Is the ADXL345/362/other capable of providing inclination data with accuracy of <1º?
This is for a consumer product so price is of great importance.
If I'm correct to get the inclination, we need to do Arcsin (acceleration).
This application note should provide useful details on how to implement the inclination sensing using a 3-axis accelerometer. The accuracy will be impacted by the offset along each axis, and the mismatch in sensitivities. For the ADXL345 and ADXL362, the offset is specified as typ, and following the example given in the application note under "Effects of Offset Error", the inaccuracy will be larger than 1 degree.
Great answer, STallur! Didn't mean to answer on top of you!
Thanks for the message!
Preserving that sort of accuracy, with a consumer price structure, will have some challenges, but may be possible. Some of those challenges may not be totally dependent on the sensors.
STallur's answer is primary, but I hope that this supplemental information helps you in asking questions. Sometimes, I find it useful to separate the errors into three different groups: initial, over temp, over time. The first two sets can be addressed via calibration (not free, but is physically possible), but the "over time" errors are often influenced by mechanical stress, which, due to normal behaviors in common assembly materials, can change over time. These behaviors can only be addressed via periodic calibration. Since the time frame between calibrations would be somewhat dependent on the application environment and mechanical stress profile on the system's PCB, we are not in a position to offer insight on "how often" calibration might be needed.
Retrieving data ...