PDA

View Full Version : Corsair HX850 Review


Oklahoma Wolf
06-14-2009, 12:33 AM
It seems an eternity since the last time Corsair released a modular power supply in their high performance HX line of units, and there has been a good deal of speculation around the net on when Corsair was going to finally release something in the 750W-850W range to better compete with the likes of the Seasonic M12D and Antec Signature units at that level.

Well, the waiting is over because Corsair is releasing two new units at 750W and 850W to fill the void. Today, I'm looking at the larger of the two units in the HX850.
http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=153

Hondacity
06-14-2009, 01:05 AM
its the power supple to have :)

dual gt300s here we come :)

awesome efficiency :)

excellent job OKW

Travis
06-14-2009, 01:13 AM
Again it's proved to be awesome. Ripple of minor rails stayed low even in CL1, which is really great.

I think the low-load efficiency test should include two to three load patterns: 50W/75W/100W.
50W seems not so pratical since it only happens on MoDT and Nettop platforms where people don't use such expensive and large units, but taking entry-level Desktop platform (and future DT platform with Hybrid mode GPU and C6 mode CPU which consumes much less power at idle) into consideration, 50~100W range should be a good reference.

Just above the blue capacitor is a series of connector solder points labeled "SR CARD." I can only assume that "SR" stands for Silent Running, which might be a godsend for us Mike + The Mechanics fans. Once they implement this feature, we may never have to shed another tear at not being able to hear Paul Carrack sing again. Thanks, CWT and Corsair!

Of course, if all it does is play Silent Running over and over, well, then we might have a problem.

You definitely know what that "SR" stands for. ;)

The Mediocre:

take the children and yourself and hide out in the cellar... I'm about to break out my Mike + the Mechanics collection and I'm not going to be shy with the volume knob

:confused: Didn't get it.

Hondacity
06-14-2009, 01:49 AM
poll - yes...the more results the betttttterererererererer

burebista
06-14-2009, 02:14 AM
Oh boy! What a PSU! And what an enjoyable review. :D
Thanks OW, amazing Corsair unit.

And for low load efficiency test I'm with Travis. A 75/100W should be tested. Usually' who needs a 7-800W PSU I guess that it's not idling at 50W. :D

Breogan
06-14-2009, 03:57 AM
Impressive ressults. Taking in consideration the estimated retail price around here, which sits around 170€, it beats the Revolution 850 (260€), the SG850 (235€) and the M12D-850 (219€) by a wide margin.

P.D. - I think the efficiency test at such a low consumption is only a gimmick... it'd be interesting to see the efficiency at a fixed 100W for all PSU tested, which sits around the typical power usage of a mid-level computer during idle status.

kiri
06-14-2009, 06:33 AM
A 30 to 35W AC Test would be nice and of course a 100W AC Test.
Taht is exactly the amount of power that my computer is pulling from the PSU. ;)

And a nice review of the new Corsair unit. You can buy PCIe to EPS12V adaptor by the way, so you only need one PCIe Connector for that second EPS12V and with six PCIe Connectors the PSUs sure has enough.

Oklahoma Wolf
06-14-2009, 09:14 AM
:confused: Didn't get it.

Try the link in my sig ;)

Travis
06-14-2009, 09:37 AM
Utube is blocked here :<

Spectre
06-14-2009, 09:45 AM
Utube is blocked here :<

Well just know this wolf has bad taste in music.

Oklahoma Wolf
06-14-2009, 10:00 AM
Utube is blocked here :<

It's part of the Silent Running lyrics.

C'DaleRider
06-14-2009, 10:12 AM
Shall I keep doing that low load efficiency test?

First, another excellent review, OW!

Second, I'd honestly say about the low load efficiency testing completely depends upon the power supply chosen.

Given that almost without exception, someone who choses an 850W power supply isn't going to be drawing 30W at idle, I'd think it's rather a waste of time for high power supplies. My rig isn't what I'd call a powerhouse or heavy power drawing rig, but it never draws under 200W, so why would I even care about efficiency at 30W/50W or such?

On the other hand, low power units, like 350W or thereabouts, could and would be the more likely units that would be used in a low power system, so then a low load efficiency test would be much more relevant.

Up to you, but I think your time would be better served with that sort of test by biasing it toward the units that would be the most likely to be used in those situations. Who'd buy a Corsair 1000HX and then use it in a system that has a max load of 150W?

Just sayin'.........

Travis
06-14-2009, 10:58 AM
I just don't know the relationship between the lyrics and the power supply, especially when it's in the summary of the review.

Oklahoma Wolf
06-14-2009, 11:05 AM
There is no relationship. Often when I can't find a mediocre point to complain about, I'll put something off topic in there.

