Meeting Details

• Location: Lobby Uni Tower
• Supervisor: ssm
• Protocol: lan
• Attendees:
  • apo
  • fma
  • lan
  • ssm
• Start: 10:00
• End: 13:15

Agenda

Drop after good flight

As investigated during the last experiment the Quadcopter is able to perform a stable flight, but after an uncertain amount of time it starts to oscillates randomly, which brings the necissity of an emergency stop. After some further review on the sensor values it is seen, that the D part of the PID-controller gives enormous values in the moment of receiving an user direction control. First attempt to correct this was to cahnge the D-value calculation formula, which was a mistake. This changing was undone after further testing.
For a better overvies what is happening the debug output was changed. The vies was limited on one axis but the output was enhanced to show all part of the PID-controller even the tick-time.
It was seen that the tick-time was floating between 2 to 11 microseconds. The problem with was to big D values was only if a uer direction control happens if a tick of 2 microseconds length occurs. During the D-value calculation the error is devided by the tick-time in seconds, thus a smaller tick-time gives a bigger D-value. Also it was observed that, if the user direction control, happens during a longer tick, the D value may be big but the copter is able to work against this.
After some further discussion it was seen that the current max PID-controller value was set to 50, which means, that in the case of max controller output a values of 50 is added/substracted of the motor values. The so far used motor value to fly was around 180, thus at full PID-controller amplitude some motors were set to 130. It was kown that the motors stop running by a values under 150 so it was clear, that the problem occured at the moment that some motors stopped running midflight.
The max PID-controller values was changed to 25 so the minimal motor values was 155.

Stop without Order

With this correction the Quadcopter was able to perform a longer stable flight but at some point it dropped to the flor as if an emergency stop had been perform but it has not. Review of the sensorvalues showed that the Gyro data at some point gave random values oscillating form -20 to 60 degree randomly. This lead to a by tick switching form -25 to 25 PID-controller value, which might have lead to motor stops. In a discussion it was decided on switching the MPU9150 sensor on the next day.