Power consumption -- real vs actual
jal at mdacorporation.com
Tue Apr 6 10:30:59 CDT 2010
> -----Original Message-----
> From: Matthew Geier [mailto:matthew at acfr.usyd.edu.au]
> Sent: Monday, April 05, 2010 5:02 PM
> To: John LLOYD
> Cc: linux-poweredge at dell.com
> Subject: Re: Power consumption -- real vs actual
> John LLOYD wrote:
> > The nameplate rating, used for electrical safety, is 7 amps for an
> R710 (100V to 120V).
> > The power supply rated capacity is 870W (for the high-output
> > The enterprise power calculator dell.com/calc says input power is
> 431 watts, maximum 578 watts.
> > Open Manage reports power consumption (at idle) is 252W.
> > Now, with this surplus of numbers to choose from, I have a few
> > First, the enterprise calculator shows maximum (578) less than power
> supply capacity (870) which is a good thing.
> > But, second, the power supply capacity is more than rated input which
> is 7A * 100V=700VA. What is the basis for the current rating on the
> Watts != VA
> I don't know what the 'Power Factor' is for a R710 power supply so
> can't check the figures line up. Generally the Watts rating is smaller
> than the VA rating. 7A x 100V is Watts not VA, so the fact that it's
Very much not true. The sticker is supposed to be VA.
The line cord is rated for insulation, and current. Devices are rated for VA, everywhere there is a regulatory regime. So you have to provide a line cord that meets the current ratings (almost always very easy in North America since there is really only one standard for line cords for 120V equipment), and a power supply that meets VA requirements.
My issue is the power rating of the device itself. Of course VA could be a lot higher than watts, e.g. motors, old switchmode power supplies, etc. Modern power supplies have to meet efficiency standards, and power-factor standards. The R710 power supplies are probably 0.95 or better PF.
(For the uninitiated, AC power, measured in watts, depends only on the current and the voltage which are exactly in phase with each other. Some electrical loads, e.g. motors, generate magnetic fields which store and release energy and result in the current being out of phase with the voltage. It's just the way it is. The current can then be considered as two parts: the part exactly in phase with the voltage, and the other part being the rest. The total current is what the power cord has to carry, although it is larger than the watts delivered. The ratio of the two is (approximately) the power factor. The total current times voltage is the VA. The part of current exactly in phase with voltage times voltage is watts.)
> lower than the rating on the power supply would be correct if the power
> supply sticker is VA and not watts....
which is the standard.
> Being in a 230/240 country creates other issues for us - things that
> come with 15A line cords as on 100v they draw more than 10A thus need
> the higher rated cord. On 230v they are below 10A so don't need the
> higher rated cord/socket, but we have to use a 15A outlet anyway or
> change the plugs on the supplied cords :-)
Oh yes, we've seen this too. Try getting customers in Germany to understand what the total power consumption of a rack is, and that they have to supply two circuits not just one. (German power distribution seems to be fairly low in amps per circuit, relatively speaking).
IEC320 cords (rated for 600V, and therefore good for 120 or 208/230) are good here. These don't require "national" power cords, and you can run most equipment on 120 or 208/230, whatever is convenient, from one distribution panel supplied with a higher-power circuit. The only exception seems to be laser printers, for which we have to supply localized versions.
My main issue is for power-constrained situations where I need to be able account for every watt and every VA, and know the modes of operation. It is something of an eye-opener to see R710 power jump from 250 to 430 watts just because the CPU is cranked up. Another eye-opener is the VA rating, no load applied, of a power strip. You might expect zero, and you would be wrong.
More information about the Linux-PowerEdge