View Full Version : Enermax Galaxy 1KW @SPCR
burebista
03-19-2007, 04:35 AM
Review (http://www.silentpcreview.com/article726-page1.html).
The Enermax Galaxy has excellent features and near-flawless electrical performance performance. The Power Guard worked wonderfully, helping raise the bar with benefits like a tidy fault detection and explanation system. Voltage regulation was executed with finesse, and ripple was superb. These are all good things. If the dual 12V transformers are part and parcel of this package, well, that's good and fine, too.
The unit was a source of terrible noise, hitting the top right around where it achieves best efficiency. Efficiency only barely reached 80% on our test rig, and it suffered badly at loads under ~100W; efficiency is the one aspect of electrical performance that's less than tops.
CAD4466HK
03-19-2007, 05:29 AM
I find this interesting because of alot of complaints about the "Power Guard" and low +5V & +3.3V loads not allowing some people too POST :(
5. LOW LOAD / CROSSLOAD PERFORMANCE
Our usual low / no load testing could not be done. That little Power Guard guy does its job quite well. We were prevented from testing at very low power or using no load starting conditions. Later on, we were held back from overdriving the system, too. Best of all, when we were in proper operating range Power Guard left us well enough alone. Included was that handy-dandy light and beep setup to explain the source of an error. A clear table in the manual explained the meaning of each color and light combination. The Power Guard shows us how it's done.
Neither could the crossload test be performed. Our usual procedure is to obtain ~75% load, almost entirely on the 12V lines, with only one amp each on the +5V and +3.3V lines. With a 1000W PSU, this is a pretty extreme load, some 750W in total. The Power Guard simply refused to let the PSU be turned on with this load, even if we built up to it by adding loading incrementally while the unit was running; it would shut down when the +5V and +3.3V lines got too low or when the 12V load got too high. If such conditions are not safe for the PSU, then the Power Guard's protective action is a good thing. Besides, it's hard to imagine how any system would demand over 740W on the 12V lines while pulling under 10W on the +5V and +3.3V lines.
"My computer has two Uber-X Graphics Cards and a kilowatt power supply!"They need to grow up and become open minded. Constricted streams lack water flow and nutrients to the area's that most need them ;)
Another - who in this modern world will expect silence at even 350W load? Thats BS and I have no idea why they don't just make a watercooled "silent" PSU to stop nagging about standard expected noise. If someone goes for a product which demands such power, they have automatically accepted noise pollution as their game. Heck, lets make a silent car that beats the Shevy with the 273mph top speed or a VanishingPoint that sounds 40dBA :crazy:
That said, there are obviously differences within a power supply from one to another - some like the SS OP1000 I saw was very noisy and others much less at the same load.
Our most power-hungry Intel 670 (P4-3.8) processor rig with nVidia 6800GT video card drew ~214W DC from the power supply under full load — well within the capabilities of any modern power supply.Well that's perfect futureproofing isn't it now :(
The efficiency was a bit of a letdown. It seems the marketing department might have been overzealous with their "80-85% efficiency from 20% to 100% load" claimVery odd and interesting how they get 53%-79% efficiency and at higher VAC input @ <30% loads, yet I've seen similar testings prove 77-80+% efficiency from lowest to highest load. Seems like there's a big difference from one to another unit, test setups, or something other which doesn't do too well for the concept of reviews. :o
I find this interesting because of alot of complaints about the "Power Guard" and low +5V & +3.3V loads not allowing some people too POST :(DFI is the only board I've heard this of so far.
CAD4466HK
03-19-2007, 05:41 AM
DFI is the only board I've heard this of so far.
I know people in person with this problem ;)
Spectre
03-19-2007, 09:49 AM
Hmmm....I have one here that I am going to be doing soon. No load starting isn't a big deal though as this isn't really something it should be doing anyway....but the low load.
SKYMTL
03-19-2007, 10:24 AM
I know people in person with this problem ;)
Add one here. It is really odd too. Enhance-labeled 400W and 500W units refuse to boot on my computer but the SAME Enhance PSUs with a Silverstone label (both 400W and 500W) boot without a problem.
KorruptioN
03-19-2007, 10:28 AM
Another - who in this modern world will expect silence at even 350W load?
