This was one of those projects that was entirely for fun but had quite a bit of engineering behind it to pull it off. Unfortunately, I can't provide a comprehensive parts list or tutorial for a build this big, but I can provide some free 3D print files to print your own 1:10 scale model.
Motivation
I've been lucky enough to make it out to Flite Fest, put on by the team at Flite Test, for the past two years. Flite Fest is a festival put on every year dedicated to all things electric RC: foams planes, drones, & helicopters. One of the main attractions at Flite Fest are their daily combat sessions, where hundreds of foam airplanes take to the skies at the same time to battle it out and be the last airplane standing. It is truly as crazy as it sounds.
The nature of Flite Fest combat has inspired some rather unique approaches to taking down other aerial opponents, which is surprisingly harder than you'd think. Some pilots build the largest planes they can (and inevitably become a prime target), while others prefer to attach streamers and fishing line to smaller builds to snag other airplanes' propellers. Regardless of approach, you don't typically see any helicopters participating in combat, despite them being excellent candidates for maximum destruction in the form of an aerial blender.
With Flite Fest's 10th anniversary coming up this year, I wanted to show up with something special. My plan was to bring a giant tip-prop helicopter to swat airplanes out of the sky.
Why this particular design? Well, conventional helicopter designs are quite mechanically complex, and thus pricey and easy to break: two things that don't make them very good candidates for combat. A tip-prop helicopter allows for some unique design liberties that make its construction much more similar to a foam airplane, and control much more similar to a drone. Plus, I already had quite a bit of experience with this particular design on my much smaller spinning drone project.
Airframe
The airframe consists of hardware store wood and 1" pink insulation foam. All in all, the raw material cost was somewhere in the ballpark of $300 for the airframe alone. The whole thing was sized to be as big as possible, while still being able to be broken down and fit in my car for transport. This ended up allowing for a helicopter with a 15-foot rotor diameter--probably in world record territory, if this type of helicopter design counts.
I used a very simplified airfoil design for the project both for ease of construction, and to thicken the leading edge of the blades for more strength and durability. It's what's known as a Kline-Fogleman airfoil, identifiable by the flat plate construction and distinctive "step" at about the quarter-chord. Efficiency wasn't the goal of this project, so I could get away with this far less efficient design in the name of speeding up the build process.
The center hub is quite literally a 1/4" plywood box with the sole purpose of housing the batteries and electronics. Some pine 1x3s serve as the blade grips, and some internal 3D prints and drywall screws help tie all of the loads from the blades together.
De-Spinning the Top
It wouldn't be a helicopter without some sort of non-spinning fuselage. For my smaller spinning drone project, I didn't bother to de-spin anything, but rather used some pulsing LEDs to indicate what the "front" was. For this project, I wanted to tackle that problem directly by de-spinning a giant red arrow so I could fly it around from the ground and have some notion of direction during the day.
Turns out, you only need two sensors to pull this off: a gyro and an encoder. The gyro measures the rotation rate of the spinning rotor and is already present on the flight controller used for this project. That just leaves the encoder to be integrated, which is a simple analog or i2c connection into the flight controller providing accurate relative angle measurement between the spinning and non-spinning reference frames. The encoder I used uses a diametric magnet and hall effect sensor to provide this measurement.
By ensuring that the relative rotation rate of the encoder measurement is equal and opposite to the measured rotation rate of the rotor via the gyro, you can perfectly de-spin the top platform, subject to whatever small amount of drift/noise your measurement signals have. To correct for this slight drift over time, you can map the pilot's yaw stick input to the rotation rate difference between the two reference frames, allowing for manual corrections and yaw control.
The code to achieve this is quite simple. Assuming you've already integrated your encoder measurement into dRehmFlight VTOL, you can use a proportional-integral controller to generate a stabilized signal for your de-spun platform's actuator. In my case, I used a brushed gearmotor and PWM speed controller, so it was just a matter of assigning this stabilized variable out to the associated PWM pin in the flight controller's control mixer.
float error = heading_rate_des_dps - (encoder_rate_dps + GyroZ);
float kp = 0.0002;
float ki = 0.0025;
integral = constrain(integral + error*dt, -0.5/ki, 0.5/ki);
spinny_PID = kp*error + ki*integral;Noise Troubles
The brushed gear motors in the center hub put out quite a lot of electro-magnetic interference due to their brush contacts creating small sparks. If you've ever seen sparks on your power drill's motor when you abruptly change speeds, you know what I'm talking about. This interfered with the main propulsion motor signal lines causing them to stutter whenever the brushed gearmotors were running. A common trick with brushed motors is to use some 104 ceramic capacitors soldered to the power terminals and motor casing to filter out this noise and clean things up.
Another noise problem that didn't present itself until my first flight attempt at Flite Fest was an issue of interference between the high power (25V, 135A+) wires out to each blade, and the propulsion motor signal lines. Turns out, running them down the blade right next to each other wasn't the best idea... This led to the motors stuttering above about 50% throttle, limiting my total thrust and keeping me on the ground. I killed a day at Flite Fest re-routing these wires to be further apart from each other with some aluminum tape in between. Lesson learned for future high-power projects, I guess.
Flights, Wind, & Destruction
The real reason I went through all this trouble was to see a giant 15ft helicopter slicing through airplanes in Flite Fest combat. The first flight went great! But due to the ~10mph wind, I couldn't keep it from drifting further and further away. We chopped up a few small airplanes at a distance, but the resulting single broken prop left much to be desired in the carnage department.
For the second flight, the wind picked up a bit more and we had the same problem with drifting before any airplanes even got into the air. I did have the means to roll/pitch the vehicle via pulsing motor control, but the very nature of this beast and the ~12mph wind gusts this time around made it impossible to keep stationary.
You can loosely approximate the exit airspeed of the rotor's generated thrust with this equation:
Vehicle Weight = 1/2 * Air Density * (Thrust Speed)^2 * Rotor Area
Solving for Thrust's Airspeed....
Thrust Speed = sqrt(2 * Vehicle Weight / (Air Density * Rotor Area))The whole thing weighed about 40lbs, so when plugging this in with the 15ft diameter rotor's area (make sure you are using consistent units when plugging things in!), the exit thrust this vehicle generates to hover is only around ~9mph . Meaning that even if I could fully pitch it over 90 degrees into the wind, it wasn't generating fast enough airflow to counter the wind speed! It was just too "floaty" to have any chance of station keeping in any wind. This is the pitfall of this type of extremely low disk-loading helicopter/drone design: you can't fly them in any appreciable wind.
Even though the wind wasn't cooperating on our last day at Flite Fest, we still took it up for one last flight with the hopes of totaling it before it drifted out of reach of the smaller airplanes. Thank goodness we did, because it certainly lived up to its name of "The Quad-I-Chopter," and the show was spectacular:
Conclusions
I hope you enjoyed this ridiculous project and maybe even learned something new about aerial robotics. As a thank you for checking this project out, I wanted to share some 3D print files so you can print out your own 1:10 scale version. The download link is available below.
And by the way--Check out Flite Test's video on this project!
Download the free .stl files to 3D print your own 1:10 scale model:
Quantities are listed in the part name.
Comments