Good article, MikeC.

As a note, the word "performance" appears twice in the first sentence of the conclusion.

Quoting: 'Help support this site, buy the Enermax Galaxy EGA1000EWL Power Supply from one of our affiliate retailers!'
No offense meant, but do you really think any genuine SPCR fan will ever buy such a noisy PSU?

That's just a bit of boilerplate that is stuck on the end of every review. The review makes clear this PSU is not appropriate for diehard SPCRers.

MikeC doesn't care who buys what. He's just saying, that if you do buy it, why not support SPCR with it?

I know this is standard advertising, but it still makes me uncomfortable, I would never do such a thing.
It's also a contradiction in terms: SPCR tested it, decided it's a PoS noise-wise but kept the standard advertising.
A small change, like 'IF you buy' instead of simply 'buy' would sort it out for me. Of course, I don't expect this small change to happen...

I don't even know where to buy a motherboard that supports enough devices to warrant this much power!! But in any case, I think products like this are indicative of where the high-end of the market is headed in the next 2-3 years. Both AMD and Intel are gearing up to offer a new platform concept where GPUs become first-class citizens like CPUs are today. High-end motherboards may offer 4 sockets, 3 of which can take a CPU or a GPU. And on the low-end, the CPU will contain graphics pipelines built right on-die.

But the question is, now that graphics is becoming the realm of the big fabbed CPU vendors rather than that of the fabless driver shops of the past, will smaller processes and more sophisticated semiconductor design keep graphics power consumption in check better than it is today? Only time will tell. Will the mainstream gaming market ever require a kilowatt to power their next-gen multi-GPU platforms? It just doesn't seem likely.

For reference, an IBM System Z enterprise-class mainframe, fully loaded, has plenty of headroom with its 1.5 kilowatt power supply. But then again, it won't run Battlefield 2...

Especially since, if they continue to test kilowatt power supplies, they'll need the money to pay their electric bills!!

welcome, arthur armstrong! great article... i enjoyed the wit and humour peppered throughout the article.

hopefully you get a better product to review next time....

fwiw, the comparatives section should be changed slightly just to please the marketing dept at Corsair. The sound recording is for a 520/620 not 500/600.

Also it'd be nice to see a small table around there somewhere showing the min and max SPL for each of the comparatives vs the tested unit.


It points to a trend of higher wattages needing higher cooling which is common sense but the justifiable use of berating people for considering such a beast would be well paired with clear obvious numbers to show the bias is purely for the benefit of the users wallet/ears.

Correction made -- re Corsair.

W/ regard to the min & max SPL dBA@1m numbers, even that is misleading, because most PSUs are pretty loud at full tilt. Comparing the M12 and Galaxy a gamer could say, "At full power the Galaxy is only marginally louder, it isn't that big a difference." What it doesn't show is that from min power to ~300W, the M12 barely changes from a super quiet 21dBA to a very modest 25 dBA. In contrast, the Galaxy goes from 29dBA all the way up to 36dBA. In practical terms, it goes from almost twice as loud to more than twice as loud. And we're not even talking about sound quality at this point.

I know that noise/power output curves would make all this much clearer. I've avoided them up to now because the temperature in the lab does vary from ~21C in winter to 26~28C in August, and the ambient temperature does affect the noise/power curve of most PSUs. Now, I'm thinking maybe we should try to do testing in a narrower temp range (say 22~24C) so that comparable noise/power curves can be shown. We'd just have to avoid PSU testing on hot or cold days.

The link to this discussion in the forum is messed up. I end up at some discussion about SilenX fans :S

I would love to have a recording of that 50dBA PSU! Just as an example of how bad it really can be.

Good review, but could we tone down the sarcasm a little bit? The first two pages are full of it. It's getting old and gets bad reputation. I understand that Enermaxes are not as good as Seasonics noise wise, but you don't have to rub it in before doing any actual testing. It just makes the article look unprofessional.

Hmmm... I thought it was mostly just in the Feature Highlights tables where we often have fun. Arthur's comments were kind of funny to me... And then, of course we have a differing reader's opinion about this too:

I think this sentence might use some rewording, as it could be quite confusing for an SPCR newbie:

First, ball bearing fans are used even in some(?) of SPCR Recommended PSUs (like the Seasonic S12-430). And it should not be a surprise that such a powerful PSU comes with ball bearing fans, as sleeve bearing is not suited to high temperature environments.