The same people that would use a S12E+ if they wanted to power a heavy load. Silence and efficiency are high on my list.
GalvanizedYankee
03-19-2007, 12:11 PM
SPCR has done good SMPS reviews for quite awhile now. My question is, why do they bother reviewing gaming or server units? Is it to feel smug about the silent units, better said, quieter units they seem to like? Anyone "needing" an 800~1000W unit will have other noise generators to mask a simple PSU fan.
Yes I did read the review, to include the first few paragraphs.
mikec@silentpcreview.com
03-19-2007, 12:24 PM
Another - who in this modern world will expect silence at even 350W load?
Low noise is a feature that's given at least lip service by just about every retail component maker today. It's not a small market segment, most people will go for lower noise if there is a choice, often willing to pay a little more. Low noise is perceived by most users as a mark of quality... luxury, even.
There are a handful of powerful PSUs tested at SPCR that remained very quiet even at ~300W load. Examples:
1) The Seasonic M12-700 started at 21 dBA@1m & climbed to just 25 dBA@1m at 300W.
2) Ditto the Seasonic Energy 12 Plus 650.
3) Corsair 520/620 stayed at 22 dBA at 300W and was just 29 dBA even at 400W.
Admittedly these are all from one mfg... but there's a couple others almost as good -- Silverstone Element ST50EF-Plus, BeQuiet, etc.
We're all grown up, btw, and there's nothing about close-minded about what was written in the article about the mythical need for >750W PSUs. I've studied many graphics card and gaming system reviews in all the major review sites and have yet to see any clear evidence that such power recommendations are valid. It's quite possible that some PSUs are unstable with some loads, but it's usually an issue of power quality, not quantity. As is usually the case, high power is actively pushed and encouraged by vendors because it benefits their bottom line. People buy into it mostly to "feel good". [An aside: Like a distant friend who just bought a V12 Jag for use in the clogged streets of HK where you mostly sit motionless in traffic and are lucky to break 30mph.]
Oklahoma Wolf
03-19-2007, 12:30 PM
I've studied many graphics card and gaming system reviews in all the major review sites and have yet to see any clear evidence that such power recommendations are valid.
Indeed - I know at least one person running an SLI rig with 8800 GTX cards off the same Silverstone ST56ZF that's powering the computer I'm using to post this. I also know for certain the Corsair 620W can power the GTX's in SLI as well. Whether or not power needs will go up with the R600 remains to be seen, but I for the most part haven't been recommending anything bigger than 750W.
Welcome aboard btw :)
GalvanizedYankee
03-19-2007, 12:47 PM
Low noise is a feature that's given at least lip service by just about every retail component maker today. It's not a small market segment, most people will go for lower noise if there is a choice, often willing to pay a little more. Low noise is perceived by most users as a mark of quality... luxury, even.
Those "luxury" fans have helped alot of low noise Antec SmartPower and TruePower units to an early grave. Some even take other hardware with them on their way out. :D
Point is, a less than unit can be precieved as being quiet luxury with a simple fan change. ;)
Welcome to this board Mike Chin. :beer:
jonnyGURU
03-19-2007, 12:51 PM
Hi Mike!
I just want to throw my 2 cents in the ring here. Not that I claim to be smarter or have more experience than anyone here, but too often I find when there is too much focus on "silent" the longevity of the equipment is sacraficed. Many of these PSU manufacturers are willing to sacrafice low noise at high loads so their product can live out their 3, 5 or lifetime warranties. I appreciate the pursuit for low noise, but I don't think many in that pursuit are really thinking of long term.
As for power consumption, I do agree that many people over guestimate their power requirements. I, for one, have a pair of 7800GTX cards running on an X-Pro 600W. Before that, I had my trusty external SevenTeam 500W powering everything for nearly a year with no issues. In my lab, I have a 7950GT OC system that can run 3DMark'06 all day long (decent scores too) and not pull more than 200W from the wall. That whole rig is powered by a 5.25" bay 300W power supply. So I'm with you on that issue. ;)
mikec@silentpcreview.com
03-19-2007, 01:10 PM
Those "luxury" fans have helped alot of low noise Antec SmartPower and TruePower units to an early grave. Some even take other hardware with them on their way out. :D
Point is, a less than unit can be precieved as being quiet luxury with a simple fan change. ;)
Welcome to this board Mike Chin. :beer:
Thanks for the welcome, and yes, you're right about those Antecs, they weren't the greatest, and they were often nasty whiners (high pitched coil/capacitor squealing).
