30 June 2013
Assembly and why Flying indoors is a bad idea
This post is part of a series: Designing a Mapping Robot
In the last post I had just gotten all of the parts of my quadrotor delivered, unboxed and tested. By now I also had a rough idea about what each thing did. Now it was just a matter of following the instructions and fitting all the pieces together right? No. Not really.
I won't bore you with the technical aspects of this. It was, hands-down, the most frustrating part of the whole process because nothing worked precisely as advertised. Actually, to be completely correct: I didn't work as advertised and very often what I thought was the right way to do something was completely ass-backwards.
No shame in that though. It's all part of the learning process.
"This thing's an autopilot so I don't need a radio right?" --Me, October, 2012
It seems I vastly overestimated how advanced the Ardupilot software is. Don't get me wrong, it's amazing stuff but it's still not completely plug-n-play. The software is constantly being updated though and even since my first attempts last December it has come along leaps and bounds.
Also I was avoiding a fundamental issue (I'm always doing that). I like to find shortcuts around deep learning. Sometimes it works and sometimes I need to double back and do things the "right" way. In this case I got royally schooled. I would need to get myself a radio and learn how to fly this beast.
Speed controllers are the little computer monsters that tell each motor how fast to spin to keep the craft level and hovering. They are too small to have a screen or any kind of visual display so they tell you they're happy by singing a little song. I completely approve of this.
It took me a good 3 days of trying to get them to do anything useful though. We're so used to visual feedback from devices.
The Turnigy Talon carbon fibre frame was pretty easy to assemble but it had a lot of screws and while lightweight seemed really brittle. This is known as foreshadowing. In the next post all my fears will be confirmed.
Flip the switch aaaaand...... nothing!
Then I had it all together and I turned it on and....... nothing. It's amazing when I look back on this moment. These days I can get the whole thing up and flying in a few minutes; troubleshooting by sight and sound any problems almost immediately. On this day though I was stumped, angry, frustrated and completely clueless. The odd-looking copter-thing in front of me was a black box of magic that I had only peeked inside. The outside of it was blinking, flashing and beeping in ways I couldn't understand.
My apartment has a great feature: an unheated, unfinished work room. I call it the MattCave. Anyway I attached a 4-point harness made of rope to the floor as a way to prevent the copter from flipping over or crashing into the ceiling/walls/me.
I can't even relate the joy when I was able to connect the radio to the base station and see that turning the brain of the copter with my hand affected what was on the screen.
Here it is Lifting off the mat for the first time:
You'll notice this is called "Test Flight 2". That's because test flight 1 ended in tears. I would have shown it but in my panic and excitement I forgot to push record on the camera. Oh well.
Here's another flight. I used my Makerbot Desktop 3D printer to make a camera mount for the Turnigy Talon 2.
So what should I call it? Flying 3D Mapping Robot seems like it could be called F3MUR. Hmm....
- Building a 3D mapping Robot 24 June 2013
- Quadrotor Shopping List 28 June 2013
- Assembly and why Flying indoors is a bad idea 30 June 2013
- First flight. Everything was going so well..... 10 January 2014