View Single Post
  #82 (permalink)  
Old 01-02-2008, 11:59 PM
monkeyboy's Avatar
monkeyboy
PPCGeeks Regular
Offline
 
Join Date: Dec 2007
Posts: 238
Reputation: 125
monkeyboy is keeping up the good workmonkeyboy is keeping up the good work
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quote:
Originally Posted by Beachbumcook View Post
Interesting results...

This is the kind of testing we have been looking for ...
Maybe the ABC battery meter program is deficient?

Maybe for the lack of knowing better, the "average Joe" has to rely on the programs and testing that one has available since we do not all have friends like yours?
I posed the technically pointed question about how a WM app like the ACB program could possibly provide accurate current/power data to the MSDN group handling power issues for WM devices. This is the response I got:
Quote:
Originally Posted by MikeCal
I have no idea how you could accurately measure power with an app. I can think of ways to infer power draw. It's relatively easy to get CPU times, LED blink rates, and knowledge of when the Backlight turns on, etc. If you knew which parts were in the device you could deduce how much power they're drawing, etc. It would be hard to tell if the audio amplifiers are powered, etc.
The way I do power analysis is with hardware. I remove the battery and attach a power supply to the battery leads. The power supply tells me exactly how much power its supplying to the phone, and dumps that data over USB to a nice little app that shows me graphs in real time. With no battery there, all power being drawn needs to be coming from the supply, so I know EXACTLY what's happening.
Mike
I believe he confirms my suspicion all along, which is that there is no way a WM app like the ACB program could provide an accurate read of the current or power consumption of arbitrary WM devices. Plus Mike implies that there are no such WM API hooks in the OS for such data monitoring. If you understood how current can be measure, it would be clear that it is unlikely that your average consumer WM device would be outfitted with what is needed to make such data available (it would add unnecessary complexity to the power mgmt circuitry and probably REDUCE the battery life in the process). In all likelihood, the ACB program is making an educated guess about power/current utilitization of the device based on CPU and I/O usage. -- but such GUESSES would not be very accurate.