Second, it's clear plastic frame that is bad for acoustics, not wings. There are excellent fans sporting clear plastic wings.



As it is now, it has too much of a "I'm smarter than you" attitude for my taste.

Yes, but take an average ball bearing fan and an average sleeve bearing fan and the sleeve fan is likely to be quieter. That is the only point that sentence is trying to make. The S12 series happens to have a very good BB fan, but I have still swapped mine for a Yate Loon, as the BB fan still has quite an audible noise signature.



um, do we actually have any evidence of this? I know this is the received wisdom on SPCR, but I have no personal experience that would verify this (always used Yate Loon black). For example, the Acoustifan C-series was a very quiet fan with clear frame.

What a worthless product for anyone who doesnt have an abnormally small penis.

However for all of the underendowed its $300 dollars well spent


Andy

I agree the tone/sound quality isn't shown in the min/max nor is the curve of the RMP/dBA ramp up shown with only two points. I was really thinking of a bigger table or a nice color graph.



Sometimes I'm just happy to have it in the article discussion thread. So long as I play devil's advocate enough to keep you talking about it, we all gain from the discussion. Especially the lurkers that just pop in here from a Google hit.

I used to think this site took it too far back when "THE article" was touting 200W PSUs (or was it 250W? it's been years and the article changes with the times). But now I am in the camp recommending PSUs in the 380 to 520W range to people that are discussing 850W to 1000W monsters.

Again thanks for all the hard work.

I also enjoyed the wit and humor in the article

I think this whole site has the "i'm smarter than you" vibe, you piss on anything that isn't "silent." (which is great imo) Previous reviews have that sort of style too.

I look foward to the reviews because of that. Most hardware reviews are a bit bland, this site spices it up a bit.

As a matter of constructive criticism, I'm in the too-much-sarcasm camp. For that reason, I found Arthur's writing style different than that of the rest of the SPCR staff. Of course, that's just my personal opinion.

Thanks to Arthur (and Mike C) for the review.

At every other site in the world, I ignore the "Feature Highlights" section because I know it's just market-speak and not relevant to the real world. However, I always read them at SPCR because of the funny comments. Love it. Keep it up!

I have numbers of a
Intel Core 2 Extreme X6800 (2,93 GHz, Dual-Core)
nVidia nForce 680i reference board
2x nVidia GeForce 8800 GTX (575/1350/900) SLI

power draw from wall (220 V):
idle 295 W
load 445 W

The X1950XTX Crossfire with FX-60 and nforce 4 is 228/401 W.

On the Cebit Thermaltake showed a 1500 W PSU and Topower showed an external 2000 W unit.
http://www.computerbase.de/news/hardwar ... 2000_watt/

Arthur's ears/eyes must be burning.

What was suggested to him when asked how he should write this first piece was: Think of it as a technical report that enlightens and entertains. Perhaps he took the entertain part a bit farther than others have done in the past, but the technical report part is still all there. Mind you, this was a tough first piece to write for SPCR, because not only are the PSU tests about the most complex, this particular PSU is more complex and ambitious than most.


Your comment surprised me a bit. I mean, the name of the site defines our specific angle on IT gear, even though our perspective forces us to look at almost every aspect of it. Most sample suppliers are well aware of our focus, and if they send products that don't meet our criteria, they know what to expect: We get disappointed, as do the readers. I know we get sarcastic sometimes, partly because of how often we see a huge gap between marketingspeak drivel and the reality... but I hope we're not strutting around with the "i'm smarter than you" vibe going all the time. That's not our intention.

---------------

Moving on to another article linked in the first page, did anyone read through The Tech Report's piece on 8800s in dual SLI?

As I mentioned in the intro, The Tech Report did some power numbers on dual SLI C2D 6800EE systems; the dual 8800GTX SLI system pulled only 373W AC & the dual X1900XTX system pulled 400W. DC power draw was ~300W and 320W. That's less than half of the min. 750W "nVidia certified PSU" for dual 8800GTX SLI.

Amazingly, the authors still wrote that although their 700W PSU worked fine, "serious" users should use a 1000W PSU (like one they showed in the article) -- mainly to ensure there are enough connectors for everything, rather than perhaps being forced to use adapters or extenders. They also failed to mention that even a good 430~500W PSU would have done fine. I found this last omission a rather big surprise.

