I want to report the results of the timelapse test I did this morning.
The camera is an A590IS, firmware 1.01b, CHDK version 719. The batteries are new Eneloops which have been full-cycled three times, the most recent cycle being three days ago. All unneeded functions (IS, etc) were turned off, and the test was run with the display turned off, and with aflock used to fix the focus. The camera was in M mode. The picture interval was set to be a nominal 4 seconds, but actually averaged 4.6 seconds. The memory card is a 4GB SDHC.
Well, it kept going like the Energizer (Eneloop?) Bunny, and I finally ran out of time and had to shut down the test. At that point, it had run for 3 hours 26 minutes, and had taken 2,679 pictures. These were 1600x1200 Superfine. When I switched the display back on, the battery voltage settled down to 2.512V. I think this means there was still plenty of juice left in them, and the test would have continued for a few hours more, or until the card filled up. My camera shuts down at around 2.2V.
It seems pretty clear that it's the display that uses up batteries, not taking pictures. Well, the display and maybe other things like histograms, IS, and so forth.
There are two things that bother me a bit about the test. The first is that the camera makes a good bit of mechanical noise if you listen closely, and my previous experience is that MF mode eliminates most of this. So I'm going to give that a try, with a plug inserted in the AV port, to see if I can minimize the wear and tear. If battery life is even half this good under MF, that would be plenty good enough.
The second problem is that even though everything was completely fixed - focus (via aflock), shutter, aperture, ISO and white balance, the pictures vary somewhat in brightness, which is a bit disconcerting when speeded up in timelapse. I HOPE this is because the pictures were taken in fluorescent light, which has flicker. If not, then my camera may not be up to this task.
The video is at this site, but you'll have to download it to view it. Just right click on the icon. Divx codec. It's just a jar of melting ice with an electric clock next to the jar:
http://drop.io/dvts5vqIn terms of battery life, this test was far more favorable than I expected. I think it means with the right preparation you should be able to do quite lengthy sequences on one set of batteries. Now I need to find out how much much difference MF makes.
