I have finished my second quadcopter with ArduIMU. This time is a "mini" version with only 28cm from rotor to rotor and has the new Jordi´s magnetometer integrated!.
This tiny machine is fantastic, with a great stability. Look at the hands-off part of the video inside house!!
Very funny :-) Video:
Specs:
Motors: 1811 2000Kv 10grams brushless motors
Props: standard 5x3
ESCs : Turnigy 6A
Batt : 2S610 Lipo
Weight (RTF) : 180grams
Again it´s a very cheap and easy setup. Because I use standard propellers I mounted left and right motors with some twist. I´d really love arduIMU flat hardware...
Look at the PPM output Rx, Is this a new 2.4Ghz receiver form DIYDrones?
No, it´s a "lighweight" (4.1grams) mix between Spektrum AR6110 Rx and Jordi´s PPM encoder. This is nice for DIY projects...
The code is improved from the old version. This one has the magnetometer integrated so now we don´t have drift
on yaw axis, better implementation for the D term on the PID controler (more responsive to user inputs), same safety improvements, fast ESCs update rate (166Hz), new PID tunning...
But this history had an interesting lesson for me... The prototype you see here are really the second one, the first one was a dissaster... (flys bad...)
Do you want to know the "small" difference between them? This:
For the first one, I use an U shaped aluminium arms to save some weight. Well this shape is very bad in torsion and generates a lot of vibrations. This vibrations saturate the gyros (our gyros are quite sensible on vibrations because the low internal resonant frecuency). You could see here a graph of the gyro output value vs Throttle and you can see clearly the effect...
I change the arms to the square shape (8x8mm) and problem gone!! Perfect stability... :-) Vibrations are our enemy (Lesson learned)
Here is the new source code: Quad1_mini_20.zip (see notes.txt inside)
Old post (first version) here
New ArduPilot Mega will be a perfect platform for Quad´s also, so stay tuned...
Jose.
Comments
Could you please describe the "mix" between the AR6110 Spektrum receiver and the PPM board. I think I see another PCB there but wonder if it is simply a way to connect the four PWM receiver outputs to the PPM board - maybe the power is connected that way?
Dave
Roy, yes left motor tilted aft and right tilted forward. The arm with the red strips is the tail. I have not notice differences in yaw between right and left (maybe a little) and also there is not noticiable effects on roll. You must think that there is a active control of the attitude all the time, so if a roll movement generate some "mix" on pitch, the control will correct this effect automaticaly.
oguz, here is a direct link to the video : https://www.youtube.com/watch?v=keAyWKbTzK8
Jose.
Thanks
Oguz
Great work and a nice post!
Let me see if I understand this properly. The left motor is tilted slightly aft and the right one is tiled forward, right (assuming the barber pole arm is the front)? When you roll to the right, does the quad tend to move backwards (and forwards for left roll)? Also, does it yaw better to the right than the left?
Thanks
Roy
Very nice quadcopter, I love it. I am really inspired now and I will start spending some more time in programming to get my own cool project.
Nice pics and info.
Thanks for the amazing quality of documentation and the source code is very readable - especially the matrix/vector math. The one thing that is remarkable with this, unless I misunderstand, is that with the PPM receiver circuit, the ArduIMU does the autopilot effort as well. In other words, perhaps for a simpler aircraft autopilot but wanting an IMU, only the ArduIMU board (+magnetometer) is sufficient for some of us who do not need the hardware multiplex for MANUAL mode but just copy in to out servos and then the stabilize/autopilot/fly-by-wire modes can also be done along with GPS input. Again, for simpler requirements (no pressure inputs), maybe the ArduIMU + PPM might be simpler/lighter and easier than the ArduMEGA - what do you think? Maybe Bill Premerlani's wind estimation could be added as well.
>Jack, Believe me, my house is small and I could fly inside. Some of the scenes on the video are from my parents house ;-)
>Bill I like to hear from you. This quad is also your "son", you are the man behind the IMU DCM algorithm...
Some times we learn more from errors than from success
>Peter, I the old project I tried with and without prop savers and find no difference at all. Anyway I made some cuts with the cutter in the prop savers (below the props) to prevent slips and also the rubbers are very hard so the props are stiff.
>Simon, Yes I´d like to keep things as simple as posible. I will find the costs to make a list. Yes, if we don´t use GPS we have a free serial port. The drift of the yaw without magnetometer is not high (the old project was without it) and it´s a good option as an standard quadcopter but because I am planning to use the GPS for navigation, the magnetometer is neccesary.
Regards,
Jose
A couple of open ended questions:
1). Since there is no GPS, is the serial port available for flight control via Bluetooth/Zigbee/WiFi?
2). What was the yaw drift like without magnotometer? I'm still not convinced one way or the other whether it is possible to tilt accelerometers to provide yaw compensation for super low cost quads.
Excellent work, I am VERY jealous!
Simon.
PS. This would make an excellent Podcast interview subject....
-
16
-
17
-
18
-
19
-
20
of 20 Next