Remember how I said I was going to make a quadrotor in the beginning of the summer? I worked on it in stealth, slowly acquiring the parts from Hong Kong dealerships (Hobbyking and other eBay sources). I got all my parts in a few weeks ago, and finished the main frame build. I foolishly didn’t take much documentation however, and my design process was pretty pathetic. It was basically just a simple CAD drawing with blocks for what I thought was important. I paid for that with the excessive weight I have packed into the frame of this thing.
With no real precision tools (mainly hand tools and a drill press), I made a fully wooden frame. It’s not even balsa wood for crying out loud. Home Depot apparently only sells dense wood, and then denser wood. So far the quadrotor looks like this:
The arms are just .75″ square pine 24″ long with 3/8″ holes drilled in a futile attempt to reduce weight. The arms have (obviously) no problem with flexing at high throttle being so overbuilt. The motors are mounted on extensions to the arms that are bolted down. At some point I will replace all the heavy steel screws in this project with lighter nylon screws if possible.
This also includes the landing gear, which I actually formed today out of Polymorph by hand. If you don’t know what Polymorph is, its a nylon-like plastic that melts at an extremely low temperature (sub-boiling water works just fine). Once you melt bits of the plastic together it becomes pliable like taffy and can easily stick to itself. It was faster to make than making some wooden structure, although I may replace the landing gear with something more effective as I go on. My handicraft turned out to be irregular, adding to the already unbalanced setup.
The main body frame is made of two pieces of particle board, which I cut out by hand with a circular saw and my friend’s coping saw using the 1:1 scale printout feature in Solidworks as a template, following the same technique as we tried for our casting attempts. Again, that too was far too heavy and overbuilt for this application. The entire setup weights in at around 3lbs, before including the weight of the GoPro Hero HD camera I plan to use for aerial video.
An Arduino Uno is used as the microcontroller for this project, with an accelerometer and a 3-axis gyro as sensors. The accelerometer will be used to maintain level flight, which I am still working on tweaking in.
As shown in the video, I need to tweak the gain values for each motor to account for the offset balance. The quadrotor lifts off at half throttle, surprising for me due to the weight of the thing. I haven’t attached the GoPro to the quad yet, but I have a feeling I’ll also need a lot of dampening to get it to ignore the heavy vibration from the motors. Notice the big bump and its effect on my *intentional* makeshift ablative landing gear. Its cheap zip-tie mountings absorb shock!
The props are mounted, by the way, like this:
Just a cap nut on a normal nut on a tension washer washer washer sandwich on the prop on another normal nut. I’m too impatient to wait two weeks for proper prop mountings so I hope I can just use the tension of the nuts to keep the props from drifting. So far at high throttle it hasn’t been bad, but the clockwise rotating motors tend to screw the propellers down towards the base of the motor drive shaft.
A few moments after I filmed the video I also happened to bump and chip a prop on the computer case you see to the left. All of this happened, by the way, in my brother’s room (bad decision) after I was too uncertain about the wind to try an outdoor flight. A word of advice: find somewhere open and closed off. You go through props very quickly. I failed to record the quad’s first actual flight outside in the wind where I managed to flip the quad over and snap a prop on a clump of dirt/grass. Use the throttle stick conservatively! (unlike me)
Electronics include a 3C lithium polymer battery, separate receiver Velcro pack, and excessive high gauge wire. To keep dirt out and to protect from potential crash damage, I installed an acrylic shield over the micro-controller. It’s pretty useless (being exposed from all other sides), but it gives a nice platform to install the video transmitter when I move to first person view.
The GoPro mounting is actually made from the case the GoPro came in. I just sawed off the sides and bolted it down to the battery plate. Since I haven’t ordered any Sorbithane for the SPIRIT MK2 shock dampeners yet, I attempted to use some neoprene in between the mountings to reduce shock. We’ll see how that goes.
I’ll be further tuning this quad throughout this week, and I’m hurrying to get my other projects finished before I have to go back to school. The 30lb combat robot probably won’t be made this year, and the walking robot may take some time before I get it battling. I’m primarily focused on SPIRIT right now, and getting my 3-axis micro-CNC going. I’m doing research at UCI Medical Center in the Department of Urology as well. Though I’m not particularly interested in urology as a field, the department focuses a lot on robotic surgery, and they have a Da Vinci surgical robot system! I have plans to make a small-scale minimally invasive surgery robot myself this coming semester, though I’ll have to learn some reverse kinematics and such. Maybe I can even find a way to mount it on the new modular SPIRIT system as a remote combat medic system. Further details will come as I make them up from thin air.