Overshoot Transient Testing - EVGA NEX1500 #1
VSB On	VSB to Full, 12V	Off to Full, 12V

Let's get some power on transient testing done, now. Not the cleanest looking 5VSB shot I've ever seen, up there, but there are no problems that I can see. There's a little spike, and it's well suppressed. The 12V shots, however, are outstanding and very clean. This is exactly what I want to see from a unit like this.

Now... the hot box. Since this unit is rated to fifty degrees at full power, that's what I'm going to shoot for. As before, I will add the fan RPM column to the table, and I will add one further column to show you the software indicated temperature.

Let's look at the temperatures, first. It immediately becomes clear to me that the thermistor we get our software reading from is in the fan's air path. I never saw anything hotter than the incoming air temperature from it, making it useful to tell how hot your case is, but it tells you nothing about how hot the power supply is. And even then, it could be more accurate. Look at test five. I got 47 degrees from my K-type thermocouple reading, while the software thought it was 41 degrees. All this with an exhaust temperature at a scorching 81 degrees.

This thing got HOT, folks. And because it got hot, the fan quickly tried to launch this unit into low earth orbit. I doubt the little thing could spin any faster (or louder) than it did.

Another effect of the heat was the decline in efficiency. No longer does this unit pass Gold. Then again, it was barely hanging on in cold testing anyway, so this is to be expected. 80 Plus load tests at room temp, so I'll give this a pass. But I'd really like to see better full power efficiency from this unit. We got Platinum numbers again in tests one and three, it's just at full power this unit has some trouble with the efficiency and heat.

Time for another look at the progressive test stability calculations. For the 3.3V rail, I see 3.6% now. A little bit of a slide. On the 5V rail, I get the same 2.9% I got in the cold tests. Finally, the 12V gives us 2% on the dot. Again, very good but not excellent. The average number, on which I will be scoring, is 2.8%.

Before we move on, here's another reminder of how I set up the scope for load testing. Both channels are always active, one showing the top SunMoon in yellow, the other showing the bottom one in blue. You can't see the yellow trace here because it's obscured by the channel B trace. This shows the EVGA unit during testing, but all shots will be from the software, rather than from the scope itself.

I'm going to zoom the scope in a bit and then let's see what we've got.

Oscilloscope Measurements - EVGA NEX1500 #1
Test #	+3.3V	+5V	+12V
LLT
Test 1
Test 2
Test 3
Test 4
Test 5
Test CL1
Test CL2

Now this is what I live for. All rails are under 20mV at all times. This is what I expect to see from class leading power supplies! It doesn't get much better than this, people, especially at this power level, and...

Sigh. There's often a snag, isn't there? This 140mV negative spike turned up during hot testing. At random intervals, the unit would emit a muted tick and then throw this onto the scope screen. It was very hard to capture, requiring much dexterity with the scope and a lot of patience. It did not do this often, but once it started, it didn't stop. And because of its infrequency, I can't quite think of it as a ripple anomaly, exactly. No, it's more of a transient anomaly. The unit acts like I've just called for a switch in load patterns, only I haven't. And the spike varies in intensity - it is under 100mV at low loads, but now present at all loads.

So, I figured the unit must be faulty. I hadn't seen this reported on any other review units. I asked EVGA if they'd care to send a second sample, and this was the main reason it got sent.

I unpacked the second unit, hooked it up, powered it on, and... dead. Click, click, click, click, click, click, click, click, click, click. Over and over, primary side shutdown after shutdown. Would not even hold 5VSB. I opened it up, found some damage inside, fixed it, and then tested it. During test three...

Lovely. That's two, and I'm still not able to determine whether its a defect in the design or a result of some kind of damage. All I know is my first one didn't do this until it got way hot, and my second one had damage right out of the box. So, I'm not going to score too much against this. I can't, because accepting a third sample would not be fair to the others that only got two chances here at the site. And I also can't go too hard on it because they can't reproduce this behavior at EVGA HQ.

All I can say for sure is that it's not my test equipment. While I was waiting for this second unit to arrive, I went start to finish on that 1250W Seasonic, which showed absolutely nothing like this. No, it's definitely something common to both units and I am not going to be able to tell you why. ARRRGH!

Let's take them apart for pictures now. I for sure want to show you what the deal was with the second sample, so you know why I can't just call this a design defect and score a giant performance hit against it.