It is generally understood that certain features, such as the GPS, draw a lot of power, but how much? I decided it was time to find out. I hacked up a battery so that I could connect a current meter in line with it, and I took data in an assortment of different states. Most measurements are averaged over seven minutes. They were taking with a Fluke 189 in logging mode and averaged in postprocessing on my computer.
I couldn't figure out how to insert a table in a post, so the following is the best I could do for formatting. You can see a nicer version on my blog at Blackberry 8820 Current Draw
GSM/EDGE 802.11/UMA Bluetooth Current, mA
Device Off 1.01
Off Off Off 2.94
Connected Off Off 5.1
Off Connected Off 8.1
Connected Connected Off 8.4
Off Searching Off 11.3
On Searching Off 16
Off Off Searching 5.8
Connected Connected Searching 10.5
Searching Off Off TBD
Use Case Current, mA
UMA Call 80
GSM Call 110
GPS On, No Signal + 37
GPS Tracking + 42
Screen Backlight, 10% + 12
Screen Backlight, 100% + 120
Keyboard Backlight + 23
The battery capacity of the 8800/8820/8830 is given as either 1400 mAh or 900 mAh, depending on which web site you read. I'll assume the former is correct for the sake of the illustrative examples below. Maybe someday if I have more time I'll test the capacity myself.
Let's say you want to calculate how long the battery will last when idling connected to the GSM network. You can look in the table and see that an idle GSM/GPRS/EDGE connection takes 5.1 mA. Divide the capacity of the battery (1400 milliampere-hours) by the current draw (5.1 milliamperes) to get the runtime, in this case 274 hours. If you also turned on an 802.11 connection, the current draw would increase to 8.4 mA, and the runtime would decrease to 167 hours. If you turned on the GPS, that would suck down another 42 mA, for a total of 50.4, or 28 hours of runtime. If you were making a call on UMA at 80 mA, and also had full-brightness backlight on at 120 mA, for a total of 200 mA, you could expect 1400/200=7 hours of battery life. You get the idea.
There are some interesting conclusions here. UMA uses less power than GSM to make a call, but more when sitting idle. Not shown in the above table is the fact that whenever you first initiate a UMA connection, the phone also apparently attempts to turn on the GPS, until it either gets a fix or fails for about four minutes. Presumably this is so that your physical location can be reported in - I find it vaguely disturbing that this is detected and (presumably) reported as a matter of course every time I use UMA.
I also wanted to test how much power it took when out of the service area, trying unsuccessfully to establish a connection, but I haven't had an opportunity to do that yet. I put the phone in a metal box, but it was still connected; I'll have to wait until I go out of town to do that test.
I hope you find this useful!
July 9: Edited to add 16 mA consumption for GSM/EDGE connected and 802.11 searching. It looks like you should turn off 802.11 when you're not using it, because having it on and searching for a network triples the idle power!