Nexus NX-4000 PSU

Power
Viewing page 4 of 4 pages. Previous 1 2 3 4

TEST RESULTS

Measurements were made at 5 output power levels: 65W, 90W, 150W, 300W and 400W. The PSU was allowed to run for 10~15 minutes at each power level before measurements were recorded. The room temperature was 24C.

A. Load on the PSU

LOAD
65W
90W
150W
300W
400W
+12V
24
36
60
144
168
+5V
20
20
40
80
145
+3.3V
16.5
26.4
39
66
78
-12V
2.4
3.6
3.6
3.6
3.6
-5V
1
2
2
2
2
+5VSR
1
2
2
4
4

B. On test bench, in 24C ambient temperature

AC Power
104W
136
210
410
560
Efficiency
62.5%
66%
71%
73.2%
71.4%
Power Factor
0.96
0.97
0.97
0.98
0.98
Fan Voltage
6.6V
7.1V
8.2
11V
11.9V
Noise*
32 dBA
33dBA
34 dBA
43 dBA
46 dBA

C. In thermal simulation case, over light bulb, with Panaflo 80L case fan on at 7V

AC Power
108W
143
218
416
560
Efficiency
60%
63%
69%
72%
71.4%
Light bulb
60W
60W
100W
100W
100W
Case Temp
27C
27C
29C
30C
30C
Exhaust Temp
30C
32C
35C
43C
46C
Fan Voltage
8.1V
8.6V
9.8V
11.7V
12V
Noise*
34 dBA
34 dBA
37 dBA
46 dBA
46 dBA

D. In thermal simulation case, over light bulb, with case fan turned off

Case Temp
31C
32C
33C
34C
35C
Exhaust Temp
34C
37C
38C
43C
46C
Fan Voltage
8.8V
9.6V
11V
11.8V
12V
Noise*
34 dBA
37 dBA
43 dBA
46 dBA
46 dBA

E. In thermal simulation case, over 100W light bulb, case fan at 12V

Output Power
Case Temp
Exhaust Temp
Noise
400W output
29C
42C
46 dBA

ANALYSIS

1. VOLTAGE REGULATION was excellent, within -/+2% on all lines in any combination of loads. It was often within -/+1%. The low and high voltage seen on each of the main lines is shown:

  • +12V: 11.75 to 12.24
  • +5V: 4.85 to 5.12
  • +3.3V: 3.37 to 3.4

This is far better than the -/+5% load regulation specified by Nexus.

It should be noted that I have no way of testing line regulation, so AC conditions are steady-state, not dynamic as it would be (potentially) in a real PC. The AC input as measured by Kill-a-Watt is usually within a couple of volts of 120V. I have no way to vary input AC voltage at this time. However, with the NX-4000's full-range AC input feature, line regulation may be a moot point.

Although over-capacity output was not tested seriously, the load was increased for a minute to force the PSU to deliver 440W, 10% over its rated maximum power: The NX-4000 withstood this abuse without a hiccup.

2. EFFICIENCYwas moderate in the sub-100W range, but improved to >70% when the PSU was pushed harder, rising to maximum of 73% at 300W. This is good performance.

3. POWER FACTOR was excellent, as it should be with the active power factor correction used in the NX-4000.

4. FAN VOLTAGE: The voltage to the fan started at a low 5.3V. It does not stay that low, however. At the 65W load, without any external heat to the PSU, the fan voltage ramped up steadily over a 10 minute period to 6.6V. In the actual use simulation, the fan voltage at 65W load was 8.1V.

The fan voltage appears to have a more-or-less linear relationship to power output and temperature. In other words, in the NX-4000, any increase in internal temperature always results in an equal proportionate increase in fan voltage up to the maximum 12VDC output to the fan.

This is quite different from the NX-3000, which keeps the fan voltage under 6V up to over 150W output, and then ramps the voltage up steeply thereafter. The behavior of the fan controller in the NX-3000 makes for a fan that varies less in noise during normal PC usage, and also keeps it at a lower speed over a wider range of power loads. The NX-4000 is less advanced in this regard.

5. NOISE measurements were still far from ideal despite the highly accurate professional SLM and 1 meter measuring distance. The test environment is too live and too noisy, and it was evident from the start that higher than normal readings were obtained. (See explanation in Test Methodology section.) Note for example, how the noise measured with the fan at 6.6V was 32 dBA, and with the fan at 8.6V, it was only 2 dBA higher. There is no question that neither is quite right. The difference should be bigger; certainly the first measurement should be lower. It is the effect of the relatively high ambient noise that's causing the high readings. Having said all that, the readings at higher levels are probably not as far off the mark.

Subjectively, the NX-4000 is quiet only at lower power levels. In a direct A-B comparison, it is considerably noisier than its brethren, the NX-3000. The fan used in the NX-4000 is not in the same class as the one in the NX-3000. It has a bit more bearing noise, and because it ramps up to higher speeds much sooner, considerably more turbulence noise as well.

The linear relationship between fan voltage/speed and temperature definitely hurts the noise performance of the NX-4000. The NX-3000, in comparison, has a dramatically different "stepped" curve in the fan voltage vs temperature curve that keeps the fan spinning fairly slowly up to a much higher temperature. This is similar to the behavior of the SF2 fan control circuit in the Seasonic power supplies.

