After much research I purchased my first ever digital camera, a Canon Powershot SX120 IS, last month. Its major point of difference from other point and shoots was the 10x zoom lens, I also like that it looks like a real camera (it's black, solid, and lumpy around the lens), and has a smattering of manual options, such as shutter speed and manual focusing, should I feel the need to tinker.
All seemed well, except that I found whenever I came to use it after a few days break the batteries would be completely dead. As recommended by Canon I was using NiMh rechargeables, so I attributed the initial problem to the use of old batteries, and purchased some low self discharge Sanyo Eneloops. These prematurely flattened too. With three different pairs of batteries I'd only get about 50 photos from each pair, whereas the manual suggested over 300 would be expected.
As an experiment I fully charged some batteries, took three shots (without flash), then left the camera to rest. Every day I turned on the camera. On the fourth day it alerted me that the batteries were low, and seconds later that camera shut itself down. Is three photographs from two fully charged NiMh batteries before they failed over four days a record?
It's interestingly easy to measure how much electricity is being used when the camera is off. Popping open the battery compartment I used my trusty multimeter to measure the current flow between the two AA batteries. As a comparison, I also measured the off current of a Canon Powershot A550 I had handy.
A550: 0.5 µA
SX120: 20 mA
Wow, my camera sucks down 40,000 times as much juice as the A550!
Going further, the battery capacity of the rechargeables was stated as being around 2000 mAh each, from which we can calculated the expected life of the battery. If the milli and micro units get a bit complicated then let Wolfram Alpha do the work. It'll happily calculate sums such as "2000 mAh / 0.5 microAmperes":
A550: 2000 mAh / 0.5 microAmperes = 457 years
SX120: 2000 mAh / 20 mA = 4 days
I don't know about the four and a half centuries, but the 4 days seems to tally with what I'm getting, It's nice when theory and experience match up.
After persuading myself that my camera was faulty I took it back to the retailer and got it exchanged for another of the same model. The first thing I did when I got the new one home was to whip out the multimeter, and I'm happy to report that its off current was so low that it was below the multimeter's ability to detect. In a week I'll have the experimental confirmation of this low off current, but I think it's safe to say it looks like it was a fault with just that single camera, and not a model-wide design fault.