This is Sick! (FSP psu)

S

scwam

Gawd
Joined
Jul 26, 2005
Messages
829
http://www.newegg.com/Product/Product.aspx?Item=N82E16817104952

Nominal loads of 440 to 475 watts for nearly an hour! I can believe my system even ran on this thing. I got minimal shutdowns only on 3dmark 06. I'm stunned that stalker burning up 460 to 485 watts just standing there while I went to take a shower was still running fine and kept on playing afterwards. Just look at all the crap I've got hooked up to it in my sig.
 
probubly because your pc doesn't actually draw that much power

My pc, opteron 146 @ 3ghz, 5 hard drives, 7800gt, etc, draws.. 119w full load. Prime95, while copying a 5gb file between drives, and 2 3dmark06's.

Most people over-estimate their power needs. If you have a Kill-a-watt, you might know the truth behind how much power people actually need, and how much they think they need.
 
Rabid Badger said:
Where are you getting those power draw figures?
Click to expand...

Who knows. The build in his sig would require 200W max from a PSU even under load. My 4300+ w/ 1GB of RAM and 7950GT OC needs only 160W. Of course, that build has SS Flash drives instead of HDD's, but still.
 
Bbq said:
probubly because your pc doesn't actually draw that much power

My pc, opteron 146 @ 3ghz, 5 hard drives, 7800gt, etc, draws.. 119w full load. Prime95, while copying a 5gb file between drives, and 2 3dmark06's.

Most people over-estimate their power needs. If you have a Kill-a-watt, you might know the truth behind how much power people actually need, and how much they think they need.
Click to expand...

I'm using this for the measurements http://www.seasonicusa.com/power_angel.htm
I've been using it on my Enermax Liberty500 for over a year. I typically get the same loads on the enermax as well. 360 watts dual burn at 3ghz, watts go higher as temps and resistance increases. 3dmark can spike to 530 watts depeding on my overclock and the temperature of the room too, averages around 420-475 usually on 3dmark06. Tested it on a few things like lamps, hairdryers, microwave, etc. Seems pretty accurate. 100 watt light bulb reports ~104 watts.

Minimal shutdown, I meant to say a minimum amount of shutdowns, ~15% of the runs of 3dmark06. Gaming nfsmw and stalker was no issue though.
 
That's part of the discrepancy. That product measures the power draw from the wall, not what your power supply is actually supplying to your computer.

You have to take the efficiency of the power supply into account. FSP claims that power supply is 70% efficient. Power supplies at that price point rarely meet their efficiency claims, I'd put it at around 65% efficiency realistically.

485 x 0.65 = 312W being supplied to the computer.

That's a much more likely number for your system's actual power usage.
 
I'm not really sure why this meter seems to show the same pull from the wall as my more efficient Enermax liberty 500 (I think it's rated 80 or 85%). I expected the readings to be actually higher than the liberty since the liberty has higher efficiency (more power draw since a lower efficiency has to be mulitplied by that draw). But yet they seem to show the exact same readings on the same system. I did notice that the Enermax PF shows .98 to .99 while the FSP always showed .64 to a maximum of .68 PF.

I noticed when I took the voltage * the amps to equal the reported V/A, mulitply that product by the average 0.66 PF it came out to the watts it was reporting. Then took that product (I guess VA?) and multiplied it by .66 it came out equal to the meter reporting.

IE: 119v * 5.8a = 690 (this product reported on the meter too)
690 * 0.66 = 455 (just wat was reported on the meter)
.
Why would the reported loads (in watts) be similar given the average .66 (FSP) and typical .99 (enermax)? I began to wonder if the meter had some way of knowing the efficiency and recalculating the final watts showed. But this idea seems crazy
 
Those FSP PSU's are pretty solid. I have never had an issue with one. But I dont' put them into much more then mid range parts. They are solid.
 
FSP/Sparkle are typically good, solid, reliable units... and FSP actually tends to under-rate their product.

I just wish they had done a better job with their Epsilon design. I wonder if they have revised it to smooth out that wicked ripple at high load yet.....
 
You must log in or register to reply here.
Back
Top