Power used

[dclive]

I'm posting here because I see this as a related topic - how much power is actually used on a given system? I bought a "Kill-A-Watt" gizmo that measures wattage used at the wall socket, so it's seen as a pretty good way to figure out how much power is actually required.

I have a HP 9207US, which is their Core Duo 1.73 laptop with 17" LCD, nVidia 7600GT system, and 1.5GB of RAM, 120GB HDD, etc...etc...

It uses about 30W from the wall (so it's actually at say 80% efficiency, so the *computer* and computer's power supply actually supplies perhaps 25W or so of power while surfing and such via remote desktop (ie using it for essentially display only, no heavy CPU usage).

When put under a load, it's using 48W from the wall (so at 80% efficiency, the computer and computer's PSU actually supplies about 38W.

25W - 38W (at the PSU) for all of the stuff above. Granted, laptop versions, but still, that's not much power required...

Something to think about when sizing up PSU requirements....

I'll have more data the next time I reboot my iMac and my Acer E700 (Q6600 + nvidia 8800GTS/320 and 4 drives box).

 

[SerpentRoyal]

Nominal load for a desktop is 140 to 150 watts.

 

[dclive]

Another data point:

iMac 2.0 Ghz Core Duo with 20" LCD, 250GB drive, Radeon 1600, 1.5GB RAM, DVDRW, bluetooth, wireless - etc...the normal Mac stuff -- is 100W at the wall; assuming 80% efficiency, it actually requires about 80W to work, this while doing some archiving operations (archiving large files) to get the CPU going. Figure about 15W less when not doing this.

 

[dclive]

I'll see if I can find something to hit both cores; I've got to believe it's a little higher than that.

 

[dclive]

Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

 

[myocardia]

Please provide links to an electrical engineer saying what you're saying, and we might believe it. Until then, we're going to do just like you do, and ignore anything that doesn't completely agree with what we already believed.

 

[John]

Originally posted by: dclive
Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

Try running 2 instances of Orthos and rthdribl at the same time. Now tell us what the kill-a-watt reads.

 

[dclive]

Originally posted by: myocardia
Please provide links to an electrical engineer saying what you're saying, and we might believe it. Until then, we're going to do just like you do, and ignore anything that doesn't completely agree with what we already believed.

That it's fine to run a PSU within its' published specs?

 

[dclive]

Originally posted by: John

Originally posted by: dclive
Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

Try running 2 instances of Orthos and rthdribl at the same time. Now tell us what the kill-a-watt reads.

OK, give me a day or two to reboot it and such, and I will.

 

[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.

 

[dclive]

Agreed - I don't suggest it's perfect - but it does seem to correspond roughly with some of Anandtech's numbers in the PSU reviews I see here with C2D systems....

 

[BoomerD]

Originally posted by: John

Originally posted by: dclive
Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

Try running 2 instances of Orthos and rthdribl at the same time. Now tell us what the kill-a-watt reads.

I tried that John. WOW! It slowed my rig to a crawl, but it handled everything with no problem. The Kill-a Watt showed about 250 watts max load during the 20 minutes I ran the test. Even more surprising to me, is that Motherboard Monitor showed the 12+ voltage at 12.15...Normally, the MBM shows the 12+ at 11.4 to 11.6v.
I still need to dig out my digital multi-tester and measure voltage that way.

 

[rise]

whats in your rg Boomer?

i pulled 350w from the wall with dual prime and 3d06.

e6400@3.00
680i mobo
8800 gts 640
4x 1024
3x 320gb
theater 650
dvd/rw
4x 120mm, 1x 80mm
swiftech pump

however, that was from a ups to a kill-a-watt to the wall so John reminded me that uses 15-25w

maybe i'll re-run directly from the wall with ortho/dribl while recording a show and a ripping a dvd.

 

[BoomerD]

It's not the most high-tech rig, but it handles the things I need it to...for now.

Pentium 4 Northwood 3.06/533 (BX80532PE3066DSL6PG) (stock clocks)
ZALMAN CNPS7700-CU 120mm CPU cooler
ASUS P4P800-E Deluxe
Visiontek Xtasy Radeon X850XT Platinum Edition AGP (stock clocks)
SoundBlaster LIVE! With ?breakout-box? front panel
3X512mb DDR400 RAM (soon to be 2X1 gb, 2X512mb)
Seagate 250 gb SATA HDD
SAMSUNG 18X DVD±R DVD Burner
Toshiba 6X DVD-ROM
Iomega Zip 250 Drive
Floppy drive (unknown)
6-80mm case fans (yep, sounds like a 747 sitting next to me...(planning on adding quieter fans with this upgrade as well)


I'm not a big overclocker, preferring instead to buy cpu/vid cards that will do what I want at the stock clocks instead ot taking the chance of FUBAR'ing things...

