Corsair HX1000W Power Supply Review

[Gabriel Torres]

Originally Posted by Olle P

I found a little mistake in the table on page 7.
In the "Input" column +12V2 and +5V should switch place to get the rest of their rows correct.

Fixed, thanks.

Regarding efficiency:
Could the lower efficiency at very high load be the result of you stressing the lower voltages more? What happens if you go heavy on +5V and +3.3V (like ~25A each) while keeping the +12V lines idle?

Cheers
Olle

This is the normal behaviour, please read this tutorial for further information:
http://www.hardwaresecrets.com/article/181

Pulling a lot of power from +5 V and +3.3 V and leaving +12 V idle makes no sense, the explanation is also on the above tutorial.

Gabriel.

[Olle P]

Originally Posted by Gabriel Torres

This [drop in efficiency at high load] is the normal behaviour, please read this tutorial for further information:
http://www.hardwaresecrets.com/article/181
Pulling a lot of power from +5 V and +3.3 V and leaving +12 V idle makes no sense, the explanation is also on the above tutorial.

Unfortunately the tutorial doesn't give any explanation... unless you think it's all a matter of temperature.
The tutorial just state as a fact that efficiency drops at high power output, nothing about why it does so. The only technical explanation, as given a couple of pages earlier, is that most components such as capacitors have their properties degraded at higher temperature. High power output usually means higher temperature and this should thus lead to reduced efficiency.

My way of thinking is that the power loss inside the PSU is greater for current drawn from the lower voltages than from the 12V line. This is based on the fact that all three voltages are drawn from the same source (transformer winding) on the secondary. They all start out as ~15V square wave AC, are rectified into ~14V DC and finally somehow "reduced" into their respective target voltages.
Unless there's another set of pulse width modulators to do most of the voltage reduction from 14V to 3.3V there will be a comparatively huge power loss in the 3.3V voltage stabiliser.
(And the Anatomy of Switching Power Supplies article ends just as it reaches this vital part of the "anatomy", so that's no help either.)

/Olle

[Gabriel Torres]

Page 8, everything below Figure 29.

[Gabriel Torres]

The results from this review were updated today to reflect AC power and efficiency measured by our new precision power meter. The updated results did not affect our conclusions or recommendation about this power supply.