By the way a 20aH battery will not supply 20A for an hour in practice, it may in theory but I am not sure down to what voltage level the batt manufacturer has taken his figures from, I would suggest that the battery will have have discharged to a level that see permanent damage.
What you need to do is figure out how many watts your load will draw. Then grab the manufacturers specs for the battery, and work out your watt/aH/cell figures from that. If for example you download a Yuasa manual it should have the kind of table you need on page 7.

Take a look at the acp site and some of their run time charts
http://www.apc.com/products/runtime_for_extendedruntime.cfm?upsfamily=21

If you graph them you will see they are not a straight line but more like a ski jump curve (to be specific it is a power curve)

They also have a selection engine that will get you to the right UPS
http://www.apc.com/tools/ups_selector/index.cfm

Chris -
   You're confusing instantaneous watts, vs measured energy in watt-hours. A UPS can say that it is a 300 watt unit - meaning that it will support a 120 volt load, with perhaps 2.5 amp current draw. It does NOT say how long it will support that load - or if it does, it mentions runtime in minutes. many UPS intended for serious commercial use have optional add-on battery packs, so you buy the UPS for the amount of current it will have to supply, and you add battery capacity to add time that the unit will support that load.
  At a previous job, in a law firm, I had a 5KW Ferrups, with four big wet-cell storage batteries. The software with it let me know what the load was at any given time, and how long the intalled battery pack would support that load. As I powered down equipment, the runtime reading would increase. Normal daily load would give me maybe 30 minutes runtime. When our building did it's annual powerdown, for about 24 hours, I powered down all the servers, but left the routers and T-1CSU/DSU's running, so I woulnd't have to restablish all the T-1 connections whent he building came back up. w\With just routers and T-1's for the load, I had a runtime more like 40 hours, which saved me a lot of work when the power came back up.
I hope this helps you understand UPS's better. Think of a flashlight that can run on AAA, AA, C, or D batteries. All the batteries have the same voltage, but contain more milliampere-hours of energy. Given the same load, the bigger battery will keep the light lit longer. for a given battery, if you use a bulb that draws less current, it will run longer. That's why the new LED flashlights are so neat - much more light, much less battery drain.

Fred Wagner

 

Yup thats what I learned. I am used to seeing watts represented as "watt hours" so was caught off guard when it said 200 watt ups the problem is that is very hard to find the actual "capacity" of these units. IE even when contacting individual manufacturers they just have no idea I can not even use there run times since the unit run times vary dramatically based on load (far less efficient under heavy load)

YES LED's Rock. I only have LED flash lights. my 4000+ bulb christmas lights were 100% LED and my bedroom and both bathrooms are 100% LED.

Chris - another reason the UPS manufacturers can't specify definitively what the runtime would be is the Power Factor. You need an EE course to totally understand it, but it relates to where the phase of the current peak is in relation to the voltage peak - if it's a resistive load - a heater, the current and voltage peak in phase, and power factor is one. If they're out of phase (ac motors involved), the power factor can vary - as low as 04. or as high as 0.7.  A lower power factor will let the UPS run longer for a given amperage load....

Fred Wagner

 

yes but they could not even tell me the amp hour rating of the batteries in the units. that should be pretty standard "from the battery" kind of stuff. I am running such a small load that it would give me a half way decent idea of how many cells I need to run this thing for the time I need.

I walk into a radio shack and it says right there on the boxes 6amp/hr 9amp/hr etc.. thats all I wanted from them hehe

The Amp/hr rating should be on the UPS unit itself, or it might even specify what make/model of battery. Sounds like you were talking to the shipping clerk - they just know the model number and the weight.... but as ttoomm mentioned, the amp/hr rating doesn't mean that much, because the output voltage will become too low to support a load before the battery is completely drained. That's why the software supplied by the vendors of higher-end UPS's knows the load curve of their units, will tell you fairly accurately the minutes remaining with the actual load you're putting on it.

Fred Wagner

 

Nerys,

You can normally find what battery your unit has in it by either downloading the manual for the unit, or going to a website like http://www.batterywholesale.com/ (where I buy my batteries) and find what replacement battery your unit takes.

Most manufacturers also publish run time charts that show run times for a particular unit at varying loads.

In general I would say that using Watt on UPS's isn't the best to do. KVA is a lot better.

As I have learned from some UPS specialists calculating any expected remaning power on the UPS over 25 min will never be correct.
Don't remember for sure why, and at that time it did sound very strange for me.
But I have gotten the explanation more the once.

Last time I got the explanation was a year ago when I needed to plan to have 160KVA redundant for 90+ min.

As stated before the power factor can be tricky. It is difficult to understand, but also difficult to get real information from the server vendors on that you can use.

UPS and power in general is tricky and thank God that UPS, Power, cooling and Diesel isn't part of my job any more.

/johnny