FR

Description

I throw myself in designing and realising a quadcopter. I have started on febuary 2011. I work on Arduino and AeroQuad cards for electronic realization between sensors and the microcontroler. Otherside the program inside is developed by myself, with at the beggining help from AeroQuad original program. The mechanical part, is DIY, in carbon fibre and aluminium tube.
I control with a joystick and the communication is not like an R/C command.

Technology

The microcontroler is an Arduino Mega2560. I have added AeroQuad shield and a wireless module Synapse RF266 (2.4Ghz, Zigbee).
The ground station is an Arduino Due with wireless module (2.4Ghz, Zigbee) and also a bluetooth one.
The quadcopter is drived throw a jostick decoded by the Arduino Due or by the laptop (old version). All communication, sending orders and receive telemetry, are done with the couple wireless module Synapse (Xbee like). I don't use any R/C transmitter/receiver protocol, like other drones.
You can see the list of equipment used

Yanonix Quadcopter Photo with 10s delay from the ground Quadcopter on a test bed Aeroquad electronic

Actual state

I believe it can fly. But it's really true, it flies. I had few major problems with stabilization (see below), but now it's the past. It is stable, strong and aggressive. I use 3S batteries and 4S batteries, with the last ones, it's amazing.
Currently, it doesn't have a GPS module.
With simple calcul, it can progress verticaly up to 10m/s.

Problems

Actual

  • Radio modules Synapse RF266 are inetressting. You program them with a Python like language, it's funny. So you can design your own procotol between Arduino and the radio module or the computer. Nevertheless it's heavy while working. The Python is interpreted throw a virtual machine in ATmega128RAF1, so you lose time. Recently, my Synapse modules have abandoned me, alone, I don't know why. So I work with new module (again, before Synapse, it was Xbee module), with which I can go down in protocol stack, and directly in C. I want it, because I need performance (huge amount of data) so I have to optimized the process. This time, I will use 2 modules in the same time. I duplicate the communication for being more reliable. The communication is as important as the rest of the quadcopter. So I will have another Arduino component which will receive all data from the two radio modules and send the best information to the central Arduino.

Solved

  • My oldest radio modules were Xbee. Unfortunately, I had sometimes transmission cut during 2 seconds. In other words, it couldn't fly, or it crashed. I suspected a weak signal from the transmitter or reception (50mW in theory). I remembered that I communicated with unicast system, not the broadcast one, in a result I had acknowledgments for every messages non-important, it overloaded useless the bandwith.
    To fix that I bought new module : Synapse RF266PC1 with a theoritical power at 100mW.
  • There are vibration that give crazy the accelerometer. I tried to place rubber but without success. In stationary flight, I was arround 0.5G instead to be at 1.0G. I could only fly with the gyroscope alone. I hadn't no automatic stablization. The problem was the scale of the sensor. It was between -2.0G and + 2.0G. Now it's set to -4.0G and +4.0G. There are filters aboard the sensors, and I have vibrations outside the band of +/- 2.0G, so the calcul is wrong, at the end.
Writed on
Loading...