Accelerometers test procedures

Offsets and sensitivity at 25C were measured after devices were mounting on the PCB using a "flipper setup" (see figure below). The "flipper" is the robotic metallic PCB holder and can be rotated (flipped) in multiple directions. Since additional stress will apply to sensor after devices are soldering to PCB, offsets and sensitivities of devices may change from values measured during production test.

The DC offset of all output channels are measured at ambient only, and the distribution histograms are calculated from these measurements.

The flipper have the capability to rotate accelerometers and align one of its sensitive axis to gravity in both positive and negative directions. To minimize alignment error, the accelerometer sensitivity is calculated with its measurement at both positive and negative 1 g orientation. For example,

where X(+1g) and X(-1g) are sensor output with +/- 1g input.

After sensor sensitivities at each test temperature are measured, their values are typically plotted against oven or device temperature. The overall sensitivity change over entire temperature range should be calculated and reported. Any abnormal behavior such us sudden jump or significant shifts near hot or cold temperature set points should be investigated.

Sensitivity distribution at 25C is also derived from this test, from the entire sample size.

Number of devices of test: 96 from 3 lots

Power supply: Vdd=3.3V

Output sample rate: 100 Hz

Number of samples per measurement: 100

Temperature ramp rate: 1C/min

All offset vs. temperature curves are normalized to initial 25C readings.

Sensor Output Linearity Test

This test is to measure maximum linearity error of sensor output. Linearity error is calculated as the difference of a measurement point to the respective value of the best fit line.

1. Device under test are mounted to a position near edge of a high precision rate table, with device’s measurement axis point toward the center of rate table, similar to the setup shown in the figure below. Measure the exact distance from center of device to center of rate table.
2. Calculate the rotation speed at each test acceleration level.
3. Turn on rate table and set it to the rotation speed based on test acceleration level. Record accelerometer outputs in the measurement axis at each acceleration input level.
4. Rotation device 180 deg and repeat step 1 to 3. Acceleration is applied to opposite direction now. Note: centripetal acceleration is always pointed toward the center of rotation.
5. Combine data measured in positive and negative orientations. A small misalignment or distance difference between two mounted orientation could cause large difference of linearity error near zero acceleration point, so data has to be post-process to remove such errors.
6. Repeat same test for other measurement axis.
7. Calculate the best linear fit straight line from test data and calculate difference of a measurement point to the respective value of the best fit line at each acceleration level. Plot linearity error vs. input acceleration and check for any abnormal shape.

The test conditions are:

Input Axes: X, Y, Z

Excitation level: 1 to 200 g

Frequency: 100 Hz

Device ODR: 3200

Number of devices under test: 15