Measuring (AC) load represented by a PC

[Don Y]

Hi Mike,

The devil is in the details.
Based on the Kill-O-Watt recommendation below
Here's what my 2.8GHz. quad-core intel does.

State Watts Volt-amps
off 3 7
booting 120 174
Idle 88 132
Sleep 10 20

A newer system would more likely have a power factor
corrected power supply. I don't have anything that new.

Well, I dug up an electronic ToU KWHr meter I designed many
years ago and tweaked the gain of the front end -- better
suited to "a few amps" instead of a few *hundred* amps.
Then, measured the PC with that *and* the current clamp
(using the ToU meter as a "sanity check" on the readings
from the current clamp).

Numbers were similar -- though *higher* than your above:
100 - 140W (idle, active). I didn't bother with "sleep" or
"off" power levels as I am only concerned with the worst
case numbers.

If I can find the optical programmer, I bring the ToU meter
back and configure it to report peaks -- then see how I can
*create* those peaks.

That doesn't include either of the two 24" monitors.
93W 95VA each.

I was expecting the monitors to be ~20% less and was
surprised to see them come in at almost this exact same
figure (despite being much smaller -- 19"). So, I'll need
to rethink whether I should recommend dual monitors on
each workstation or just go with one. I could also point
to the extra counter space that they would consume. But,
the killer is having to support another KW! Sheesh!

(also makes me wonder what I'm drawing in my office with
6 monitors on! :< )

 

[mike]

(also makes me wonder what I'm drawing in my office with
6 monitors on! :< )

Measuring will scare you.
Since I got the ability to measure, I drastically reduced my
power consumption. Decommissioned my entire vampire stereo system,
put all the computers on short sleep intervals. I've got three
3W led lamps that light the entire place 24/7.
My neighbor calls me "the mole".

 

[Don Y]

Hi Mike,

Measuring will scare you.
Since I got the ability to measure, I drastically reduced my
power consumption. Decommissioned my entire vampire stereo system,
put all the computers on short sleep intervals. I've got three
3W led lamps that light the entire place 24/7.

I like/need LOTS of light! E.g., 10 "A-base" fixtures in the kitchen
alone. Obviously, had to move those to CFL's as "indecents"
would have made it unbearably hot! My bathroom has another 8...
and I still manage to cut myself shaving! ;<

My neighbor calls me "the mole".

We burn lots of KW. There are always at least 3 PC's running:
- this machine for email/WWW
- a machine that provides basic services for my private internet
- another machine that I use to FTP/HTTP stuff while this machine
is seeing other use (e.g., so SWMBO doesn't complain that the
machine is busy)
Plus several 100M & 1G switches to enable the network fabric.
Surprising how much power they eat!

When working, at a bare minimum, I can fire up a single display
on an X-terminal for a total cost of ~120W. Most of the time,
I'll fire up a pair of displays, though (so closer to 220W).

If I need a Windows machine, we're closer to 400W for two displays
(disk drives, etc.)

If I have to bring a server on line, this can creep up to double that
*without* a display!

Of course, these are in addition to the first three boxes -- and
ignore if SWMBO needs to use her laptop. Or her workstation!

[When I'm bringing up a new OS/app in a new bit of hardware,
its not uncommon for me to have two workstations running (so that's
4 monitors) plus an X terminal (2 more monitors). Thankfully,
some of the devices can be located in other rooms so I'm not
pumping extra BTUs into that little space! Though it does get
very comfortable on cold winter nights! :> ]

The real kicker is my working hours -- all over the map. So, its
usually a PITA to power down (or sleep) a machine as I will probably
be using it throughout the "day" -- assuming it isn't busy chewing
on some problem (e.g., "make all" or -- *gasp* -- "make world").

OTOH, we do try to economize on peripherals that don't need to be
up much. E.g., I have an LJ5p & LJ6p that we use for casual
printing. Nowhere near the power hog that the color laser,
Phaser or even the LJ4+ are! And, they sleep automatically so
I don't have to worry about powering them down if I forget
(though they are powered off when not in use).