I agree that simply running a quieter, slower fan on an average PSU will just make it die sooner, which is why we examine the noise and the temperature rise throughout the power range. Seasonics, which remain at the top of the quiet heap, have been very reliable and long-lived for me; they've been used in most of the systems I've built for myself and friends over the past 4 years. They have the best fan curve: Virtually no fan speed rise to a certain temp, then a fairly rapid rise beyond -- ie, the classic elbow. The latest Silverstones are similar, along with a handful from other brands. High quality components and good heatsinks/cooling desgn can be used to bring overall noise down without affecting longevity. The extra headroom allows lower airflow to be used.
mikec@silentpcreview.com
03-19-2007, 02:02 PM
Hi Mike!
As for power consumption, I do agree that many people over guestimate their power requirements. I, for one, have a pair of 7800GTX cards running on an X-Pro 600W. Before that, I had my trusty external SevenTeam 500W powering everything for nearly a year with no issues. In my lab, I have a 7950GT OC system that can run 3DMark'06 all day long (decent scores too) and not pull more than 200W from the wall. That whole rig is powered by a 5.25" bay 300W power supply. So I'm with you on that issue. ;)
Hi Johnny,
Good to hear your experiences and agreement on this power issue.
It is not just that "people over guestimate 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 (http://www.slizone.com/object/slizone_build_psu.html). They range 750W to 1100W.
The Tech Report did some power numbers on dual SLI C2D6800EE 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). They also failed to mention that even a good 400W PSU would have done fine.
It looks like people are believing the >750W "requirements" because that's what the experts and the companies are telling them. 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.
Spectre
03-19-2007, 02:07 PM
Hi Johnny,
Good to hear your experiences and agreement on this power issue.
It is not just that "people over guestimate 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 (http://www.slizone.com/object/slizone_build_psu.html). They range 750W to 1100W.
The Tech Report did some power numbers on dual SLI C2D6800EE 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). They also failed to mention that even a good 400W PSU would have done fine.
Mike you are assuming here though that everyone is using a good 400w PSU that actually has neccesary ~75% of it's capacity on the 12v rail/rails with no derating curve. That simply isn't realistic for most users.
mikec@silentpcreview.com
03-19-2007, 02:37 PM
Mike you are assuming here though that everyone is using a good 400w PSU that actually has neccesary ~75% of it's capacity on the 12v rail/rails with no derating curve. That simply isn't realistic for most users.
But it's more realistic and affordable than a 750W or 1000W unit, no? ;)
Most 400W ATX12V v2.xx PSUs have at least 300W capacity on the 12v lines. OK, perhaps 400W is optimistic. Call it 500W then. It's still way below the 750W min recommended, isn't it?
yeah, but the moment they start saying that a good 500 is enough, people will start running sli setups on $30 "500W" units, as "a 500W is enough". i think that part of the reason they only reccomend 750W+ units is to take stupidity into account.
Mantralord
03-19-2007, 03:46 PM
yeah, but the moment they start saying that a good 500 is enough, people will start running sli setups on $30 "500W" units, as "a 500W is enough". i think that part of the reason they only reccomend 750W+ units is to take stupidity into account.
Except that the units they recommend are specific and expensive units.
Also I'm glad MikeC is here.
Except that the units they recommend are specific and expensive units.
true, but do you really think that the average user is going to know that?
there is a big role here for people on forums and reviewers to teach people the difference between good and bad models, but the large majority of people think that all psu's with the same wattage rating will be equal, and will not stop to think about the price difference.
Makalu
03-19-2007, 05:59 PM
Well I certainly agree that the typical 8800GTX SLI system doesn't consume more then around 400W. Lots of people have measured it and that's what it is...albeit it could theoretically go a bit higher if and when DX10 is fully optimized. Perhaps another couple of 12V amps. But anyway, what I wanted to mention here is that I think...not sure...but I think that Nvidia is only certifying units for 8800GTX SLI if they come with the necessary 4 PCI-E cables since using adaptors would open up a big can of worms. There are units that Nvidia has certified for less then 8800GTX systems that do now come with 4 PCI-E connectors and are capable of powering the systems but I think they aren't certified because the manufacturer's haven't submitted the new 4 cable versions. Just speculation here on my part though.
