Computer power usage

Posted by Robert Morein on October 17, 2005, 10:42 am

I used a Kill A Watt to analyze my setup. There were some surprises:

1. Modern 21" Sony based WEGA monitors use only 80 watts a 1600x1200, 60Hz.
This isn't much more than a flat panel.
2. CyberPower 1500VA UPSes measured at 30 watts no load, fully charged. It
was quite a disappointment that they are such power hogs. Is there anything
more efficient in this power range?
3. Cable modem, routers, firewalls, VoIP appliance, fiber optic links
totaled around 80 watts.
4. A Dual Athlon with an Nvidia 5900 came in at a mere 225 watts while
running in 2D mode. PS is a 600W PF neutralized Silenx, which is made by
Fortron.
5. A P4 2.6/400 with an Nvidia 5200 came in at 150 watts, in 2D mode. Supply
is a less efficient 400W Computer Warehouse, not PF neutralized, probably
also made by Fortron.
6. A Via C3 low power machine, which I constructed using a Supermicro socket
370 mb, for use as an email/text processing always-on machine, was a bit
disappointing. In spite of the fact that this processor idles at just a
couple watts, total power draw was 100 watts. Video card is a fanless 2D
only Matrox G550. Power supply is a PF neutralized Silenx 350 watt unit.
Where is all the power going?

Total power draw with just the Via machine two 21" monitors is 383 watts. I
can knock off 80 watts by turning off one of the monitors, but the total is
still disappointing.

Comments?

Posted by Danno on October 17, 2005, 12:23 pm

  Found the same with my C3 integrated board (1-Gigapro); used far more power
than I expected it to.
  The best consumption I've found is from an IBM Thinkpad 560, P100, DSTN,
no optical or floppy drive: in normal use, runs between 11 and 20 watts
(19 - 30 Volt-Amps). It works well for basic tasks.
  You may want to measure your computers again, but set your Kill-A-Watt
meter to Volt-Amps instead of Watts. The true power consumption of electronic
equipment doesn't measure well with the "Watts" setting. For instance, I have
a 19" NEC CRT, which comes in between 60-90 watts, but the VA measurement
shows an average of about 150VA. Quite a difference. The watt measurement
works well for things like lights and coils, but VA is better for things like
computers and stereos.

--
Solar-powered Linux: Slackware 10.0, 2.6.13, Averatec 3250H1 w/accel S3Uni
RLU#272755

Posted by Vaughn on October 17, 2005, 7:12 pm

     I am a bit lost here.  Do you understand the relationship between Watts,
VA, and Power Factor?  (True Power = Apparent Power  X  Power Factor.)  If (for
example) your KAW reads 75 Watts and 150 VA, that implies a Power Factor of 2.
2 is a terrible power factor, but about normal for today's electronics with
switching power supplies.  Is that about what the KAW is indicating?  If so, it
is working properly.

Vaughn

Posted by RF Dude on October 17, 2005, 8:45 pm

PF is a number less than or equal to 1.0. Resistor loads are 1.0. Then you
have leading and lagging PF depending on if it is capacitive or inductive.

Posted by Vaughn on October 17, 2005, 9:09 pm

     <Blush> You are exactly right.  VA  X  Power Factor = True Power.  Danno's
monitor has a power factor of .5 (not 2).  I don't have my KAW here at home to
compare, but the figures he presented still sound about right.  I don't have any
expensive test equipment to compare my KAW to, but the figures I have seen on
mine for Watts, VA and PF seem plenty believable enough for normal work.  YMMV

Vaughn