Vibrations will be different when a prop changes relative position to the wind, so prop vibration is not relevant to accelerometer calibration. Prop vibration has tremendous influence on gimbal performance and image stability but where calibrating accelerometers for flight stability purposes that vibration is irrelevant,
unless the aircraft is in flight during calibration. Vibration is a combination of forces, and a cyclic vibration during an accelerometer calibration does really bad things to the calibration.
Your Pixhawk clone calibration instructions mimic the instructions provided for Arduino calibration processes. As you mention you have performed additional research relative to accelerometer calibrations you have probably noticed that most of them reference the use of a "table" of some sort to establish a known baseline that is level, immobile, and used as a flat plane to place an accelerometer on in different positions during the calibration process. An accelerometer being calibrated is placed on the test table to measure the initial gravitational value and re-positioned according to manufacturer instructions, which will likely comply with an ISO standard, to obtain reference values for different orientations.
Most calibration processes include the use of gravity for the initial baseline value as gravity is almost, but not quite, a known constant. There are minor variations of the gravitational field that will be encountered at different geographical locations. Those requiring an "absolute" calibration obtain those values and factor them into their computational process but our toys don't require that level of precision. We should recognize that
any movement of an accelerometer during the calibration process causes a change in the measured force vector, so movement of a sensor prior to establishing a zero reference point invalidates the calibration process.
When and where larger UAV's employ avionics suites employing accelerometers, the avionics module is initially calibrated on a calibration table. The calibration values are recorded and input to the avionics where they can be reviewed prior to every flight to verify that normal drift has not impacted accelerometer force values. If they have been impacted there are offset values that can be input to "re-zero" an accelerometer. If an accelerometer fails to demonstrate force values that fall within the tolerance after being offset they avionics suite is declared unfit for use. All of this is performed in a static test.
There's a tremendous amount of math used in the calibration process but unless we are acting as a certified calibration test lab we don't need to know it for multirotor calibrations, we only need to follow directions provided by a manufacturer. If a manufacturer provides instructions that are in conflict with commonly accepted and widely recognized calibration procedures we might want to question the manufacturer in order to obtain explanations that qualify their reasoning.
We should also consider we are dealing with Yuneec, a manufacturer with a long history of developing incomplete, inaccurate, and poorly written user manuals, that has never admitted to being in error, that has never admitted to having product or component defects, and fails to provide owners with probable causes of failure after performing warranty repairs.
Me thinks a lot of the debate about the static versus dynamic calibration method stems from a general lack of understanding of what an accelerometer is, what, how, and why they do what they do. Among the multirotor community there is also a considerable lack of aeronautical knowledge relative to flight, which in turn causes confusion with developing a general understanding of an autopilot. If people would individually research accelerometers they would develop a much better understanding after a review of various types of accelerometers and the methods used to calibrate them than they will obtain from a multirotor product forum. I'm including a like to one manufacturer's calibration instructions to provide a visual example of a calibration table.
Calibrating 3-Axis Accelerometers
