I set up a pair of Dell PowerEdge 2950 III servers last week, each one running a pair of Xeon 5410 CPUs (2.33GHz quad-core), 8x2GB FB-DIMMs, and a pair of 15k rpm SAS drives hooked up to a PERC6 - peak power draw registered by BMC has been 293W so far on one box, and 276W on the other.
None, but desktop systems aren't large banks of FB-DIMMs either, nor multiple quad-core CPUs or rows of 15k rpm fans. The most power-hungry graphics card today is well under 200W power draw. You really, really have to work to exceed 500-600W power draw on a modern computer.
I like most of your reviews very much guys and I have a suggestion/wish for the future reviews.
Can you test (if possible, of course) Inter-Tech CobaNitrox IT-7750SG PSU?
I would also like to see the tests of FSP Epsilon series PSU.
Epsilon has already arrived and will come soon. CobaNitrox might not be interesting for most readers since (I think) they're only available in Germany.
Nice to hear about Epsilon. I have seen about 87-89% efficiency about Epsilon series on several web-sites and I would like to see professional tests about that PSU.
Yes, CobaNitrox is only available in germany, but it comes relatively cheap compared to other high quality PSUs and it is very interesting does it offer same functionality as other PSUs in the same class do?
Sweet jesus, I did certainly not expect those graphs for the 3.3 and 5v lines, with even slight increases, and as for the 12v, well, it stays bang on target.
Corsair, even though they've entered the market late, have seen issues with other PSUs and created the best PSUs in every segment - it seems they can do no wrong.
Assuming a 500W draw on the wall socket, a change from 85% to 86% efficiency, the power lost inside the power supply unit (as heat) changes from 75W to 70W. While this isn't an important figure by itself, this can make is run cooler or be a bit less noisy.
If you take into account the power delivered to the internal components (not the one drawn from the wall socket), this improvement in efficiency is (a little bit) better than that: for a 500W internal load, you would use 588.2W from the wall with the 85% power source, and 581.4W for the 86% unit. While the total difference is small (7W or so), it is still there.
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
17 Comments
Back to Article
Gholam - Wednesday, April 30, 2008 - link
I set up a pair of Dell PowerEdge 2950 III servers last week, each one running a pair of Xeon 5410 CPUs (2.33GHz quad-core), 8x2GB FB-DIMMs, and a pair of 15k rpm SAS drives hooked up to a PERC6 - peak power draw registered by BMC has been 293W so far on one box, and 276W on the other.HOOfan 1 - Wednesday, April 30, 2008 - link
and how many graphics cards are you running?Gholam - Thursday, May 1, 2008 - link
None, but desktop systems aren't large banks of FB-DIMMs either, nor multiple quad-core CPUs or rows of 15k rpm fans. The most power-hungry graphics card today is well under 200W power draw. You really, really have to work to exceed 500-600W power draw on a modern computer.Powervano - Tuesday, April 29, 2008 - link
I like most of your reviews very much guys and I have a suggestion/wish for the future reviews.Can you test (if possible, of course) Inter-Tech CobaNitrox IT-7750SG PSU?
I would also like to see the tests of FSP Epsilon series PSU.
Christoph Katzer - Tuesday, April 29, 2008 - link
Epsilon has already arrived and will come soon. CobaNitrox might not be interesting for most readers since (I think) they're only available in Germany.Powervano - Tuesday, April 29, 2008 - link
Nice to hear about Epsilon. I have seen about 87-89% efficiency about Epsilon series on several web-sites and I would like to see professional tests about that PSU.Yes, CobaNitrox is only available in germany, but it comes relatively cheap compared to other high quality PSUs and it is very interesting does it offer same functionality as other PSUs in the same class do?
Christoph Katzer - Tuesday, April 29, 2008 - link
http://www.planet3dnow.de/artikel/hardware/netztei...">http://www.planet3dnow.de/artikel/hardw...etzteile...HOOfan 1 - Tuesday, April 29, 2008 - link
Another CWT....Powervano - Tuesday, April 29, 2008 - link
Good review, but not the same high quality one as I see here, at AnandTech...JarredWalton - Tuesday, April 29, 2008 - link
Note the author. ;) But of course, that's about a year and a half old.Powervano - Wednesday, April 30, 2008 - link
Oh, my bad :-D Sorry Cristoph. But your new reviews are way better and I even think best of all :)Powervano - Tuesday, April 29, 2008 - link
Thank you! :)piroroadkill - Tuesday, April 29, 2008 - link
Sweet jesus, I did certainly not expect those graphs for the 3.3 and 5v lines, with even slight increases, and as for the 12v, well, it stays bang on target.Corsair, even though they've entered the market late, have seen issues with other PSUs and created the best PSUs in every segment - it seems they can do no wrong.
Corsair, I salute you.
ineedaname - Wednesday, April 30, 2008 - link
Don't forget that corsair PSU's are actually rebranded Seasonic PSU'sDrMrLordX - Thursday, May 1, 2008 - link
Isn't the HX1000 a CWT unit?Christoph Katzer - Thursday, May 1, 2008 - link
Yes it is. Corsair buys from several companies...Calin - Tuesday, April 29, 2008 - link
Assuming a 500W draw on the wall socket, a change from 85% to 86% efficiency, the power lost inside the power supply unit (as heat) changes from 75W to 70W. While this isn't an important figure by itself, this can make is run cooler or be a bit less noisy.If you take into account the power delivered to the internal components (not the one drawn from the wall socket), this improvement in efficiency is (a little bit) better than that: for a 500W internal load, you would use 588.2W from the wall with the 85% power source, and 581.4W for the 86% unit. While the total difference is small (7W or so), it is still there.