370forlife
06-14-2009, 11:53 AM
The random babble he puts in the reviews makes them very funny and different, part of the reason I always read the reviews here on jonnyguru.

Oklahoma Wolf
06-14-2009, 12:27 PM
The random babble he puts in the reviews makes them very funny and different

Less tedious to write, anyway ;)

RagingDragon
06-15-2009, 12:04 AM
I like the low load efficiency test. The 5% load may be a bit extreme, but that makes it a worst case result, so it's OK with me. But perhaps others would prefer a 10% load test?

It's good to see some serious competition for the Antec SG and CP series - the competition should make top quality power supplies more affordable, and hopefully encourage other brands/manufactures to increase their quality as well.

Redbeard
06-15-2009, 01:16 AM
As a consumer speaking - I like the idea of showing efficiency at 10% load. I agree that 5% is a bit extreme.

I like the idea of showing the real efficiency curves though - from 10% to 100%. It would also be nice to show real-time power draw of system components, though. So people would stop assuming that their system draws 50W at idle and 1000W at load. One of the first things I was surprised by, a few years ago, was how the difference between idle and load on high-end desktop systems wasn't as drastic as I assumed it would be.

It's big, sure, sometimes double, but it wasn't like 110W idle and 500W load. A system that drew 500W load was always almost half of that at idle.

But again, it's been a couple years since I did this testing and I suppose now with Hybrid SLI and other low-power consumption things it might make more difference now.

blivit4
06-15-2009, 07:00 AM
With things like Speed Step with the CPUs and similar schemes in the video cards the difference between idle and load can be over 3:1. HT4U (http://ht4u.net/reviews/2009/power_consumption_graphics/index13.php) did direct power measurements on 20 current video cards and some of the high end cards have > 1:4 idle/load ratio.

I know with my HTPC (E5200 with a OC to 3.2GHz & Speed Step enabled, and a 4670 video card) I draw from the wall 60w @idle, 85w watching 1080p TV via a tuner card and 165w running Furmark and Orthos concurrently. With my 80+, 400w PS it's not even getting to 20% load watching TV (most common task).

HOOfan_1
06-15-2009, 08:20 AM
Notice that the picture of the unit on the back of the box, is not sleaved all the way into the housing....obviously Corsair must have seen the feedback on the web and gone back to CWT and required them to sleave all the way into the housing.

I guess saying the capacitors are "solid-state" is better than saying they are "Solid Snake". :D

Super Nade
06-15-2009, 08:36 AM
Personally, I don't really care for efficiency at 50W or 100W. I', more concerned about how well the unit regulates. Yep, I'm an evil tree burner..;)

EekTheCat
06-15-2009, 08:52 AM
Personally, I don't really care for efficiency at 50W or 100W. I', more concerned about how well the unit regulates. Yep, I'm an evil tree burner..;)
What he said.
Good job, OW.

Smirnoff
06-15-2009, 09:21 AM
Notice that the picture of the unit on the back of the box, is not sleaved all the way into the housing....obviously Corsair must have seen the feedback on the web and gone back to CWT and required them to sleave all the way into the housing.


Dunno, but to me, the unit on the packaging looks more like the HX1000, because the fan is smack in the middle, unlike on the 850. And the power button which you can almost make out, is the big chunky type. :confused:

HOOfan_1
06-15-2009, 09:38 AM
Dunno, but to me, the unit on the packaging looks more like the HX1000, because the fan is smack in the middle, unlike on the 850. And the power button which you can almost make out, is the big chunky type. :confused:

You're right, the power stich also rocks horizontally instead of vertically, and the blue connectors are right next to the hardwired connectors, unlike on the HX850

Mantralord
06-15-2009, 02:30 PM
On one hand, there's no point in being overly paranoid about a computer's power usage since it's such a meager portion of the total power usage in a household. On the other hand, the overclocking/PC enthusiast community (like the audiophile, ricer, gamer, et cetera communities) is full of idiots that deserve nothing less than to not be allowed to build systems that draw more than 250W of power. Emphasizing low load efficiency indirectly (in its own almost-insignificant way) pushes the agenda of shutting up the script kiddie overclockers. The end justifies the means for me here, so I vote yes: keep doing the low load tests.

HOOfan_1
06-15-2009, 06:00 PM
2 one star reviews on newegg because it smells like paint :rolleyes: and the moderators let that tripe get by?

"it might have lead and be dangerous"...my guess is that author ate plenty of lead paint chips when he was a kid.

sn_85
06-16-2009, 02:18 AM
2 one star reviews on newegg because it smells like paint :rolleyes: and the moderators let that tripe get by?