Hopefully, sometime next year, I'll make the big jump to a Conroe processor and of course, new motherboard, ram and PCI-E video card.

Yesterday, I tested the system load by running 3DMark05 while downloading a large file and burning a CD and only hit 285 watts briefly during the process.
The Kill-a-Watt is plugged into a power strip, rather directly into the wall, but that shouldn't make any difference in the readings...should it?

 

[rise]

i miss my northwood 2.4c thats still a fine rig but do you still use that zip drive?

i don't think the power strip would add anything but i'm sure someone will correct me

 

[BoomerD]

Originally posted by: rise
i miss my northwood 2.4c thats still a fine rig but do you still use that zip drive?

heh-heh...almost never since I started burning CDs and DVDs...I have some misc. drivers and patches stored on the zip disks, but that's about it. I think I use the zip maybe 2x per year...at the most.

 

[BoomerD]

Well, my system certainly didn't like that...I ran Orthos in 2 instances, plus rthdribl while hooked up to a multi-meter. Power (as measured through a molex connector) for the most part, it ran in the 11.4-11.45 range...until I fired up rthdribl...then it dropped to 11.30 volts...ran for a few seconds like that, then the video card went "wonky" and started putting out nothing but pixelated garbage...
Mind you, nothing in my system is O/C'd...I'd hate to see what O/C'ing the cpu and video card would do to things...
Just more proof this Ass-poor psu needs to be replaced.

 

[rise]

lol, whats your psu or did i miss that?

 

[BoomerD]

It's not a good one...Aspire 500 watt...came with my case. In fact, I've known for some time that they weren't rated very highly, (do I get an award for the understatement of the year?) and have suspected that it may be the cause of some of the little "glitches" I experience from time to time. Nothing serious, and could be software caused as well, but MBM and other "software monitoring" have shown low 12v ranges, so that started my doing some investigating a couple of weeks ago.

BTW, Kill-a-Watt shows the power factor to be 0.64...


OMG! I always knew this PSU wasn't "top of the line", but I didn't realize it was quite this bad:
http://www.hardwaresecrets.com/article/413

 

[NeoPTLD]

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.

Ok, then, you'd have to copy both the current and voltage waveshopes for an entire cycle of the voltage, then divide them into about a hundred small sections, and multiply each section up and add the products, then divide by the number of sections, that's if you got a standard analog scope. With a digital, you can probalby get a program to perform integration function from acquired data.

 

[John]

Boomer, it looks like you really tortured your Aspire psu. Your 12V is way below ATX spec, and I can only image how bad the ripple is which could be damaging your hardware. Have you considered purchasing a quality psu?

 

[PurdueRy]

Originally posted by: NeoPTLD

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.

Ok, then, you'd have to copy both the current and voltage waveshopes for an entire cycle of the voltage, then divide them into about a hundred small sections, and multiply each section up and add the products, then divide by the number of sections, that's if you got a standard analog scope. With a digital, you can probalby get a program to perform integration function from acquired data.

Yea, I am from the new school where digital is in

 

[BoomerD]

Originally posted by: John
Boomer, it looks like you really tortured your Aspire psu. Your 12V is way below ATX spec, and I can only image how bad the ripple is which could be damaging your hardware. Have you considered purchasing a quality psu?

Yeah John...we've been discussing it here for a week or 2... Most likely, I'll be getting one of the Corsair HX 620's, even though it's overkill for me now...

 

[pcslookout]

Originally posted by: John

Originally posted by: dclive
Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

Try running 2 instances of Orthos and rthdribl at the same time. Now tell us what the kill-a-watt reads.

I tried that with my current system and I pulled 290 watts from the wall. I think my current power supply is perfect for my system for upgrade room!

Intel Dual Core 6420
Epox mobo
4 GB of DDR 2
EVGA Geforce 8800 GTS
WD Raptor 150 gb
Hitachi 500GB
DVD burner

 

[PurdueRy]

Originally posted by: pcslookout

Originally posted by: John

Originally posted by: dclive
Just ran Kill-A-Watt on the Acer E700 yesterday (4x SATA drives, Quad-Core Q6600, nVidia 8800GTS/320, 4x 1GB RAM sticks, Hauppauge dual tuner). 250W typical (it jumps around), 295W max observed - this "at the wall", so the machine (assuming 80% efficiency) is actually using about 240W, max, and about 40W less (about 200W) typical.

Try running 2 instances of Orthos and rthdribl at the same time. Now tell us what the kill-a-watt reads.

I tried that with my current system and I pulled 290 watts from the wall. I think my current power supply is perfect for my system for upgrade room!

Intel Dual Core 6420
Epox mobo
4 GB of DDR 2
EVGA Geforce 8800 GTS
WD Raptor 150 gb
Hitachi 500GB
DVD burner

So about 240W power actually needed. Sounds about right.