Corsair HX750i 750W Power Supply

REVIEW INFORMATION
SUPPLIED BY: Corsair
MANUFACTURER: Corsair
PRODUCT: HX750i 750W
PROD LINK: HX750i Product Page
PRICE: $169.99 @ NewEgg
Price is at the time of testing!

Now, with this being a power supply review and all, I’m fixing to do some load testing. One SM-268 is enough to load down this unit today, so that’s what I’ll use. The rest of the test gear will be the same as I’ve used about a billion times before in most of my other reviews.

HX750i 750W – StandBy Tests
Test # +5VSB DC Watts/
AC Watts
Eff. P.F.
Load Tests
1 0.5A 2.52W/
3.19W
79.0% 0.268
5.03V
2 1.5A 7.50W/
9.46W
79.3% 0.477
5.00V
3 3.0A 14.9W/
19.2W
77.6% 0.574
4.97V

First up, standby load testing. Here we see this unit doing well above average on all three tests, but not being really groundbreaking by getting above the 80% mark. Such units are few and far between right now, though, and it will probably stay that way for quite some time yet.

Voltage regulation is a respectable 1.2%. Excellent results.

HX750i 750W – Low Load Tests
Load
Level
+3.3V +5V +12V -12V 5VSB DC Watts/
AC Watts
P.F. Eff.
Load Tests
10.1% 1.5A 1.5A 4.8A 0.2A 0.5A 75.4W/
86.7W
0.974 87.0%
3.28V 5.03V 12.08V -12.12V 5.02V

Before I babble about the numbers in this test, please observe that the load level is a clickable link. I did have Corsair Link up and running, and yes it did work immediately this time out. The link shows you the screenshot taken during this test. Turns out that on these HXi units, Corsair link does only allow minimal control over the fan and whether or not multi-rail overcurrent protection is enabled. No enabling the OCP or setting limits for each 12V cable – you need an AXi for that much control.

At any rate, we can now compare the software directly against my measurement gear. Now, remember… no built-in monitoring solution is going to match up well to a $300 power meter. Better accuracy is not possible without paying for it. In this test, power going out is reasonably on target while the incoming power draw is considerably off. This led to the software actually reporting lower efficiency than what we got. Additionally, while the 12V and 3.3V current draws were represented somewhat accurately, the 5V draw was grossly overestimated.

But that’s just for the low load test. Perhaps the software accuracy improves when you go up in power a bit. Let’s find out.

HX750i 750W – Cold Load Tests
Test # +3.3V +5V +12V DC Watts/
AC Watts
AC
Input
Eff. P.F. Fan
RPM
Intake/
Exhaust
Progressive Load Tests
1 1.5A 1.5A 11A 150.3W/
165.7W
117.6V 90.7% 0.986 0 28°C/
28°C
3.28V 5.02V 12.08V
2 3A 3A 22A 297.1W/
321.5W
116.9V 92.4% 0.994 0 28°C/
30°C
3.26V 5.01V 12.04V
3 3.5A 3.5A 28A 375.5W/
405.8W
116.4V 92.5% 0.995 0 29°C/
32°C
3.26V 5.03V 12.02V
4 5.5A 5.5A 45A 596.2W/
651.7W
115.4V 91.5% 0.997 712 27°C/
43°C
3.26V 5.03V 11.96V
5 7A 7A 56A 743.1W/
820.7W
115.7V 90.5% 0.997 896 28°C/
44°C
3.25V 5.00V 11.93V
Crossload Tests
CL1 18A 18A 0A 153.8W/
183.9W
118.6V 83.6% 0.989 696 28°C/
39°C
3.27V 5.00V 12.07V
CL2 0A 0A 62.5A 749.9W/
826.7W
116.7V 90.7% 0.997 920 28°C/
42°C
3.23V 4.98V 11.92V

Once again, each test number is clickable to compare against the software. And power efficiency calculating has indeed improved a fair bit. It’s still not accurate, but it does arrive at mostly an accurate calculation of the actual efficiency number. Which, by the way, was an easy pass for Platinum. No complaints there.

Unfortunately, the accuracy of the 5V current draw measurement continued to be a problem. In test one, it wasn’t so bad, but by test five it was off by three whole amperes. It’s too bad because otherwise, the software was pretty darn good at keeping tabs on the current draw of the other two main rails. 12V current was dead on target the entire time, while the 3.3V rail only really messed up in the first crossload test. Where, strangely, the 5V current display decided to behave itself for only that one test. Weird.

But enough about the software for now. We need to whip out the calculator and see how stable the real measured voltages were at the ends of the cables, where the software can’t measure too well. The 3.3V rail managed 0.9% stability – excellent. The 5V rail did 0.6% – even better. The 12V rail came in at 1.2%. Also excellent. The average was 0.9%. So, you’re not going to get the stability of an AXi unit out of this thing, but there’s no shame in these numbers at all. This used to be, until relatively recently, the best you could expect from anyone.

Let’s take another look at the software before moving on.

Nothing’s changed here… the customizable main screen still shows us a tiny case lacking the space for all the stuff an AX1500i supports. I have that unit powering a mining rig right now… that rig has so many video card dials and readouts along that left side pane I didn’t even try to customize the display.

Here’s a look at the low load test power screenshot with a configuration tab open. The ability to set alarms for voltages is displayed. I confess I never really bother with that side of the software.

Of course, you can fiddle with the fan if you want to. And when you’re done mangling some poor song by using a fan to fiddle with, you can even control how fast it spins once you have it back inside the power supply.

CORSAIR LINK IS STILL RUNNING!!!!!!!!!!!!!!!!!!!!!!!!!!!

Just thought you’d like to know. Corsair Link certainly thinks I have to be reminded every bleeping time I minimize the software via tooltip. Annoying? Yes. Unwelcome? Yes. Would I like the ability to shut off that reminder? Oh, hell yes. But I’m too sexy to worry about it.