I’ve been playing around with the idea of a foam board UAV for so long now that it was almost a relief when I finally settled on a design. But that was just the start of the challenges…

The Versa Drone Project

There’s nothing new about a home-built drone. There’s nothing new about a foam board wing. There’s not even any real magic to putting an APM into an airframe and handing over some or total control of the flight to the computer. So why have I spent so much time on this project?

Several times over the past few months I’ve seen articles come up on my Twitter feed about 3D printing and how some bright sparks have been printing airframes using them.

Image courtesy of dronemakers.com

In the first a team from a UK university had built and flown (as a glider) a wing that was entirely 3D printed. It looked awesome and sported some impressive aerodynamic theory. They then spoiled this a bit by going on to say how they were now trying to work out how to fit an EDF in it. Errr… probably not going to be so cool when it runs out of battery at the first waypoint! The article heralded the design as the “first disposable drone”.

In the second a team from Spain had printed a wing for a Bixler or something similar and posted a video of it flying around beautifully. The article went on to describe how by being able to print this wing in only 38 hours it would now be possible to print disposable drones on demand in remote locations as and when needed.

Hang on… 38 HOURS?!?!

I got to thinking: In 38 hours, and with the amount of money it costs to buy just the filament for a 3D printer, I could build a foam board wing, fly it, crash it and build another 2 to replace it!

…and so the Versa Drone was born!

I had spent so long playing around with different designs for a foam board UAV that I had lost site of what I had in front of me. My Blunt Nose Versa Wing variants all sported the ideal qualities for a UAV/RPAS design.

• Manageable wingspan
• Plenty of wing area for a light wing loading
• Stable and docile flight characteristics
• Lots of room in that extended center section

I immediately pulled out the plans and started making adjustments to ensure I could fit and access an APM flight controller, telemetry radio and associated electronics. The idea was to set the aircraft up in much the same way as my FPV ship but this time to have the GoPro (or other camera) able to face downwards so it could shoot video and still images for mapping/surveillance purposes.

Building the Versa Drone

I’m not going to go into great detail here as if you know what a Versa Wing is then you know there are plenty of great build videos and tutorials out there. I will briefly detail however some of the modifications I made:

“Bunny Ears”

The bunny ears are the extended winglets that first featured on my Blunt Nose Versa Wing FPV ship. The original intention was to improve lateral stability by increasing the vertical surface area. This does seem to have worked to a certain extent although continued testing has revealed that they actually make the model more susceptible to turbulence on windy days. They also do make the model a little less practical for transport.

Elevons

This was an idea lifted directly from models like the RiteWing ZephyrII. By making the ailerons larger than a standard Versa Wing and making the outboard portion deeper it moves the center of pressure towards the wingtip. This improves handling at low speed at the expense of increased torque on the servo at higher speed and deflections. In many hours of flying though I have never seen one of these elevons fail.

Underbody Protection

Although designed to be cheap and disposable, I still wanted this airframe to have a modicum of resilience to the rigors of sometimes rough landing areas. To this end I covered the underside of the center portion of the model with 3mm “corflute” plastic. I then further improved this by adding shallow skids to the margins of the center section and also about midway out along the wings to protect the slight protrusion of the aileron horns. Besides protecting the underside of the wing these skids also protect the valuable downward-facing camera lens from landing scratches.

Gear Mounting

Not really a “modification” but different from a normal Versa Wing. The HobbyKing “HK Pilot” APM is mounted centrally on the wing above the balance point. To give me as much depth as possible I actually cut down to the corflute base and put small “walls” around the hole giving a snug and solid mounting for the flight controller. Whilst I had intended to put the telemetry radio out in the wing a change in cabling meant that for this first version it sits beside the controller in the center section. The GPS is simply mounted on the outside of the wing for maximum reception.

Setting up the HK Pilot APM

I need to be clear that I am not using a 3DRobotics APM or Pixhawk in the Versa Drone. This is, like everything else I use, sourced from HobbyKing and in this case I am using the “HK Pilot” controller. That said, it uses the Open Source APM 2.5.2 architecture, Ardupilot firmware, I used the Ardupilot instructions to set it up and I’m running the whole show through Michael Oborne’s Mission Planner software – so basically anything I say here would serve just as well for a genuine 3DRobotics APM board.