Here is the noise data for the NX-3000 from Nexus. Note that this graph is not identical to the one shown in the original SPCR NX3000 review. The measuring microphone distance here is the correct 1 meter; in the original test, it was conducted at 0.6 meters.

Now here's a noise graph for the NX-4000 of data obtained under the same conditions. Note that the power test points are not identical to those in the NX-3000 graph. The virtually straight line shows clearly the linear relationship between fan voltage/speed and temperature I refer to.

The claim of 21.7 dBA @ 1 meter is probably correct at turn-on, but it is significantly higher under a normal low load (65W). The above graphs shows ~32 dBA at a power level of 160W with an ambient temperature of 27.4C. This jibes reasonably well with the result obtained in my live room testing with the PSU in the open at 24C in Table B above. But the conditions in a normal PC case are much closer to that of Table C above, where the temperature is ~30C, where the noise level at 160W jumps 3 dBA to 37 dBA.

Another thing to consider is that people who buy a 400W PSU likely do so because they need the extra headroom. The performance at 60% power level and higher may be more pertinent to examine. Again, referring to Table C, the noise level measured at 300W output was 46 dBA, which is likley too high a reading because of the live room and reflections. The Nexus spec sheet chart shows 39.5 dBA in 27.4C ambient. This is certainly not very quiet; it is less than 1 dBA less than at full power.

At these 300-400W power levels, air turbulence dominates the sound, which is noisy like most PSUs tested. And perhaps it should be: With current technology, it does not seem realistic to expect low noise in fan-cooled PSUs when running loads more than ~200W. It's interesting to note is that the Nexus factory noise data shows that for 300W output, the NX-3000 and NX-4000 have virtually the same noise. But at all lower power levels, the NX-3000 is significantly quieter.

TEMPERATURES were measured for this PSU during the in-case thermal simulation tests because adequate cooling was an issue for the SilenX 400W PSU reviewed last month. That model is based on the same Fortron as the NX-4000, so it was interesting to explore whether a powerful fan makes for a thermally more stable power supply.

The short answer is yes. The NX-4000 kept itself cool enough and ran continuously without a single problem under all the tough conditions it was subjected to in my usual demanding heat simulation testing.

The in-case thermal simulation tests were done with a lit light bulb in the case. Either a 60W or a 100W bulb was used to simulate the heat generated inside a running PC. Two sets of tests were done: One with no case fan, and another with a Panaflo 80mm back panel fan (our reference fan) running at 7V. In both cases, the PSU behaved well and stayed cool enough to keep the exhaust air well under 50C. It remained, in the worst case, a good 10C cooler than in the SilenX, which shut off due to thermal protection at 57C.

Why the additional fan?

  • The Panaflo 80L case fan at 7V does not add any more noise than emitted by the PSU under normal conditions.
  • Almost every system has at least one case fan, so this is a realistic real-use application.
  • The case fan voltage was selected to provide some airflow while its noise remained masked by the PSU fan noise.

The results show that the extra fan lowered case at all loads, as would be expected. Curiously, PSU exhaust temperatures for the 65, 90 and 150 watt output levels were improved but there was virtually no difference at the higher 300W and 400W levels. It's not clear exactly what this means. Perhaps the internal heat of the PSU was the primary source of temperature gain the the PSU at the higher loads?

A final comment on load testing: Full power testing of PSUs for any length of time is a very demanding test, generally tougher than what real use conditions can demand. SPCR's bench testing is steady-state and can be extended indefinitely until the PSU burns; in real world applications, PSUs in PCs rarely get anywhere close to this kind of abuse (except maybe in a serious server room, which is a different application altogether), and the power demand on them varies up and down in a much more dynamic way, with average power loads rarely exceeding 150W for desktop PCs.

CONCLUSION

Despite being based on the same platform as the ExoticPC's SilenX 400W 14 dBA model, the Nexus NX-4000 is a different beast. It does not take high honors in the noise category, but provides excellent, stable power at quieter than normal noise levels. It is noisier than its smaller brother, the NX-3000, and considerably more powerful. It is wise to provide good quiet case cooling to keep the NX-4000 at the minimal noise level, as its thermal / fan speed curve is more geared to best cooling than lowest noise. For those needing the high 235W of the combined +3.3V and +5.5V lines and the 180W of its 12V line, plus good peak headroom above these levels, it is a good choice.

The NX-4000's strengths:

  • heavy build quality, with massive heatsinks
  • high power output with high peak capacity
  • excellent stability and voltage regulation
  • excellent self-cooling
  • good directed airflow design,
  • good quality active PFC with universal voltage input
  • reduced noise

The NX-4000's weaknesses:

  • linear fan speed control
  • noise on the high side for SPCR
  • fan quality could be improved

The NX-4000 is a solid high power PSU well suited for use in a high power desktop PC. Recommended with caution for PC silencers seeking the lowest noise. Optimized case cooling is recommended in order to keep the PSU fan from spinning up to high noise levels.

Our thanks to Nexus for the NX-4000 review sample and their kind support.

* * * * *

Discuss this review in our Forums.



Previous 1 2 3 4

Power - Article Index
Help support this site, buy from one of our affiliate retailers!
Search: