Setset... very accurate in some aspects in a techincal brief, but none of that has to do with the problem with the moment and it's battery life
The problem is pretty simple in the end..
All these devices can't utilize a direct volt or amp meter.. They utilize an ADC circuit and then use a calculation to determine voltage... they then integrate the voltage curve over the nominal voltage and create a linear decay curve that represents the majority of the voltage except for the very top and very low (very top you have a massive voltage dropout between max voltage and nominal voltage and the same at the bottom... massive dropout after nominal voltage into minimal voltage).... Now this is all fine and dandy and works VERY well in the end.. Except that for the moment they are utilizing an INCREDIBLY low resolution ADC which doesn't include a buffer region... The behold II on the other hand uses a 9bits calculation and has an incredibly HIGH buffer range per voltage (moment outputs ADC values as 10^3 while Behold II outputs adc values as 10^4)... This causes a problem when you have a large demand on the power system, since the voltage is a order of magnitude less then what it should be, any small drop is percieved as a huge voltage change...
I have written all this up over at SDX and explained how i modify the ADC values from stock to integrate a better buffer region so it doesn't drop off to 15% so fast BUT my new decay curve is to linear for my liking (it is based off the numbers samsung calculated and they aren't right (i haven't had time to actually recalibrate the curve) so with the modified cl14 kernels, it isn't as accurate as it was with teh cj05 kernels (since i modified both of them and the cj05 code's decay curve looked incredibly better.. but the values were off in my calculations)...
I have wrote all this to samsung and asked them to switch over to 9bits for higher resolution and got 1 response from the actually code programmer that they would look into the possibility of utilizing this... but no promises