MEMS 3-Axis Accelerometer (ADXL345 / MPU6050)

What is a 3-Axis MEMS Accelerometer?

A 3-axis MEMS accelerometer measures proper acceleration—the rate of change of velocity experienced by a body relative to free fall. It senses both static acceleration forces (such as the continuous 1g downward pull of Earth's gravity) and dynamic acceleration forces (such as impact, shake, vibration, or rapid robotics movement).

Built with microscopic polysilicon surface-micromachined structures, MEMS chips like the ADXL345 and MPU6050 translate physical mechanical deflection into differential capacitance, which is digitized via onboard high-resolution ADCs and transmitted over I2C or SPI serial buses to microcontrollers.

Electrical Best Practices & Noise Mitigation:

• When wiring the MEMS 3-Axis, ensure power leads are kept short to minimize stray loop inductance and inductive bounce during rapid current draw changes.
• If using an analog sensor output with a 3.3V microcontroller like an ESP32 or STM32, verify the maximum sensor output does not exceed 3.3V. If operating from a 5V supply, insert a 2:1 resistive voltage divider (e.g. 2kΩ / 1kΩ) on the signal line.
• Place a 100nF ceramic capacitor across the sensor VCC and GND pins as close to the module package as physically possible to shunt high-frequency ripple generated by nearby motors, servos, or Wi-Fi transients.

Inside the chip, a microscopic proof mass is suspended by polysilicon springs between fixed capacitive plates. When subjected to acceleration or gravity, inertia displaces the proof mass, altering the differential capacitance between the moving fingers and fixed plates. An internal phase-locked demodulator measures this delta-C and converts it into voltage.

When stationary on a flat table, the sensor measures +1g (approx. 9.81 m/s²) along the Z-axis and 0g on the X and Y axes. By computing trigonometry angles on the static gravity vector, the exact Pitch (tilt forward/back) and Roll (tilt side-to-side) angles can be derived without needing a gyroscope.

// Production-Grade I2C Accelerometer Reader (ADXL345 / MPU6050)
// Platform: Arduino / ESP32 (volt-X Lab Implementation)

#include <Wire.h>

const int MPU_ADDR = 0x68; // I2C address of MPU6050
int16_t rawAx, rawAy, rawAz;

void setup() {
  Serial.begin(115200);
  Wire.begin();
  
  // Wake up MPU6050 (clear SLEEP bit in PWR_MGMT_1 register 0x6B)
  Wire.beginTransmission(MPU_ADDR);
  Wire.write(0x6B);
  Wire.write(0x00);
  Wire.endTransmission(true);
  
  Serial.println(F("MEMS Accelerometer Initialized via I2C (400kHz)"));
}

void loop() {
  // Request 6 bytes starting from register 0x3B (ACCEL_XOUT_H)
  Wire.beginTransmission(MPU_ADDR);
  Wire.write(0x3B);
  Wire.endTransmission(false);
  Wire.requestFrom(MPU_ADDR, 6, true);
  
  if (Wire.available() >= 6) {
    rawAx = (Wire.read() << 8) | Wire.read();
    rawAy = (Wire.read() << 8) | Wire.read();
    rawAz = (Wire.read() << 8) | Wire.read();
    
    // Scale factor for ±2g range: 16384 LSB/g
    float ax = rawAx / 16384.0f;
    float ay = rawAy / 16384.0f;
    float az = rawAz / 16384.0f;
    
    // Calculate pitch and roll in degrees
    float pitch = atan2(ax, sqrt(ay * ay + az * az)) * 180.0f / PI;
    float roll = atan2(ay, sqrt(ax * ax + az * az)) * 180.0f / PI;
    
    Serial.print(F("Ax: ")); Serial.print(ax, 2); Serial.print(F("g | "));
    Serial.print(F("Ay: ")); Serial.print(ay, 2); Serial.print(F("g | "));
    Serial.print(F("Az: ")); Serial.print(az, 2); Serial.print(F("g | "));
    Serial.print(F("Pitch: ")); Serial.print(pitch, 1); Serial.print(F("° | "));
    Serial.print(F("Roll: ")); Serial.print(roll, 1); Serial.println(F("°"));
  }
  
  delay(100);
}

Code Architecture Breakdown:

• Direct Register Addressing: Communicates directly over the I2C bus with register 0x6B to wake the power management state machine.
• Atomic 6-Byte Burst Read: Reads X, Y, and Z acceleration registers in a single transaction, preventing skew between separate axis measurements.
• 16-Bit Signed Integer Stitching: Merges high and low bytes `(Wire.read() << 8) | Wire.read()` to reconstruct true two's complement binary readings.
• Trigonometric Pitch and Roll: Uses atan2() for full four-quadrant angle resolution without division-by-zero singularities.

✅ Engineering Advantages

• Provides absolute drift-free gravity reference vector when stationary.
• Microscopic semiconductor footprint fits into the smallest wearable and IoT form factors.
• Extremely low current consumption allows continuous battery operation for years.

⚠️ Physical Constraints & Limitations

• Sensitive to external mechanical vibrations and motor noise that corrupt tilt angles.
• Cannot measure yaw (rotation around the vertical gravity axis) without a magnetometer.
• Double integration of acceleration to find position suffers from quadratic drift error.