I’d never set one of these beauties up before so it was a bit hit and miss at the start. There are some fantastic online instructions at the Ardupilot website (http://plane.ardupilot.com/) so I’m not going to replay a blow by blow account of how it’s done. I did however learn some lessons that might save you a bit of time should you try it yourself…

• Don’t fry the board! – Later versions of the Ardupilot firmware highlighted a problem with some boards where the 3.3v regulator that supplied power to the sensors had blown. This was happening on both genuine 3DR boards and the HK Pilot and was caused by trying to power the board, GPS unit and receiver using only the USB cable whilst setting it up. When you go to plug your board into the computer either unplug the accessories or provide a separate power supply. I did this damage to my first APM board and it is possible to fix it with some fiddly soldering and a new 3.3v regulator.
• Don’t use the wizard to set up your APM in Mission Planner – A strange thing to say but the truth is the wizard doesn’t give you a break. If you use the wizard you are left trying to complete the live compass calibration and accelerometer calibration with your 1200mm wing attached to a computer by a USB cable. You have to use the USB cable to load the firmware but everything else can be done with telemetry. You can do one of two things easily:
  • Load the firmware and calibrate everything before you put the APM into the airframe.
  • Load up the firmware, then switch to a telemetry link to complete the calibration.
• TIP – if your compass appears to be broken. Take all the magnets out of your hatch. It is amazing what a massive effect a small magnet in the wrong place can have on your navigation systems!

Flight tuning the Versa Drone

I already knew that the airframe would fly well from my FPV ship so the only “maiden jitters” came from knowing I had a lot of cash sat in that center section should I crash or lose the model. Fortunately a good throw had her flying perfectly straight out of the gate and it was into the long and often arduous task of setting up and tuning the flight controller.

Again, I’m not going to go into micro-detail as you can follow the same instruction I did online (link above). I did learn a few tips and tricks along the way though…

• My FPV ship has quite large throws on the elevons but uses massive amounts of expo to tame them. This is no good for a UAV, especially if you are using Autotune to tune it. The controller doesn’t know you are using expo so once it takes over your flight controls can be twitchy at best and dangerous at worst.
• Set the model up to fly like a trainer. You’re not fitting an APM to fly aerobatics so why set the plane up like it?
• Fly and completely trim your model before you do anything with the APM. That includes calibrating the radio on Mission Planner. If at any time you change anything and your trims change you should go back and calibrate your radio again.
• Autotune is great but stick with it. You need to “teach” the APM to fly with at least 20 full stick deflections in all directions – that’s why the high expo and large throws is a bad idea! I was flying for at least 10 minutes to complete my autotune flight and it is still far from perfect.
• Follow the instructions! I didn’t the first time around and the plane was uncontrollable and tried to fly away when I switched it into an automated flight mode.

The first autonomous flight…

After a number of tuning and testing flights I was ready to go for a full GPS waypoint mission. A successful test of the “Return to Launch” flight mode told me the navigation and control systems were working so there was no more excuses.

In the end, the first autonomous flight of the Versa Drone was a bit of an anticlimax! The UAV flew a nice circuit around the field and then switched to loiter mode until I took control again and brought it down for a landing.

Anyway, a few flights later I managed to capture some footage during a test of the DO_JUMP feature (which makes the UAV repeat all or a portion of the mission for a specified number of circuits)…

This was intentionally a one man mission as I wanted to prove that it is possible for one person to handle the flight from start to finish.

I think the tuning could be tighter but the Versa Drone still held the course pretty well, despite the strong wind blowing on the day.

Conclusion

What I have created here is a 1200mm span UAV that is able to carry a small camera for mapping or surveillance missions without any problem. It has a take-off weight of only 1.1kg so will easily fit into the new CASA 2kg weight class and will be able to be operated commercially without any certification or license required (when the new rules come into effect).

Its two 2200mah 3-cell LiPo’s provide enough power for 30-45 mins duration at an average 15 m/s.

All this is achieved for (considerably) less than the price of a new 3DR Pixhawk controller.

38 hours to print a plastic wing? I think the foam board community could teach the world a thing or two about creating a truly “disposable” drone!

Where to next?

Despite the slightly epic nature of this article, this is only really the tip of the iceberg for me and there are huge amounts of information I haven’t covered here.

There is certainly more work to be done improving the platform and tuning the airframe. I want to try different wing sections and maybe even some KFm airfoils for comparison. I also want to bring the size down to make transporting the wing even easier.

Watch this space…