[geeks] Got the CurrentCost meter installed

[Patrick Finnegan]

On Tuesday 18 January 2011, gsm at mendelson.com wrote:
> On Tue, Jan 18, 2011 at 01:44:19AM -0500, der Mouse wrote:
> >I think that means they can't sense anything but current, thus
> > making it impossible for them to sense true power if your effective
> > total load is not resistive (no power factor correction) and
> > definitely not capable of correcting for things whose draw is not
> > sinewave - such as switching power supplies, to pick a totally
> > random example. :)
> 
> The point is does it really matter? The device in question is used to
>  estimate power consumption not truely measure it.

I think that we're arguing about different things.  My point is that 
it's doing the measurement in a way that isn't consistent with how the 
power utility does, and potentially gives you much different numbers 
than what you'd actually pay.  So, here's some numbers go go with Mouse 
and my concern; hopefully this will help illustrate my point.

I have a quad-socket Opteron box.  Plugged in, but turned off, the 
metered PDU shows that it draws about 1A at 240V.  So, 240VA.  However 
the utility company is billing for Watts, and the actual used standby 
power is probably no more than about 20W.  So, the power usage shows up 
as about 10 TIMES what is actually being used. (As a side note, this 
means I have a pretty horrible power-factor, which utility companies 
really dislike).

I've got a bunch of machines that I leave plugged in, but turned off, so 
that I can use WOL or similar methods to turn them on when I want to use 
them.  So, something that meters like this can easily show a power usage 
that a few times what is actually being used in a situation like mine.

This is why I want something that measure actual Watts used, not just 
Amps (or even volt-amps).  Some inexpensive plug-in meters like the 
Kill-o-Watt (not sure if it can, but another one I have can) can tell 
the difference between Watts and Volt-Amps, and can be useful for 
comparing VA and Watts on a single device.

Pat
-- 
Purdue University Research Computing ---  http://www.rcac.purdue.edu/
The Computer Refuge                  ---  http://computer-refuge.org