One of the current trends among micro-quadcopter enthusiasts involves modifications that purposely keep their flying machines at ground level. This floor-hugging tweak is called the Tiny Whoov. It is a micro RC hovercraft built around the uber-popular Blade Inductrix quadcopter.
There are several different ways to get a Tiny Whoov of your own. A cursory web search reveals many step-by-step tutorials, a few manufactured conversion kits, and even an off-the-shelf hovercraft from Blade, the Inductrix Switch. All of the links I found are based on the Inductrix (or one of the many clones).
I wanted to build a Tiny Whoov, but I still enjoying flying my Inductrix in stock form. So I wasn't keen on clipping its wings. Undeterred, I decided build my own variation on the quadcopter-to-hovercraft theme using a different micro-quad. I improvised a simple design while taking copious inspiration from the Tiny Whoov.
The Tiny Whoov uses only the front two rotors as lift fans for the hovercraft. The rear rotors are used to propel and steer the vehicle via differential thrust. With this setup, the same control and gyro settings that work as a quadrotor will also work in hovercraft mode…sort of. More on that later.
For the quadcopter, I used my venerable Heli-Max 1SQ. You may recognize it from my recent Christmas tree drone. Although I didn't know it at first, the modular construction of the 1SQ made it ideal for this project. But don't worry if there isn't a 1SQ on your shelf. The basic components of most mini-quads are the same. I suspect that you can crank out a similar hovercraft with whatever quad you have on hand.
There is nothing fancy about the design or materials that I used. Most of the hovercraft hull was constructed with cheap foamboard and hot glue. One sheet of foamboard from the dollar store is adequate to build several of these things.
Design On The Fly
I'm sure there is much written about the design and optimization of hovercraft. I ignored pretty much all of it. With the Tiny Whoov as my muse, I sketched out a hovercraft hull sized for the dimensions of the 1SQ. My arrowhead-shaped structure measures 8" (203mm) long, 8" (203mm) wide, and 1.25" (32mm) tall. In hindsight, I probably could have made it a bit smaller in every dimension.
Each arm of the 1SQ is held in place with a single tiny screw. I removed the relevant fasteners and motor connections, then carefully detached only the rear arms. The remainder of the 1SQ sits on top of the hovercraft hull. I had to cut a relief hole to fit the stock battery compartment on the belly of the 1SQ. A few dabs of hot glue secure the quad frame to the hull.
I made small foamboard gussets to mount the rear quad arms to the hull at a 45-degree upward angle. The arms are placed as far back as the motor wires would allow. Hot glue would probably work fine here as well, but I used 5-minute epoxy just to make sure the spinning parts would not break loose and wreak havoc.
Air from the front fans enters the hull through circular holes located just below the motors. I'm sure I could improve efficiency with nicely shaped propeller shrouds, but it works fine as-is. The air flows out of the hull through slits around the perimeter of the bottom side. This exiting air creates the air cushion that allows vehicle to "hover".
Initial tests indicated that the foamboard-only hull worked well enough to elevate the vehicle. It performed much better, however, when I added a skirt to help trap the air under the hull. I made the skirt from 2mm-thick EVA foam and attached it to the hull with Foam-Cure glue. The skirt protrudes below the bottom of the hull .25" (6.4mm). To make the skirt more flexible, I used scissors to add a fringe to the exposed area.
I originally placed the 1-cell 260mAh LiHV battery just forward of the 1SQ's control board. This location made the hovercraft nose-heavy, as evidenced by a slight nose-down stance. I used scrap foamboard to create a battery mount above the 1SQ control board. The battery is secured with Velcro.
Testing and Tweaking
My first step after completing the hovercraft was to lash the hovercraft to my workbench and perform a control system check. I noticed that the yaw control was backwards. Although that seemed a little strange, I didn't think much of it. The controls appeared to work fine after I reversed the yaw channel on my Tactic TTX850 transmitter.
Things got weird again when I set the hovercraft on the floor. I slowly added power to the fans and noticed that the hull quickly floated on the air cushion and began translating forward. That was all good. The problems came when I tried to turn the hovercraft. Whenever I applied the tiniest yaw inputs, the vehicle would immediately begin spinning as if I had jammed in full yaw control. This happened consistently, no matter which direction I tried to turn.
