As I sit here watching numerous Joule Smasher Led Flashers drain their batteries I got serious about calculating where the power was being used. No real surprises at the top end: it's all about how long the led stays on and the length of the off period.
For the ATTiny13 part, using the Watchdog timer for timing only gives you the ability to generate coarse time intervals. Using boost control with a period control by the micro while running I get the following sensitivity analysis when the led is on for 10ms:
So a 1% change in the time in the off state, albeit impossible to achieve with a watchdog timer, is as big a difference as a 1% change in led current or a 1% change in time the led is on. No surprise there.
It appears there is little I can do in this application to improve the average current consumption except change the led on period. If I improved the efficiency of the booster by 2%, a big achievement from where it currently sits at 76%, then I will have made the same improvement as if I shortened the led on time by 0.3%.
But the maths did support controlling the voltage booster chip. But that's the subject of another post once the trial results are in.
No comments:
Post a Comment