• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Some interesting readings from my Seasonic Power Angel

Zap

Zap

Elite Member
First off, Seasonic's page for the Power Angel.

I had a chance to use it and it's pretty cool for a geek like me. I tested it on my gaming rig. Note that the system is a bit different now. Here's the parts:

Opteron 144 running at 2.7GHz
AMD quad heatpipe heatsink
Thermaltake Starforce fan on CPU heatsink
Biostar Tforce6100-939 motherboard
eVGA Geforce 7900GTO video
Seagate 400GB SATA 7200.10 HDD
LG 16X DVDRW
Fortron FSP300-60GLS power supply
A+GPB aluminum mATX slimline case (with mods)

My readings are not exact because sometimes the numbers didn't stay steady. What I noted down were the ones that DID stay steady, or the peak that I saw (when doing something like during booting).

Power Factor - .98
AC frequency - 59.9Hz
AC voltage - 118.7v
Plugged in to wall with nothing hooked up - 1W
Computer plugged in, but shut off - 5W (that LED on the motherboard? 😛 )
While booting into Windows - 156W
idling with CPU fan on low - 109W
idling with CPU fan on medium - 111W (all other readings at this setting)
idling with CPU fan on high - 114W
3Dmock06 Deep Freeze HDR/SM3 test - 195W
3Dmock06 Red Valley CPU1 test - 151W

My "guesstimate" for total possible system draw is ~235W. I get that number from taking the difference between the CPU test and idle (151W - 111W = 40W) and adding it to the video card test (40W + 195W = 235W).

So, in theory if I were to run software that maxes out my 2.7GHz CPU and my stock clocked video card, my system will be drawing 235W from the wall. Efficiency is rated at 75% by Fortron, so 75% of 235W is ~176W.

Now, that's fairly low wattage, but I believe that my system needs it mostly in +12v. Reason why I say that is because with an older Enermax 270W PSU with 16A +12v, I was barely able to run the system with an ATI Radeon X850XT. With a lower power draw 7900GT, I was able to run overclocked a bit on the CPU. With an again more powerful 7900GTO but with this Fortron that is more +12v heavy and does 14.5A + 8A on dual rails with 22A max I can fairly easily run the CPU at 2.7GHz.

My 2 second no-heavy-thinking conclusion: A reasonably powerful single GPU gaming rig does not actually require a huge wattage power supply, just one that is stable and has enough power on the proper voltages. My own configuration draws 235W from the wall and the components only draw 176W from the power supply during load. While idle, system only draws about 83W from the power supply, or 111W from the wall.

Comments? Suggestions? Rip me apart for my sacrilege in saying not everybody absolutely needs a 1kW PSU?
 
I'm in agreement with you Zap. I also have a Seasonic Power Angel and tested my system.

Opteron 165 @ 2.6GHz
TT Big Typhoon
Asus A8N-E
eVGA 7900GTO
2GB RAM
2x250GB Samsung P120 SATA
Pioneer DVR-110
Antec SLK3000B with 2x 120mm fans and a fan controller
Antec NeoHE 430W

Running 3DMark06 I never pulled more than 200W from the wall. So assuming 80% efficiency, that's 160W power draw on load. So I am in complete agreement with you when you say people overrate their needs with power supplies. Just make sure you have a good quality once that puts out stable voltages and you'll most likely be fine.
 
I don't know how accurate the load information is that my APC UPS spits out at me, but here are my observations for two systems that are on an APC RS800 UPS:

First figure is the raw value from the UPS; Second figure in brackets is the first multiplied by 0.78 to take into account PSU efficiency.
idle: 220w [~172w]
3 instances of Euler Prog*: 300w [~234w]
3 instances of Euler Prog* + Doom3**: 315w [~246w]

* Euler approximation program: I wrote in a little 20 line prog in C that estimates a certain differential equation with a ridiculously small delta-x value and compares the answer with what it previously calculated continuously in an infinite loop. System A runs two instances of this program while System B runs only one, so three total.

** Doom3: just me running around killing stuff on the alpha labs sector 4 level on System A.

System A: AMD X2 4200+, Asus a8v deluxe rev2, 4x512MB Corsair VS PC3200, eVGA 6800Ultra AGP, m-audio revolution 7.1 + onboard sound enabled, seagate 300GB 7200.8 SATA, wdc raptor 74GB/8MB, Seasonic S12-500, 3x120mm case fans, arctic cooling freezer64 pro hs/f, Fedora Core 5 x86_64.

System B: AMD Opteron 144, Biostar TForce 6100-939, 2x512MB PC3200, onboard video/text-only, no sound card enabled, 5xPATA + 1xSATA hard drives, Seasonic S12-380, 3x120mm + 1x80mm fans, arctic cooling freezer64 pro hs/f, Slackware-current.

(Everything is stock across both systems)

While I can't say how much power is being consumed on each rail, and the above figures efficiency figures are very loose, regardless, my information seems to back up the OP's conclusion that most systems probably don't need huge power supplies just yet. SLI/Crossfire systems with watercooling and six hard drives is another story, of course.
 
I had another though about this. I'm wondering how fast the Power Angel (and smthmlk's UPS) updates the information. Potentially there may be times when there's a spike in power demand, if for instance EVERYTHING somehow gets in use for a split second at 100%. Would these devices display that? They don't hold a peak value AFAIK, just "realtime" so it depends on me looking at it and reading the LCD on it. So, though my system doesn't seem to draw a huge amount of power, I don't think I'd be comfortable running it on a PSU that can barely put out the needed power (let alone on the proper rails).
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top