Likewise, NAS boxen are off unless in use. Ditto for any external
drives. Though I need to build a switched outlet strip for all
the damn wall warts so I don't have to keep unplugging them when
not in use! Why the hell hasn't someone made this, already??

 

[JW]

That plus you could get a NIST traceable HP 3458A for somewhere in the $6k
range and know what it is doing to 8 1/2 places.

Well, as long as it doesn't draw more than 1A. Full scale AC current is
stated as 1.050000, so I believe it's only 6.5 digits. AC voltage is 7.5
digits.

 

[tm]

Well, as long as it doesn't draw more than 1A. Full scale AC current is
stated as 1.050000, so I believe it's only 6.5 digits. AC voltage is 7.5
digits.

Jeez, pick a few nits, why don't cha.


tm

 

[JW]

Jeez, pick a few nits, why don't cha.

Sorry. Test and measurement is all about being precise.

 

[mike]

We burn lots of KW. There are always at least 3 PC's running:
- this machine for email/WWW
- a machine that provides basic services for my private internet
- another machine that I use to FTP/HTTP stuff while this machine
is seeing other use (e.g., so SWMBO doesn't complain that the
machine is busy)
Plus several 100M & 1G switches to enable the network fabric.
Surprising how much power they eat!

When working, at a bare minimum, I can fire up a single display
on an X-terminal for a total cost of ~120W. Most of the time,
I'll fire up a pair of displays, though (so closer to 220W).

If I need a Windows machine, we're closer to 400W for two displays
(disk drives, etc.)

If I have to bring a server on line, this can creep up to double that
*without* a display!

Until recently, I just let the computers run when I was awake.
I have a PDA that constantly monitors and graphs the operation
of the gas furnace. You could easily tell from that graph
what time I woke up and turned stuff on.

I rarely have anything running when I'm not at the system.
I tweaked everything to sleep when not in use.
When they're all sleeping, the total baseline consumption for the house
is below 100W. And any can be brought back in a few seconds.

 

[Don Y]

Hi Mike,

Until recently, I just let the computers run when I was awake.
I have a PDA that constantly monitors and graphs the operation
of the gas furnace. You could easily tell from that graph
what time I woke up and turned stuff on.

I "went to be" hours ago (technically, I'm still *in* bed) but
have been "up" briefly every hour or two to check on the progress
of one of the machines and tickle it on to the next activity
(then, back to bed for another "short nap")

This tends to be the pattern here much of the day -- tickle a machine
and then move back to continue work on another. Of course, that
"move" may just be from one window served by machine A back to other
window(s) served by machine B. (You quickly learn to make sure
you know which machine has the focus lest you do something you
*really* don't want to do! :> )

I rarely have anything running when I'm not at the system.
I tweaked everything to sleep when not in use.

I'm impatient. I get annoyed if a screen saver locks the current
session and I have to retype a password, etc. Or, if an external
disk spins down and I have to wait for it to spin back up before
I can examine a directory thereon.

When they're all sleeping, the total baseline consumption for the house
is below 100W. And any can be brought back in a few seconds.

When the ACbrr is regularly pulling a few KW, refrigerator and
freezer chest each running on demand, any "conservation" that
I might undertake is easily in the noise.

But, its good to see the monitors "sleeping" (reawakened just by
moving the mouse) each save me 100W. *That* adds up!

I wonder how many watts burn (here) in idle wall warts, LED
indicators (smoke detectors, charge indicators on PDAs, etc.),
backlit displays (alarm clock, etc.)? I know it's never dark
in the house even with all the lights off, new moon, etc.!

<frown>

 

[whit3rd]

On 10/6/2013 12:51 PM, tm wrote:

Can you attest to their accuracy regardless of load characteristics?

I've A/B compared them with multikilobuck AC test equipment, they're
good to a percent or so; more than accurate enough to check VA and W
for planning your AC requirements. There are multiple models, and
some depth of items measured (you have to push the buttons to get
to all the readings).