On March 10, 2019, Ethiopian Airlines Flight 302 crashed, killing 157 passengers and crew. This was the second fatal crash of a 737 MAX aircraft.

Ethiopian Flight 302 data shows pilots’ struggle

Shortly after takeoff, the left angle of attack sensor goes haywire, (1) triggering cockpit alerts about unreliable airspeed and altitude readings and making the control column shake in a false stall warning.

At just shy of the 2-minute mark, when the pilot retracts the flaps and turns off the autopilot, the bad sensor activates MCAS (Maneuvering Characteristics Augmentation System), (2) which pushes the nose sharply down three times.
The pilots then hit the cutoff switches, turning MCAS off. But in the ensuing minutes, the forces on the tail make it impossible to manually pull the nose up.
At the 5:43 mark, they turn the cutoff switches back on, providing them electrical power to move the tail, and make two quick nose-up movements. (3) But then MCAS returns with a sharper nose-down movement.
The plane crashes to earth 6 minutes after takeoff (4).

——————————

What went wrong: The angle of attack sensor

The angle of attack sensor sends data to the MCAS on the angle of the plane’s nose. Too steep of an angle can cause the plane to stall. If the sensor sends false data, MCAS will active prematurely.

How the angle of attack sensor works

With the 737 MAX’s automatic system cut off, forces on the horizontal tail could make it very difficult for pilots to swivel it manually

In both the Lion Air crash and the Ethiopian Airlines crash, the MCAS was found to be a key contributing factor.