"it might have lead and be dangerous"...my guess is that author ate plenty of lead paint chips when he was a kid.

lol. that's the smell of factory fresh, not buying used an having it smell like an ash tray or a cat. i hope he wouldn't complain about a "new car" smell after buying one fresh off the lot.

370forlife
06-16-2009, 03:08 PM
lol. that's the smell of factory fresh, not buying used an having it smell like an ash tray or a cat. i hope he wouldn't complain about a "new car" smell after buying one fresh off the lot.


Or buying his comprehensive sandwich and complaining it dosen't smell like socks.

Yanz
06-16-2009, 04:33 PM
i'm enjoying the review. thx OW !

rip97000
06-17-2009, 02:54 AM
great review OW!!!

is there any chance at seeing a picture of you hotbox setup?

burebista
06-17-2009, 03:46 AM
It should be very close to this (http://www.jonnyguru.com/modules.php?name=Testing_Methodology&op=Page2). :)

Minerva
06-18-2009, 10:19 PM
Wow nice box!

rkl525
06-19-2009, 10:07 PM
Hardware Secrets now has the HX850 up for review on their site, saying "
This power supply standby (+5VSB) output, however, is flawed. It reached a ripple level outside the proper range all the time except when we pulled 850 W or more from the unit."

http://www.hardwaresecrets.com/article/741

rip97000
06-19-2009, 10:24 PM
does gabe really take apart the psu first? and if so why? shouldn't he test it they way that he got it from the factory?

Travis
06-19-2009, 11:02 PM
Of course he tests it before desoldering it.

HOOfan_1
06-20-2009, 11:23 AM
same with Oklahoma Wolf...he tests it, desolders it and then puts it back together.

rip97000
06-20-2009, 11:52 AM
it just seems weird cus he starts off the review with that

Smirnoff
06-20-2009, 01:19 PM
Just the way he posts the review I guess. If he took it apart and put it back together before testing, he could run the risk of affecting the test results.

Oklahoma Wolf
06-20-2009, 01:21 PM
You'd be surprised how infrequently the disassembly affects the results. Usually mine test out almost exactly the same once I get them back together. Sometimes they'll scope even better if I find problems inside and fix them first.

Travis
06-20-2009, 01:22 PM
Disassembly phase is the process and test phase is the result. It's logical that process comes before the result, while practically load tests always come before disassembly.

Redbeard
06-22-2009, 11:32 PM
Just in case anybody is interested - the HX750 and HX850 were just 80 Plus Gold certified. I might be wrong but I think we're the only guys actually shipping our gold certified high-wattage PSUs. (well, and CWT, since they are the ODM).

Travis
06-23-2009, 12:18 AM
http://www.80plus.org/manu/psu/psu_detail.aspx?id=25&type=2

Smirnoff
06-23-2009, 05:11 AM
Gold? Holy smokes, that's awesome. :D

Stefan Payne
09-22-2009, 11:30 AM
Hm, could it be that corsair uses a different fan on the produtction units??

as you can see here (http://www.hardware-academy.org/testberichte/7/19-high-end-netzteile-im-vergleich.html) it seems that there is an Ong hua HA1425H12B-Z fan with 0,5A/12VDC in the production HX750...

ef3393
09-22-2009, 08:55 PM
of course the low load test is even more important than active mode, Intel will release the new power management C6 requirement soon the CPU will not consume power during idle mode and the power supply must capable with 12V/0A, It's great if JonnyGuru keep doing low load test 10% and 5% to find out the future proof power unit.
________
Vaporizers For Cheap (http://vaporizer.org/reviews)

kaninja
10-08-2009, 10:36 PM
So I have a question about the HX750/850 now. I have read that Corsair asked the 80Plus Group to degrade the rating of these PSU's to Silver? I also read that Corsair designed the PSU to meet the Silver requirement but when 80 Plus tested the units they actually qualified for the Gold certification. However the units were so close to the borderline of the Silver/Gold requirement border that Corsair decided to market the PSU's as having a Silver rating. My question is, is the above info true?......I wouldn't want to perpetuating FUD if I don't have to. Thanks.

HOOfan_1
10-08-2009, 10:54 PM
Yes it is indeed true.

Confirmed by a Corsair employee on this website...if I could find the thread where he stated it, I would link it

I can link this though
http://channel.hexus.net/content/item.php?item=19218&page=3

kaninja
10-08-2009, 11:01 PM
Thanks for the link. I had read that before but for the life of me I couldn't find the models they were talking about anywhere on that page. Maybe it's time for me to visit the optometrist.