Spectre
03-19-2007, 07:33 PM
Except that the units they recommend are specific and expensive units.
Also I'm glad MikeC is here.
And that doesn't keep people from reading a box and going ooooohhhhhh shiny 680w Aspire Beast...it'll run my 8800GTX.
The right 400w or 450w unit will run it no problem. The problem is the majority of 400w and 450w units aren't and even some of the units on the SLI cert page are bluntly crap compared to other uints of equal or less capacity available while units far more capable aren't certified for higher levels because they don't have enough connectors.
It is not just that "people over guestimate 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 (http://www.slizone.com/object/slizone_build_psu.html). They range 750W to 1100W. Even worse, there are these calculators telling everyone they need half kW PSU for machine which barely peaks above half of that.
The right 400w or 450w unit will run it no problem. The problem is the majority of 400w and 450w units aren't and even some of the units on the SLI cert page are bluntly crap compared to other uints of equal or less capacity available while units far more capable aren't certified for higher levels because they don't have enough connectors.I suspect SLI certification consists only from meeting some arbitrary paper requirements.
BTW, Here's best pick ups from one thread in Finnish MuroBBS:
Pentium-D@4.05, X1800XT Crossfire, 4HDs: Antec Neo HE 430W
FX60, overclocked X1900XTX Crossfire: Antec Neo HE 380W
Now what might have been recommended ratings for those PCs?
The same people that would use a S12E+ if they wanted to power a heavy load. Silence and efficiency are high on my list.Thats a quieter consideration not a dead silent consideration that I mention when you quote me. ;)
Low noise is a feature that's given at least lip service by just about every retail component maker today. It's not a small market segment, most people will go for lower noise if there is a choice, often willing to pay a little more. Low noise is perceived by most users as a mark of quality... luxury, even.Yes I realize Mike and I agree on many points. What I don't agree on is judging a PSU on the merits of it acoustics primarily. If it meets the rated labelling and does well in voltage regulation, build quality, ripple/noise signal departments, PFC and efficiency, and noise is not like the leaf blower I only once in a green moon want to hear; I am damn happy about it :) But this doesn't include striving for dead silence or below 35dBA @1m with powerful systems for me and that is what I comment on and question.
The SeaSonic units you mention I actually own and can attest to their silence being one of the remarkable features. ;) Do I like that? Of course!
But if I've got a system running 70-100% load, I don't consider noise to be the primary crucial factor in connection with a PSU. My first would be many more, like performance :p and then after all due variables I would definitely consider noise. I wouldn't demand it to be whisper quiet but something like 45dBA below will do just fine. This is where I turn to your testings and ratings, so thanks :D Although IMHO if commentary was more based on performance/quality first and then noise, it would do the PSU more justice as to what it was built with in mind.
We're all grown up, btw, and there's nothing about close-minded about what was written in the article about the mythical need for >750W PSUs.I have no issue with that at all actually, but what I commented on was this:
"(nVidia IIRC)...and others are meant to spur those who feel inadequate unless they have more horses under the hood: "My computer has two Uber-X Graphics Cards and a kilowatt power supply!" We think it's silly and wasteful."
Everyone who gets a plus >750W PSU doesn't need to and don't necessarily have that attitude. That is kiddy minded outlook and may only apply to a handful so the generalization wasn't of mature nature.
The Tech Report did some power numbers on dual SLI C2D6800EE 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). They also failed to mention that even a good 400W PSU would have done fine.
