Power used

[dclive]

http://anandtech.com/cpuchipse...howdoc.aspx?i=3066&p=2

Above link shows with a Quad-Core 6600 system and 8800GTX graphics (and single drive, 2 RAM chips, etc.) about 215W are used under load, so a standard 300W PSU would give plenty of headroom.

170W are used normally (when not under load) with the same system.

For those that disagree: Are AnandTech's tests wrong?

 

[SerpentRoyal]

No way an overclocked Quad will pull 215 with one HDD and optical drive. I'd say 300 watts average. Course the final number will depend on Vcore setting.

 

[dclive]

Read the link.

260-277W is the number you're looking for (overclocked, roughly, under load) and the URL is http://anandtech.com/cpuchipse...howdoc.aspx?i=3066&p=4.

 

[MrOblivious]

Originally posted by: PurdueRy
It would be much better to use a 2-CH Oscilloscope to measure instantaneous current and voltage. At least then you know exactly what the transient behavior looks like.

I don't own a Kill-a-watt but I doubt they have that great of a sample time. You could be missing short peaks in the power and have no idea you were. Or the voltage or current waveforms could be very noisey when the PSU is being pushed fairly hard and you wouldn't know.

The kill-a-watt is at best a good tool for estimating idle power draw.

You are exactly right the Kill-A-Watt misses transient peaks of at least 10ms duration...and even at an idle draw they can be woefully wrong with a percentage of power supplies on the market:


http://img.photobucket.com/alb...e_c_t_r_e/100_1218.jpg

http://img.photobucket.com/alb...e_c_t_r_e/100_1976.jpg


The one thing I have noticed is it is consistant when it works. So for a unit it reads it is reasonably accurate....but for units it does not work with you get the above.