
Oem psu's, Fortron, Sparkle
Note, these recommendations are not complete
Recommended Premium Brands
These cpu's should surpass your expectations.
PC Power & Cooling A little noisy, a little old fashioned but they set a power requalion standard with the 510 that no one else meets.
PCP&C Turbo Cool 510 SLI $219 An extra 30.00 for 24 pin and pci-e connector over the same spec standard 510's (around 190), but if you can afford it, the tightest regulation you can buy.
PCP&C Ultra Quiet 470 $99 Their consumer level psu, not 12vATX v2.x but strong 12v rail. Not ultra quiet, no pci-e, not as tight as the 510's but very solid.
SeaSonic Set a standard of quiet, efficiency and performance approached by a select few.
S12 line 330-600w $59-159 Will easily support my recommendations (330w equivelent to 350w etc.) Current model 80+ certified, 600 model supports SLI
Recommended High performance Brands
These psu's should surpass your expectations, but with a noted fault
Antec
TruePowerII (Each individual psu specs are on the psu's page) Load Regulation 40% better than standard, stand with their customers. No AFPC on US models
Tagan 230v areas only
All LinesVary by connectors and bling. Have everything, would be Premium but my Black 480U01 was very sensitive to 115v power and would switch off before UPS (2) could respond (Love it with that exception). Seperate ground cable makes suitable for non-metal cases.
OCZ PowerStream520Adj The ability to maintain tight regulation to advertised wattage, and individually externally adjustable rails makes this PSU an overclockers dream. However, this psu is not ATX12 v2.x and, at @90% load and above (enough to power 2 overclocked VnF3s with 6800GT's through 3dMark05) heat levels can become high enough to cause shutdown and/or cause long term damage to the psu if used at that level.
Recommended Solid Performers
Suitable for even heavily configured systems, some care needed to match your needs.
Fortron Blue Storm
Recommended Budget PSU's
Will meet your expectations if picked with care.
Sparkle
Sparkle FSP300-60N $32 For those times when you have no cash. For lightly configured system only. If it wasn't FSP, (always seem to outperform their rating), it wouldn't be here. Look for higher output (up to 400w, new model so may be tough to find for a while) models for a real deal.
FSP350-60THA $50 A more flexible starting point. (Look for 400)
FSP450AX-PN
Link to user discussion and their psu's http://forums.pcper.com/showthread.php?t=402356, the first thread on this subject. http://forums.pcper.com/showthread.php?t=402356
note tho that antech has started buying lesser psu's from other OEM's and as such should be avoided for now since you duno if u are gonna get a good one or a squeeler.(bad caps)