I actually reported that link to this forum pointing out that most people recommend too much etc. ;) But this is where I'm going to stop you, and consider this another myth debunked which comes with strong belief usually. Facts don't always have to be so bipolar as beliefs seem to be nor carved in stone. One top system may not have pulled <250 or <350W from the power source in total, but although >750W is too much, people recommend 400W PSU's just like that for the similar setups and its inacurrate. Systems differ drastically and some pull more on very similar setups. For those people we're talking about, near enough every one of them will OC. This increases the power draw. And the ho-hum; a PSU working at 50% load is much easier for it and runs cooler (less stress>less heat>less noise>more lifetime) than when its at higher stress loads, let alone 100% max and the consideration of future proofing, I'm sure you know. Here, let me show you just one instance where one set of figures for total power consumption are not applicable universally and do not agree with "400W plenty for any" notion. OC'd quad core and GTX SLi system pulling 546W under stress, etcetera:
http://forum.coolaler.com/showthread.php?p=1674466#post1674466
http://forums.slizone.com/index.php?showtopic=405
http://www.extremetech.com/article2/0,1697,1937996,00.asp
http://www.extremetech.com/article2/0,1558,1932947,00.asp
http://www.anandtech.com/video/showdoc.aspx?i=2870&p=18
http://tomshardware.co.uk/2006/11/30/brute_force_quad_cores_uk/page4.html
http://tomshardware.co.uk/2006/11/30/brute_force_quad_cores_uk/page12.html
Also see what the AMD lab tester had to say about power requirements. Thats just one or two links I saw as I linked one today morn to here. There have been more and user tests that affirm, one individuals system may only require 400W AC draw while another will need 600W, with very similar best system setups. I've seen this many times with my own eyes so I can't go in-denial about it. If I can get someone to get my similar setup back home working and running, I'll provide you figures of my power draw which are higher than 650W too.
I also believed what you do primarily, but figured it out the hard way. One of my own systems showed exact signs of suffering with a SeaSonic S12-650 about 4-5 months back which I did report in a few places online. It wouldn't even open Windows Explorer (Vista) alongside the browser (10 tabs) and Thunderbird with OC'd QX6700, watercooling, XFX8800GTX in SLi etc and without shutting off. The PSU is fine and I have it running in a system right now for a FX57 with X1950XTX and a bunch of components. Moving to XP SP2 gave the same results with 3 browser tabs open and running work related software called Houdini. After troubleshooting everything it finally came down to nothing, but the PSU being underpowered somewhere. Tried everything I had that was 650 or below from server PSUs and work PSUs, nothing worked. Only had a Enermax Galaxy 850 free from a workstation that had higher wattage and popped that in, system ran without a glitch even during benchmarks! I know Jon from here and others faced very similar problems and solutions at least once.
To date, I only know a few sites/people, such as a team in Germany and Russians site at X-bit Lab that do accurate testing of GPU power consumptions. Their tests reaffirm such findings along with yours that its not a "mine does 200W full load so will everyone elses" sort of situation. Its trickier with differences.
It looks like people are believing the >750W "requirements" because that's what the experts and the companies are telling them. 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.Yeah I agree with you there. What they do is work out a guesstimate and then add 30%-50% to that :D
If mine needed 300W though, I wouldn't give it 300W but at least 100W more. Looking at price per watt, efficiency, future proofing (eight cores and so on), addition of components, will still hinder me from getting a Jeantech 250 over a SeaSonic 330, taken in terms of power. Furthermore, when kept under typical-high loads for lengthy periods, all electronic components wear out by time and power output, capacitance, efficiency (etc) decreases with product quality deterioration. For a person like me running critical work 18hr/7 at least, I'd add a little more total watts for security at least. I'm not a gamer BTW, its work related and thats why more impedance testing (http://cp.literature.agilent.com/litweb/pdf/5950-2949.pdf) needs to be carried out on integrated circuit designs and components than currently being done I feel... thats just my opinion though.
Gah! A bit lengthy but my original one was lost when forum error'ed.
mikec@silentpcreview.com
03-21-2007, 12:52 AM
No argument from me about your (and others') troubles with SLI setups. I haven't had direct hands on experience with such problems. All I can do is shrug and say I'm glad I don't have them. ;)
What I will say is about what SPCR recommends: I assure you the top ranks are represented by quality products that tested very well electrically and were also very quiet at least to to a mid power level 2-300W. Just because we consider noise the ultimate pass/fail doesn't mean we disregard all the other criteria. We wouldn't keep adding tests for them otherwise. :D
Spectre
03-21-2007, 10:01 AM
I suspect SLI certification consists only from meeting some arbitrary paper requirements.
;)
vBulletin® v3.8.1, Copyright ©2000-2012, Jelsoft Enterprises Ltd.