Corsair HX650w 650W ATX PSU

Simulated Load Testing

Corsair HX650w

To provide accurate and consistent results in all of our PSU testing, Overclock3D uses professional grade DC electronic load equipment capable of placing a sustained load of 3690w across a total of six rails (including +5vsb and -12v) on the PSU!

This is achieved by using a combination of SunMoon and Analogic electronic load equipment which allow us to adjust amperage loads in increments as small as 0.01A while also measuring voltage and wattage readings on-screen.

During today's tests, we will be placing the Corsair HX650w under 25%, 50%, 75% and 100% load levels inside a hot box regulated to a temperature of around 50°C. Additional 'Cross Load' and a 'Max Load' tests will also be performed under these conditions to simulate how the PSU reacts to heavily uneven loads as well as running above its specified output.   

Corsair HX650w Results @ 50°C
   +3.3v  +5.0v  +12v  +5vSB  -12v  AC Watts /
 DC Watts
 Efficiency  Intake /
Δ Temp
Test 1
5.25A 5.25A 9.50A  0.75A  0.20A 190w /
 86.84% 50.4°C /
3.37v 5.08v 12.10v 5.01v  -11.94v
Test 2
10.50A 10.50A 19.00A 1.50A  0.40A  377w /
 87.26%  50.3°C /
3.34v 5.04v 12.06v 4.97v  -11.95v
Test 3
15.75A 15.75A 28.50A 2.25A  0.60A  567w /
 86.59% 50.2C /
3.30v 5.00v 12.02v 4.93v -12.00v
Test 4
21.00A 21.00A 38.00A 3.00A 0.80A 771w /
84.43%  50.6°C /
3.26v 4.95v 11.96v 4.89v -12.04v
Test 5
21.00A 21.00A 1.00A 0.00A 0.00A 240w /
75.00% 50.8°C /
4.0 °C
3.27v 4.77v 12.12v 5.01v -12.13v
Test 6
 1.00A  1.00A  57.00A  0.00A  0.00A 716w /
87.98%  51.0°C /
3.38v 5.07v 11.96v 5.00v -12.68v
Test 7
 21.00A 21.00A 44.00A  3.00A  0.80A 858w /
84.14%  51.2°C /
3.25v 4.94v 11.96v 4.89v -12.10v
Corsair HX650w Performance Overview
 +3.3v Diff.
 +5.0v Diff.
 +12v Diff.
Avg Effic.
Noise Rating
-3.26%   -2.55% -1.15%  86.28% Low 

Looking first at the performance overview chart directly above, we can see that the HX650w manages an average efficiency of 86% in the standard tests 1-4. This is pretty much in line with Corsair's advertised efficiency of 85% given a small margin of error on our part and Corsair's tendency to under-hype their PSU specs. Of course, this is also the first place where we can truly see a difference between the HX650w and the HX620w, with the latter 'only' achieving ~80% efficiency, which to be fair, was good for its day. Voltage stability is also very reasonable on all of the main rails, with the highest fluctuation of 3.26% being well within acceptable tolerances.

Moving on to the first cross-load scenario in test 5 shows that the HX650w isn't too keen on running without a load on its +12v rail. However, as I've said before, this particular test isn't something likely to be reproduced in a real world scenario (who would buy a 650w PSU to only run it at 180w load on the +3v/+5v rails?). So really the results here are somewhat forgivable. Cross-load test 6 on the other hand needs no forgiveness at all, with the HX650w delivering near on it's full output on the +12v rail alone, and with efficiency results approaching almost 88%!

The final test, is of course the one that every PSU manufacturer fears the most....MAX Load. In this test the PSU is pushed as far as it can go before it either engages its safety features, or spits more flames out of its exhaust grill than the Batmobile. Recently I've had quite a few of the latter, but thankfully the HX650w didn't follow suit and managed to deliver a respectable and sustainable 722w. Attempting to go any higher simply resulted in the PSU shutting down safely. Phew.

Corsair HX650w Scope Results @ 50c
   +3.3v  +5.0v  +12v
Test 1
Test 2
Test 3
Test 4
Test 5
Test 6
Test 7 (MAX)

Ripple refers to the small fluctuations in voltage that occur in all PSU's on a millisecond scale. Large quantities of ripple on a PSU can damage or kill components in your PC over time and therefore testing a PSU's performance in this area is just as important as any other part. Unfortunately ripple cannot be measured using a standard multimeter and requires a specialist device called an oscilloscope. The results you see above are taken from our Rigol 25Mhz 400MSa/s scope that can save 'screenshots' of its data to a USB memory stick.

ATX specification states that ripple should be no higher than 50mV on the +3.3v/+5v rails and 120mV on the +12v rails, and judging by those flat lines in the screenshots above, the HX650W certainly had no trouble at all meeting these requirements throughout almost all of the tests. Only in the cross-load test 5 does the ripple on the +5v rail hit 70mV, but once again we need to remember that this particular load scenario is highly unlikely to ever be encountered in the real world. The rest of the results on the other hand are exceptional with the +3.3v and +5v rails rarely budging from around 10mV and the +12v rail only hitting 42mV in test 6 when the PSU is already running 70w higher than its rated output. Absolutely awesome work Corsair!