So it is not just that macho gamers overestimate their power requirements -- I mean that's perfectly normal. But they're actually being told they need >750W PSUs for dual 8800 SLI directly by nVidia. (There are no PSUs below 750W rating in SLIZONE's list of certified components for dual 8800GTX SLI. They range 750W to 1100W.) And expert sources like the Tech Report.

The SLI certification looks like marketing bs & collusion to me -- nVidia working hard with PSU partners to push higher power, more profitable models to the consumers... when nowhere near that much power is needed.

PS -- jojo4u, you posted while I was writing.... Too bad that Computer Base article doesn't list the PSU used. The AC power demand is 72W higher than the very similar system setup used by The Tech Report; perhaps the CB PSU was not as efficient?

They don't list it, but on page one they tell about it:

PC Power&Cooling 1KW Quad

EDIT:
Keep in mind, this PSU ran on 220 V.
IMHO Techreport recommends solely the 1 kW PSU because of the 4 PCIe power connectors.
Nice article about overclocked quad-CPU and 8800 GTX SLI setup on 500 W PSU: http://www.theinquirer.net/default.aspx?article=36066

my comment on this whole humour and professionalism discussion, is that dishonest marketing deserves brutal and unrelenting ridicule. when it came time to be objective and serious, i feel that the author did so. but it's all a matter of personal tastes i suppose...

but the one thing that confuses me is that the last time SPCR posted an unfavourable review of a product (see: http://www.silentpcreview.com/forums/vi ... hp?t=33055), i remember asking why they even bothered to do it, given that SPCR has a policy not to post products that fail miserably. the answer then was that too much time had already been invested in it to allow it to go to waste.

now again we have a product that fell well below the SPCR standard yet is posted anyways. the strange part is, that it was almost a foregone conclusion that this thing would be hellishly loud. so why post the review?

the only reasons i can gather from the article itself, were that a) it was a chance to test out the revamped testing rig, and b) there were many requests from the SPCR community. but i can't imagine there being that many people would actually consider this to be a potentially quiet component...

just curious as to the current policy on scathing reviews...

Sometimes we scratch our heads too!

Here, both a) and b) were true -- though I am sure no one thought it would be a quiet machine, there were some who hoped it might be OK at moderate power levels (you know, before the helmets and guns go on along w/the headphones.)

But there's also c): Persistence on the part of the sample supplier. This was not the first but the second sample of the Enermax Galaxy that was pushed to us. At some point we bow and say OK, we appreciate that you value our review, we'll do our job.

There's also d) some educational and hand-on learning factor, too: If we're swimming up river regards recommendations for power, it helps to be able to say, yes, we've looked at the whole range of products available, not just a select few.

And finally, we're not blind to e) the potential of an extreme product such as this to draw traffic, including new traffic, to the site. In media as in academia, publish or perish is the rule...

This is a good point. When people come on saying "Hey why don't I get this mammoth PSU?" we can point to the review. If we didn't have that they would just accuse us of being blinkered and not willing to consider PSUs that don't fit the low-power SPCR "orthodoxy".

A 1KW PSU isn't that much of an overkill, actually it pales in comparison to the OCZ 2000W PSU.

When will the insanity end?

Assuming 75% efficiency at 2 kW, that PSU at full load is enough to make my safety switch trip

Looking back at the most recent PSU reviews, I see:
Enermax 1kW
Lian-Li 600W
Tagan 700W
Antec 550W
Ulltra X-pro 600W
Corsair 520W and 620W
Antec 430W (from a series which also included 380W, 500W supplies)
Dark Power Pro 430W
Seasonic M12 700W (from a series including 500W, 600W)
Seasonic S12 550W, 650W.

Where are the ~300W PSUs? The fanless ones? The 150W SFF ones? I feel like I'm seeing a constant progression of "here's a monster PSU - it is really noisy" reviews. What is the point? I was never going to consider that PSU anyway.

I suppose ~420W is OK if it is the only way to get the features you want (e.g. 120mm fan, modular cabling) and the efficiency is OK in the 50 to 150W range.

If you're going to make a big comparison table (as others have suggested in this thread) I suggest efficiency and noise at 50W and 150W would be the appropriate statistics to use. (maybe stretch to 250W for the quite-but-not-silent gamers.)