550W Roundup: Three PSUs at Different Prices
by Martin Kaffei on October 28, 2010 4:00 PM ESTOCZ Voltage Regulation and Quality
| +3.3V regulation | |
| Load | Voltage |
| 10% | 3.39V (+2.72%) |
| 20% | 3.38V (+2.42%) |
| 50% | 3.34V (+1.2%) |
| 80% | 3.30V (0%) |
| 100% | 3.28V (-0.61%) |
| 110% | 3.24V (-0.91%) |
| +5V regulation | |
| Load | Voltage |
| 10% | 5.03V (+0.6%) |
| 20% | 5.02V (+0.4%) |
| 50% | 4.98V (-0.4%) |
| 80% | 4.94V (-1.2%) |
| 100% | 4.91V (-1.8%) |
| 110% | 4.90V (-2%) |
| +12V regulation | |
| Load | Voltage |
| 10% | 12.37V (+3.08%) |
| 20% | 12.34V (+2.83%) |
| 50% | 12.29V (+2.42%) |
| 80% | 12.25V (+2.08%) |
| 100% | 12.21V (+1.75%) |
| 110% | 12.15V (+1.25%) |
+3.3V as well as +12V start very high with +2.7% and +3.1% over the desired value, though this isn't necessarily "bad". +5V starts lower, but there is not much room for the typical drop. Even at 110% overload all rails are still within specification, and +12V never drops below 12.15V, which is nice (particularly if you're pushing an overclock).
Ripple and Noise
| +3.3V ripple quality | |
| Load | Ripple and noise |
| 10% | 14.10mV |
| 20% | 14.40mV |
| 50% | 20.80mV |
| 80% | 21.30mV |
| 100% | 27.90mV |
| 110% | 30.60mV |
| +5V ripple quality | |
| Load | Ripple and noise |
| 10% | 10.60mV |
| 20% | 16.10mV |
| 50% | 19.70mV |
| 80% | 27.10mV |
| 100% | 34.80mV |
| 110% | 40.90mV |
| +12V ripple quality | |
| Load | Ripple and noise |
| 10% | 23.30mV |
| 20% | 31.60mV |
| 50% | 34.40mV |
| 80% | 40.20mV |
| 100% | 55.80mV |
| 110% | 76.10mV |
While the voltage regulation was good, the ripple and noise doesn't fare as well. 40.60mV is close to the 50mV limit on +5V; 76.10mV at +12V is okay but still worse than many other power supplies. The 30mV result on 3.3V is also near the limit. All are still within specification, but there's nothing particularly exceptional here; this is just average performance from any decent PSU.
Facebook
Twitter
RSS
82 Comments
View All Comments
Mathieu Bourgie - Thursday, October 28, 2010 - link
I'm with everyone here, crap power supplies are a plague and with the marketing focusing on the Watts "output" (many PSU are overrated as Martin pointed out with the Techsolo Black Mamba 550W), it just gets worse for uneducated consumers.I wish that ATX specs were tighter and that if anything, PSUs should be regulated by an independent commission whose approval you need to sell a PSU (not going to happen, I know), or at the very least to display their label (like 80 PLUS).
Requirements being based on being capable on delivering promised output (At 50C, not 30C or 40C...), ripples within a certain reasonable range, overcurrent/undercurrent/surge/etc. protection to prevent both fires and to protect the PC components, correct wire gauge and other important things that don't come to my mind right now.
In short, a certification that would tell you that PSU has what you should look for and will work reliably as intended. Just like 80PLUS with efficiency, what a great certification. Not perfect, but it's a big step forward.
Stuka87 - Friday, October 29, 2010 - link
Typically the UE and CE certifications should be that governing body when it comes to PSU's. If it is typical for the PSU that burned to have those issues, it should never have gotten certified.Although there is the "Fake" CE mark that looks very similar to the european CE mark, but actually stands for "China Export". I have seen some cheap electronics (and cheap open frame PSU's) with this mark on it before. And of course those PSU's tested out to be crap when ever I hooked one up to my power analyzer.
But an ATX specific spec would be awesome.
brybir - Thursday, October 28, 2010 - link
"This PSU is not available in the US, but it's still a nice representative of the crap-section."While this may be objectively true (as it appears to be), I suppose I expect Anandtech articles to read less like blogs and more like professional evaluations that give the facts up front and well reasoned, supported and defensible conclusions at the end. At least that is what I have come to expect on Anandtech....please dont turn into Tomshardware (i.e. poor writing with sensationalism everywhere).
JarredWalton - Thursday, October 28, 2010 - link
I'll just chime in here by stating that Martin is located in Germany and as such is not a native-English speaker. I heavily edit most of his reviews, but it can get a bit tedious at times to rewrite every little phrase. Perhaps I should have been a bit more creative up front, but I let this description of the Black Mamba go through and apparently you guys didn't like it, so I've reworded it to be a bit more professional. Maybe I was just reminded of Michael Meyers: "If it's not Scottish, it's CRAP!" :-Dlenghui - Thursday, October 28, 2010 - link
Well, great job translating, Jarred, and great job writing, Martin. Great review.I happen to own an OCZ Fatal1ty 550W and a ModXStream Pro 500W, so this article intrigued me. I have owned both for over a year, running 24 x 7 x 365. No problem so far (touch wood).
brybir - Friday, October 29, 2010 - link
Thanks for changing that. I read that whole article and enjoyed the comparison as I am pretty uninformed about PSU and why one is better than the other, but that one little phrase just kept coming back into my mind when I thought about the article. Keep up the good work!softdrinkviking - Sunday, October 31, 2010 - link
it's not "you guys" that don't like it, it's just one prick.you and Martin are both doing a great job, and i absolutely floored that
you are catching shiite for that one sentence.
i mean, how stiff can you possibly be to take offense at that?
Calin - Friday, October 29, 2010 - link
A 550W power supply that blows at 300W actual output is certainly "crap-section" - so, in this specific case, I understand the feeling.dertechie - Saturday, October 30, 2010 - link
Yeah, I actually enjoy the black humor often found in PSU reviews.EddyKilowatt - Tuesday, November 2, 2010 - link
Well the thing performed poorest on the tests, and failed at half rated load... those are supportable and defensible results by my standards.I don't have a problem with Martin's exasperation creeping in a bit early in the review. Everyone with knowledge needs to take a stand against the cheap crap that's always threatening to flood the market and drive the good stuff out. Martin's comment might've been out of line if the Techsolo had under-performed by 10%... but it *failed with smoke* at just half rated load! That's crap, and I'm glad he called it such.