«Prev 1 2 3 4 5 Next»

Most Recent Comments

15-01-2010, 03:05:32

The banners are out, the drinks are flowing... Whats the occasion? Jimbo is finally back in the reviewers chair and is taking a look at the Corsair HX650W power supply.

Continue ReadingQuote

15-01-2010, 03:57:33

Great review Jim, your test area must look like a NASA lab room :P Great to see you back mind ;-)Quote

15-01-2010, 04:42:40

Cheers mate. And yea, when people come round they do give me a bit of a strange "are you a terrorist or something" look Quote

15-01-2010, 05:01:22

Is the 7 year warranty applicable to the modular (HX) series only, or does it cover the TX PSU's as well? local etailers have variously listed 5 and 7 year warranty's. Not a major as I'm sure either would last the distance. Awesome review.Quote

15-01-2010, 05:10:31

The TX950w I've got here has a 5yr warrenty. So I'd imagine that applies to the rest of the TX series too.Quote

15-01-2010, 07:32:32

good review, makes me glad i bought 1 !Quote

15-01-2010, 12:28:07

nice review jim, im upgrading my psu pretty soon and these corsairs are defintely the way 2 go but i was going to get the HX750 one since its more future proofQuote

15-01-2010, 16:35:54

Originally Posted by name='thestepster'
nice review jim, im upgrading my psu pretty soon and these corsairs are defintely the way 2 go but i was going to get the HX750 one since its more future proof
The corsair HX series PSU's are quality products. You really can't go wrong with any of them.Quote

15-01-2010, 21:01:14

exactly its just the pennies that cant stretch, i was thinkin about 850+ but being a jobless tax scroungers just now its my best optionQuote

01-09-2010, 18:17:29

All - Any thoughts on these items relative to replacing my stock PSU with one of these animals? Dell's notorious for arranging its grilles and mount points so as to only "allow" Dell replacement parts. I have an XPS Studio 435T/9000 (Core i7 920) and am worried about 3 things.

1) Form and fit (not afraid of a little metal work, however)

2) cable management (more along the lines of reach and airflow than aesthetics), and

3) the PSU's ability to handle Stepped Approximated Sine wave or Pulse Width Modulated sine wave UPS's (battery side only, not AC side).

These Dells are notorious for crashing when presented fake sine waves from their power sources. Any idea if the Corsairs care if it's a pure or fake sine wave? Ever test that? If interested pick up an APC BX1300G and see if your PSUs puke when it goes into battery mode... I'd be real interested in the result!

Great report, Jim.

Cheers, BZ

02-09-2010, 03:34:17

Hey BZ,

To be honest I have little (read:no) experience with aftermarket PSU's in Dell machines. The last I heard tho was that they switched around certain cables on the ATX connector so that only their PSU's would work and anything else would go *bang*. Not sure if this is still the case though?Quote

02-09-2010, 05:20:51


All the recent Dell machines I have replaced PSUs in have had standard ATX (albeit crappy) units in them. I would think yours is no different, the only issue I can forsee is if there is enough room in the case for a new one, since Dell OEM PSUs tend to be on the short/small side and the Corsair units are a little bigger. Remember you'll need some cable clearance too. The only newer Dells that I've worked on with non-standard PSUs were smaller form factor machines, which is almost to be expected, though even some of the larger towers (probably not XPS) have been more narrow than normal.

Never seen a Dell with the ATX connectors switched round wrongly yet, I would hope that they'd also change the plug so that a standard one simply wouldn't fit in that case, who knows if they do or not.

I can see many complaints on the net about your PSU having problems like you mentioned, but no issues with Corsair units in the same way, so I would think a decent quality unit would solve the problem or not present the issues you are concerned about.

Remember, if you buy a unit from the internet, you can use the Distance Selling Regulations to get it returned if you find out it won't fit, but you do only have 7 days to do that and obviously you mustn't damage it.

Hope that helps.

I have a Corsair HX620 ready for RMA (still) and by the sounds of how they operate I may get back one of these newer units, glad to hear it's at least an improvement on the old one.Quote

02-09-2010, 07:46:05


Thanks for the feedback and sharing your observations/experience. I've been inside the XPS tower - it seemed pretty roomy, bet then again I wasn't trying to replace the biggest blob inside the box at the time. Based on your and Jim's reply I guess the best thing to do is get a multimeter out and check everything before trying to slip it in place. Agree - not changing the plug while messing with polarities and voltages would be criminal. I'm building my next box...

Thanks again both. I'll report back after the Corsair arrives (tomorrow as scheduled).


BZ (from across the pond - Northern Virginia, USA)Quote

02-09-2010, 07:54:43

I think I must be showing my age...

It seems that starting after September of 1998 Dell defected from the cause of industry standardization and began using specially modified Intel supplied ATX motherboards with custom wired power connectors. Of course they also had custom power supplies made that duplicated the non-standard pinout of the motherboard power connectors.

