Up in the Air
Several years into his hobby of building unmanned aerial vehicles (UAVs), IEEE Member Anish Mohammed earned some geek cred when he met William Premerlani, who’d written some of the initial open-source software that galvanized the do-it-yourself drone community. “He called me a true UAV addict!” Mohammed recalls, laughing.
Mohammed, an independent management consultant in information and data strategy in Woking, about 50 kilometers southwest of London, has degrees in medicine and information security. He has long maintained a side interest in robotics. Shortly after his 2002 move to the United Kingdom from his native India, he began tinkering with design kits from Vex Robotics and Lego Mindstorm. He went from building remote-controlled airplanes and helicopters to building drones. While traditional remote-controlled aircraft are controlled manually via a wireless network from the ground, drone models are programmed to fly themselves.
BUILDING AND FLYING DRONES
TECHNOLOGY STRATEGY CONSULTANT
“I was looking for ways to reduce the number of crashes and found a software component that programs an autopilot takeover if the craft flies out of preprogrammed boundaries set with GPS,” he says.
He began to collect the individual components for a quadcopter (four-rotor) drone and built one from scratch. That was in 2010. Since then, he has built 10 multicopters—aircraft with more than two rotors—including an octocopter made with eight rotors, as well as two fixed-wing aircraft. The rotors of his largest multicopters are more than a meter in diameter. Mohammed flies multicopters both manually and autonomously.
“I was really interested in all the algorithms that help hold the multicopter in the air and fly and allow fixed-wings to fly along a programmed path,” he says. “There’s a whole bunch of challenging math and theory for doing that.”
Mohammed says he spends 10 to 20 hours each week working on or testing his drones. He flies his copters with fellow enthusiasts in a nearby park.
“It’s the configuration—making it stable—that takes the most time,” he says.
The cost for DIY drones depends on “how geeky you want to get,” he says. Both fixed-wing and copter drones can cost from US $300 to $3000, depending on their accessories. Higher-end models, for instance, might include a video camera that enables the flier on the ground to see where the aircraft is headed in real time.
Mohammed says it’s exhilarating to build something and make it fly: “After all the long hours and wanting to give it a rest, the day I get a drone off the ground and into the air, it’s like, ‘Wow! I got it right!’”
IEEE Member Richard Schwartz has spent a good half-century playing wind instruments: the recorder, which he picked up in college, the oboe after graduating, and the flute since his mid-50s (he’s now 71).
So when he stopped by a favorite music store this year in Santa Monica, Calif., he couldn’t help but notice some musically subpar models of simple flutes. “I thought I could do better than that,” says Schwartz, an electrical engineering consultant since leaving the aerospace industry in the 1990s.
Home from the music store, he scoured the Internet to learn how to make his own plastic flutes, then tried his hand at making them from polyvinyl chloride pipe. PVC is relatively easy to cut, drill, and file, and it’s readily available in hardware stores.
He once made a high-pitched piccolo out of PVC pipe to play in a local parade, but a flute is more difficult, he says, because pitch errors are more obvious in its lower frequency range. After his first dozen attempts foundered because the second-octave notes weren’t on pitch, he decided to make 100 flutes in 3 months from standard half-inch (1.27-centimeter) outer diameter PVC pipes. “That’s what I figured I had to do to teach myself,” he says.
Schwartz uses Flutomat, a computer program that calculates the position and hole diameters. “For most wind instruments, each finger position can produce several notes, depending on the mouth position and blowing pressure,” he notes. “But Flutomat does not calculate harmonic frequencies, and the high notes kept coming out flat. I had to find a way to correct that.”
One problem with flutes of PVC pipe is that they are straight cylinders; actual flutes are barrel-shaped. “I put a half-inch pipe with a slightly smaller internal diameter on the head end, observed the result, and did some calculations to determine how much of the thicker-walled pipe was needed to get the flute perfectly in tune,” he says.
Schwartz can produce a flute in about an hour. “The time-consuming part is tuning them,” he says. “You have to adjust the size and position of each hole to get the right frequency.” The flute also comes cheap: home-repair stores sell PVC pipe in 6-meter lengths for US $5.
As he gradually gives his batch of PVC flutes to friends, he is turning his attention to making PVC variations of piccolos that were popular in blues clubs during the late 1920s and early 1930s. “I’m trying to see how good I can make them,” he says. “There are a lot of design compromises and trade-offs.”
Schwartz has applied his engineering know-how to his hobby.“I figured out that the equations for a sound wave traveling through a pipe are the same as for an electromagnetic wave traveling through transmission lines,” he says. “At first I thought I was brilliant. But after a Google search, I learned that someone discovered that same thing 50 years ago and wrote a paper on it. But the scholarly music community ignored him, and nothing came of it.”
A perfect model would use computational fluid dynamics to figure out how the air flows through the flute, but then, he adds, “the computing really gets ugly.”
For the time being, Schwartz is using a popular antenna and transmission-line finite-element model to solve the sound wave’s differential equations and to study the best models for the features of the flute, such as the head cavity and tone hole.
Photos: Po-Wah Yau; Ketung Hsiao