Validating MEMS and other sensors

-April 14, 2015

Sensors find places in almost every consumer, industrial, and automotive product these days. For example, mobile devices have accelerometers, ambient light sensors, gyroscopes, etc., and automobiles have airbag sensors, tire pressure monitors, proximity/pedestrian detection sensors, and more. No matter what applications they go in, ensuring sustained operation of sensors throughout their life-cycle is extremely important.

This paper discusses the various platforms used during sensor product validation. Depending on the type of sensor under validation (“SUV”), one or more of the below platforms are applicable for evaluation. The details of each of these setups are described below.


MEMS validation platforms

The below sections explain the validation platform used for exciting the MEMS portion of sensors and analyzing the device response.

(Un)Powered mechanical shock

The system shown in Figure 1 below is meant for providing mechanical shocks to SUVs in either powered state or unpowered state. The shock varies from few hundred gees to more than 4,000g levels.

In both the unpowered and powered mechanical shock tests, the offset (output of each axis at 0g level) of the SUV is observed on an oscilloscope before and after the shocks. The difference (as the name suggests) is whether the SUV is in powered or unpowered state while the fixture (with mounted SUVs) is being subjected to a mechanical shock.

The offset observed after the shock would determine the effect of shock on SUV. There are also options of providing the shock at various temperatures by controlling an inbuilt oven.

Figure 1. Mechanical shock system

Ball-drop test

Here, the SUV is subjected to shock by means of a ball being dropped from a height closer to board-mounted SUV screwed on a hanging metal sheet.

During and after this ball drop, the offset for each axis is continuously monitored to see the impact of the ball drop on the device output.

Figure 2. Ball Drop Setup

Shaker test

In this test, the board-mounted SUVs are screwed on a fixture in a shaker system capable of sweeping the SUV with varying g-level and different frequencies in sinusoidal/random pattern. The fixture can be either hexagonal or square in shape with boards mounted on each side (say, n) of fixture. These boards have multiple SUVs (say, m) soldered on it and. This way to allow multiple SUVs (n*m) be tested in parallel. Offset of all the SUVs is observed simultaneously along with the shaker’s stimulus for device behavior analysis.

Figure 3. Variable Frequency Shaker system

Bend test

Here a board with SUV is bent using the system as shown below. This it to observe the effect of deflection on the device offset.

Figure 4. Bend board system

Drop test

In this test, the SUV is tied to a specific weight and is dropped from a known height. After the drop, the SUV is placed in a validation board and checked for its functionality.

Figure 5. Drop Tower setup

ASIC & Communication I/F Validation platform

All the ASIC blocks are validated platform includes the usual test equipments like clock generators, power supplies, oscilloscope and/or a NI-PXI based test solutions. An automated environment can be built using Labview to validate the different ASIC blocks like clock generators, voltage regulators, ADC.

In order to validate the I2C/SPI interface, a I2C/SPI master can be used to send the commands to control the SUV and read responses from it. The master solution can be developed in such a way the timings of the interface can be controlled and validated across a range.

These tests are done at varying voltage and temperatures to ensure behavior across these conditions.


Sensors validation is complex because of the intense involvement of various setups/equipments to ensure their guaranteed operation across life-span. This paper explained briefly the various setups used for mechanical and ASIC validation of sensors.



IEEE Standard for Sensor Performance Parameter Definitions (IEEE Std. 2700-2014)

Loading comments...

Write a Comment

To comment please Log In