I have several UPSs here at home and am also concerned about their efficiency. Each is running through a wattmeter (Kill-A-Watt) and I can check their power consumption that way. Ideally I would also monitor the load side and compare that to the power drawn to get an idea of the efficiency, but I haven't gone that far yet.
One thing to be aware of about UPSs is that there are a few different types, ranging from 'standby' to 'online'. A 'standby' UPS is routing incoming power straight to the load (perhaps through surge suppressors), and also using some incoming power to keep the batteries topped up. Only when line power is absent are the batteries drawn upon. In this situation the overhead is only that extra wattage required to keep the batteries topped up. Most of the low-cost consumer-grade UPSs are of this style. My smallest UPS (Conext 900 AVR) works like this, and by removing all load I can measure that overhead wattage, in this case it is 8 watts. In theory, this overhead would remain at 8 watts regardless of load, and only go up when the batteries are depleted as a result of a loss of line power, until the batteries are again fully charged.
At the other end of the range from 'standby' is the 'online' UPS, where the load is powered completely from the batteries, and incoming power is only used to keep the batteries charged. Clearly this configuration will suffer in efficiency, with three main areas of loss: converting AC to DC to charge the batteries, batteries themselves do not return 100% of the energy they receive, and of course the DC-to-AC conversion on the output side. Online UPSs have several advantages over standby UPSs including zero transfer time and complete isolation from the incoming power, removing the need for surge/spike/brownout protection for the attached load. Their use is usually reserved for more critical applications.
Here is a webpage I found outlining the differences between the various types of UPSs:
www.gtechna.com/e/products/Powerware/online%20vs%20interactive.html
and here is an article I found discussing the issue of UPS efficiency:
www.electricity-today.com/et/issue0102/i01_ups.htm
I have one 'online' UPS, a 1500VA rackmount model by Tripp-Lite. It backs up my entertainment center which includes a PC, a couple of DVD Recorders with hard drives (which could be corrupted if power is lost while in use), and an HD cable box which will lose its program information if it loses power. With no load it draws about 25 watts, but what I haven't measured is how much that overhead increases in response to adding load.
I also have a 'do-it-yourself' UPS fashioned out of a combination battery-charger/inverter and a huge deep-cycle 12-volt battery (265 amp-hours). This setup is used to power the blower for the gas-fired furnace in our house in the event of a long power outage. Since this is not an application that requires instant switchover, I leave it unplugged except for once a month to top up the battery, and it usually runs for just 5 minutes before going into trickle-charge mode. This approach might be worth considering for someone with a UPS that powers something that is not turned on very often and for not very long (once daily at most perhaps). The energy required to keep a battery topped up all the time is typically much more than what it takes to bring it back to full charge after it has been unused for a period of time. Granted it is more of a hassle to do it that way, but if your PC is turned all the way off for most of the day, also turning off the UPS will most likely result in a reduction of energy consumption.
Edited by bobkart - Sun, 06 Jul 2008 08:19:51 GMT