An even bigger crime than simply using non-standard power connectors is that only the pinout is non-standard, the connectors look like and are keyed the same as is dictated by true ATX. There is nothing to prevent you from plugging the Dell non-standard power supply into a new industry standard ATX motherboard you installed in your Dell case as an upgrade, or even plugging a new upgraded industry standard ATX power supply into your existing Dell motherboard. But mixing either a new ATX board with the Dell supply or a new ATX supply with the existing Dell board is a recipe for silicon toast. How do you like your fried chips, medium or well done?


03-09-2010, 10:11:43

Originally Posted by Jim View Post

I think I must be showing my age...


So - I ordered an HX750W from the net and after doing so thought it prudent to call Corsair to discuss this before installing. Explained my situation to the quite knowledgable tech; mainly the story on the factory PSU, speed bumps people had run into, and my problem with the OEM power supply not liking the "dirty" AC signal coming from my UPS's inverter. He said he'd never seen that problem with a Corsair, so that was the first bit of good news. I was torn between sticking with the old PSU and upgrading to a pure sine wave UPS (less risk) - or - buying a new PSU and sticking with my relatively inexpensive UPS (much less cost despite making a major upgrade to the PC itself). Whilst speaking with him he urged me to verify (with Dell) that my motherboard was ATX12V 2.2 (& backward) compliant, had a 20+4 pins main power plug and either an 8 pins or 4 pins power plug for the CPU. So off I went first to Intel to read the ATX 2.x spec. Found pinouts and other information useful to non-Dell users - since Dell only publishes scant marketing level descriptions (hardly full specs). So then off I went to (useless) Dell. The only thing they could tell me was their part number (F217J) for my original PSU - couldn't recommend any others as replacements. I cross-referenced that to a Flextronics Model #VP-0900073-000. Couldn't verify that it was ATX 2.2 compliant, but found a few posters out there who claimed to have successfully replaced this exact model with a number of differing brands, all ATX 2.2 compliant PSU's. The new Corsair arrived yesterday and down went the PC to the workshop. I cracked the machine open, cleaned and dusted, took a bunch of pictures and measured stuff. Decided it would fit and it was safe to open the new Corsair.

Wow - was I thrilled when I opened up that box! High quality nylon pouch for the modular cables, packaging like I've never seen before (velvet-like bag containing the PSU), wire ties, extra mounting screws and a very well written installation guide - even a metal "Powered by Corsair" logo tag for the outside of the machine. I was really impressed. So, the new PSU was about 1/2" shorter and about 7/8" longer, but turns out those dimensions don't matter much in my chassis. True to form, Dell's external mounting holes in the chassis didn't line up with the PSU. I extended the upper right chassis hole by making it a bit oblong (up and to the right about 1/8") and drilled a new lower left hole (about 5/16" directly below the original one). New PSU slipped right in and the holes lined up with their threaded mounting screw bosses in the PSU case. Because the Corsair isn't as tall, it didn't rest on a stamped-out horizonal chassis tang that the original rested upon, but there was no give or movement of any kind once the four screws were securely in place. The cables are some of the best quality I've ever seen. Took about 15 minutes to route stuff, another 10 minutes to tidy and close things up, and back up to the office I went.

Fired right up (I like my chips medium, thank you Jim). UPS monitoring S/W said I was burning 195W - about 50W less than before (I suppose that's a manifestation of the unit's efficiency... cool!). Then off to test the el cheapo APC Back-Up system (el cheapo - US$169 - not that cheap). Ran the self test and voila. It worked. This PSU's so quiet - what a purchase.

Jim and alexhull24 - I owe you a debt of gratitude. Thanks for the prudent warnings and good advice - and Jim - thanks for the exceptional review the 650. While I didn't necessarily write this for just you gents, I thought you might appreciate the detail and the news on the XPS Studio 435T's ATX 2.2 compliancy. Maybe somewhere along the line Dell got the message... Then again, probably just an "oversight" on their part.

One footnote. Someone out there blamed Active PFC as the root cause of my problem between the PSU and the UPS. Not true. The old PSU is a switching PSU. Says so right on the label. Glad I didn't rule out Active PFC. Had I followed that "advice" I probably would not have bought this and may have ended up with the same problem... There you go, Jim. See if you can show that Switching PSU's can't handle the Step Approximated or Pulse Width Modulated sine waves. Maybe that's the root cause... Maybe it's just a piece of junk Dell OEM PSU that's the problem.



Attachment 2844Quote

Register for the OC3D Newsletter

Subscribing to the OC3D newsletter will keep you up-to-date on the latest technology reviews, competitions and goings-on at Overclock3D. We won't share your email address with ANYONE, and we will only email you with updates on site news, reviews, and competitions and you can unsubscribe easily at any time.

Simply enter your name and email address into the box below and be sure to click on the links in the confirmation emails that will arrive in your e-mail shortly after to complete the registration.

If you run into any problems, just drop us a message on the forums.