EVGA 450BT 450W Power Supply

PROD LINK: 450BT Product Page
PRICE: $44.99 @ NewEgg
Price is at the time of testing!

And here we go… powering up the FastAuto for the last time, even though it seems like it got here just yesterday. I’ll be using all the usual gear, of course, much of which will be going to a new home with said FastAuto once the boss makes that determination. For the time being, I’ve trained a crack team of vicious box elder bugs to guard all this precious test equipment until the day comes for it all to make friends with someone else.

Shall we begin? Of course, we shall.

EVGA 450BT – STANDBY Load Tests
Test # +5VSB DC Watts/
AC Watts
Eff. P.F.
Load Tests
1 0.5A 2.55W/
83.2% 0.179
2 1.25A 6.33W/
80.3% 0.356
3 2.5A 12.53W/
75.9% 0.493

Standby tests come first, of course. Here, we find the unit managing above average efficiency all around, though nothing too groundbreaking.

We also see some fantastic voltage regulation to the tune of 1.6%. That’s a good number for any unit.

Let’s run the main tests and see if there are any other good numbers to speak of.

EVGA 450BT – Cold Load Tests
Test # +3.3V +5V +12V DC Watts/
AC Watts
Eff. P.F. Intake/
Progressive Load Tests
1 1A 1A 3A 45.2W/
117.9V 81.2% 0.774 27°C/
3.329V 5.09V 11.84V
2 1.2A 1.2A 6.5A 89.5W/
119.1V 84.8% 0.951 27°C/
3.328V 5.10V 11.82V
3 3A 3A 16A 220.5W/
118.3V 85.9% 0.981 28°C/
3.319V 5.10V 11.81V
4 4.5A 4.5A 24A 330.9W/
117.4V 84.5% 0.992 28°C/
3.312V 5.10V 11.79V
5 6A 6A 33A 451.7W/
116.4V 82.3% 0.997 29°C/
3.305V 5.10V 11.78V
Crossload Tests
CL1 10A 10A 0A 82.3W/
118.7V 76.0% 0.953 28°C/
3.300V 4.80V 12.40V
CL2 0A 0A 35A 395.4W/
117.9V 83.0% 0.995 29°C/
3.492V 5.38V 11.26V

Okay, we have some numbers. Not too many good ones, though.

Fortunately, we did get our Bronze level efficiency pretty easily. No tests need help from me to get there, and even in test one, the number’s not so bad. But… test one is not a clean result. What we have here is a power meter fooler, but only at really low loads. Been a while since I had one of these, and it was only able to fool it until test two. There, the power meter… (sunglasses) had its number. YEAAAAAAAAAAAH!!! Stop groaning… you can be free of my dad jokes and obsolete references tomorrow. Hey, I’m serious… keep groaning, and I’ll whip out a Peter Popoff reference. I’ll do it, so help me God. Come to think of it, God probably would help me with that one.

But the good news cannot last, for we need to talk about those voltage readings. The box promised us 2%, but that was never going to happen the way I test and score things now. This is a cheap group regulated design, and those are often hysterically bad at doing zero loads on any one rail. And in this case, we have worse results than usual. Like the 450BV, this unit is far more comfortable with one crossload test than the other, and that test is once again CL1. You know, where things are skewed to represent ancient 5V heavy rigs.

Seriously, nothing is out of spec on CL1, but everything is out of spec on CL2. All voltages, including the 3.3V. Don’t run this with a modern rig, peoples. As a result of this, the load regulation numbers work out to 5.82% on the 3.3V, 11.60% on the 5V, and 9.50% on the 12V. Mediocrity all the way along, and if you ever see your rig pulling numbers like CL2 has, you will be shopping for new hardware in short order. And this is actually possible now with modern builds that don’t disable advanced sleep states.