It took me several minutes to figure out the problem. The root cause had to do with a seemingly insignificant trait of the 1SQ. Quadcopters are configured with two rotors on diagonally opposed corners that spin clockwise (CW) and two on the remaining corners that rotate counter-clockwise (CCW). I'm not sure how this standard was set, but most quads have CW rotors (as viewed from above) on the left-front and right-rear corners. That leaves the right-front and left-rear locations for the CCW rotors.
In a stable hover, the torque from one pair of rotors is cancelled by the torque of the opposing pair. Yaw control is achieved by managing the relative RPM of these pairs. For instance, a command input to yaw the aircraft counter clockwise causes the left-front and right-rear (CW) motors to speed up while the right-front and left-rear (CCW) motors are slowed. The resulting torque imbalance causes the desired CCW yaw movement.
When converted to a hovercraft, a yaw command causes the same RPM shifts. Yet, the differential thrust of the vertically-oriented motors in the rear has much more control authority than the torque delta. That works fine in most cases because we would still want the right rear motor to speed up for a left (CCW) vehicle turn. The same control input results in the same vehicle movement.
The 1SQ, however, has its motors configured in reverse to the standard. This is of no consequence when operating the 1SQ in its intended role as an aircraft. But converting it to a hovercraft caused the vertical rotors to operate opposite the intended manner. A left (CCW) yaw input actually made the left-rear motor speed up and the hovercraft turned to the right.
While reversing the yaw channel in my transmitter seemingly corrected the swapped command inputs, this change was not recognized by the onboard gyros. The gyros are hard-coded to speed up the left-rear motor for a CCW turn. Whenever the gyros sensed CW movement of the hovercraft, it attempted to counteract this motion with its hard-coded CCW commands. This actually resulted in more CW rotation. The divergent yaw behavior I was experiencing suddenly made sense.
Correcting this issue was as simple as "un-reversing" yaw control in my transmitter and swapping the rear motor outputs on the 1SQ control board. The right-rear motor is connected to the left-rear output of the board and vice-versa. Thankfully, all of the motor connections use tiny plugs rather than soldered joints. So the swap was really simple to do.
Driving the Hovercraft
After correcting my initial yaw problems, I found that the hovercraft is surprisingly easy to drive. It took me a little while to adjust to the controls. But I was soon piloting the hovercraft from room to room with no problems.
The gyros do a good job of keeping the nose pointed where I want. Those vertical motors also have plenty of authority for snappy turns. I expected the hull to slide through turns like a drift car, but that is not the case. The rear end stays planted almost as if it has tires.
I typically only use the left control stick (throttle and yaw) for driving. Adding power has the combined effect of lifting the hull onto the air cushion and propelling it forward. I'll jockey the right stick every now and then to see what effect the pitch and roll controls might have to get the hovercraft out of a corner or over a fold in the rug. Sometimes it helps, but usually not.
I've only driven the hovercraft indoors thus far. This thing can hit pretty impressive speeds across the kitchen floor. I just have to make sure I leave enough room to slow down before hitting the wall (or table, or my foot, or…)! The foam skirt makes a reasonable brake when I pull back on the throttle.
While the hovercraft will travel over short carpet and rugs, it does so slowly. I usually run it only on my tile and wood floors. It can clear door thresholds, rug edges, and other obstacles as long as there isn't an edge more than about .25" (6.4mm) high.
Overall, this hovercraft has been a fun winter project. Constructing the hull was cheap and quick. Plus, I got to exercise a few extra brain cells while sorting out the yaw issues. The completed vehicle is a zippy little indoor toy that I can take off the shelf whenever I need a quick sanity break. Interestingly, the foam skirt also works great for sweeping up dog hair. But I don't think I'm ready to consider this a cleaning tool just yet.
If you've built a Tiny Whoov or a DIY quad-to-hovercraft project, please share your experience in the comments section.
Terry is a freelance writer living in Buffalo, NY. Visit his website at TerryDunn.org and follow him on Twitter and Facebook. You can also hear Terry talk about RC hobbies as one of the hosts of the RC Roundtable podcast.