Engineer Spotlight: Tanky Drone’s Bogdan Ryabyshchuk Talks Crowdfunding and Racing Drones
We talk with Tanky's founder about nigh-invulnerable 3D-printed armor, how to destroy a GoPro, firmware and batteries, who is the best pilot, and the inevitable power-to-weight tradeoff of flying things.
Drone design is a challenge in many ways. Co-founders of Tanky Drone, Bogdan and Yuki Ryabyshchuk, have developed a camera-guided racing drone the size of your hand. AAC's Jeremy Lee chatted with Bogdan to find out how.
Tanky is a Kickstarter campaign that puts all the power and speed of a racing quadcopter kit into a ready-to-fly package and a high-performance protective shell. You can also tilt its onboard FPV camera while in flight, which is more than just a gimmick—it can change the way you fly. A great deal of thought and refinement has gone into the design, along with "best of breed" components from across the Remote Control community.
We talked with Bogdan, Tanky's founder and chief engineer, about nigh-invulnerable 3D printed armor, how to destroy a GoPro, firmware and batteries, who is the best pilot, and the inevitable power-to-weight tradeoff of flying things.
AAC: Is this your first Kickstarter?
B: Tanky is our first Kickstarter project, but this is our second Kickstarter campaign. We had a Kickstarter campaign for Tanky in July that raised $50k but didn’t get to our total goal. I am happy to say that we have beaten our first campaign in the first 24 hours this time around. We like building and designing things; marketing is not something we are as proficient at, but we are learning fast.
AAC: How's it going?
B: We are really excited and very grateful to all of our backers. This time around it’s going amazingly well! Hopefully we can keep the momentum going and get Tanky out into the world as soon as possible.
AAC: Which of you is the better pilot? Do you have races?
B: We love flying together and racing on occasion. Our favorite thing is to chase each other around a field. As far as who is the best? Chris, our test pilot, is on a whole different level—by far the best in our team. I guess that’s why he is the test pilot!
Bogdan (on the left) demonstrating Tanky
AAC: Your prototypes seem to have printed parts for Tanky's armor. What have you learned about fabricating drone shells?
B: The prototypes are indeed 3D-printed using SLS technology via Shapeways. They are made out of extremely tough nylon. We printed the parts in several colors and finishes and what we found is pretty interesting. By far the toughest material is the white, unpolished nylon.
The unpolished white nylon has properties similar to the DAL props. The plastic does not snap; you have to bend it back and forth over and over again to get it to give. We crashed the unpolished white Tanky into trees, the ground, and even rocks. The plastics have deep scratches on them and lots of grass stains, but they are all still in one piece.
We also printed using the polished material in white and other colors. The material Shapeways uses for their polished prints is still nylon, but I believe they make it out of a slightly harder material or cure it in a different way to make it harder. The result is a stiffer print that is MUCH more brittle. The polished prints can survive a tumble down a grassy field, but that’s about it.
For prototyping reasons, we are sticking with unpolished white prints. For production, we are going to try using a colored polycarbonate blend. We tested some injection-molded parts made out of PC, and they are incredibly tough. It’s the same stuff they use in bulletproof glass!
AAC: You've mentioned in the past that setting the angle of the camera subtly affects how you fly. Can you elaborate?
B: It’s kind of like driving. When you are learning how to drive, they tell you to look ahead at where you want to go and you will automatically go there. It sounds like magic, but then you realize that it really works.
With camera tilt, it is the same thing. Your brain naturally wants your view to be level with the horizon. So, if you set the angle of the camera at 25°, your brain will tend to tilt the whole craft by 25° to compensate. The steeper the angle, the more your brain will compensate by increasing the forward tilt of the craft. This directly translates into your cruising speed, the speed your drone is going at while the horizon is level in your view.
By remotely adjusting the angle of the camera, you can set your cruising speed. If you are just starting out, you can set it really low, maybe 10° or so. As you get better, you can increase it more and more, but the cool thing is, if things get too fast for you, you can easily tilt it down and slow down.
What I like to do is set it to 45° and fly full speed for the first 2-3min or so. When the battery starts sagging badly I turn the camera down to maybe 15-20° and go on a nice and slow cruise around the park for another 2-3 minutes, really getting everything out of that battery pack. It’s fun to fly fast, but it’s also really enjoyable to take it easy and explore a bit.
Yuki using the remote control system
AAC: Tell us about 3D flying (flying upside down). Can Tanky do it and are its motors reversible?
B: We haven’t tried it because, quite honestly, we are very skeptical of non-collective pitch 3D flying. However, both the ESCs and the FC will support it, and the camera doesn’t much care if it’s upside down, so it will work just fine. You will need to run 3D props, of course.
3D flying is really cool, but it’s really something best done with a CP quadcopter, like the Stingray, or a CP helicopter. On a normal drone, you are putting a tremendous amount of load on the motors when you ask them to reverse mid spin, and the lag between input and reaction at the zero point is just really frustrating to anyone who has ever flown a proper CP Heli. That said, Tanky could be set up to do it.
AAC: Do you need to fly more delicately when Tanky is wearing the GoPro Hat? Are there any plans to extend the 'protective armor' to also contain the action cam?
B: Those GoPros are crazy tough! In most cases, what kills your GoPro isn’t cracking the case but the internal g-forces damaging the image sensor or other electronics. The way we protect against that is by mounting the GoPro on rubber isolated mounts. In a hard enough crash, the rubber mounts will rip out and the GoPro will disconnect and tumble on its own. To keep it a little safer, we recommend getting soft rubber TPU bumpers for the GoPro. They are light and really absorb the inertia of hard impacts well.
AAC: Is it possible to sacrifice raw speed for extra flight time by using a bigger battery? Or do the thrust/weight/power ratios end up about the same?
B: We tested different battery sizes and 1300mAh 4S strikes the perfect balance. You get a little more flight time with 1500mAh, but the improvement is negligible. For racing, where you only need to finish a few laps and your in-the-air time will be under 2 min, a 1000mAh pack works really well. It gives you just a bit more speed at the expense of flight time. But the differences are tiny. It’s like you point out, with a bigger battery you have more weight to carry around, so a lot of the extra battery capacity goes to carrying around the battery itself.
The most striking example of this is the very first drone I built, a 300mm tricopter. On that setup I was getting almost the same hover times with a 2200mAh and 4000mAh 3S packs. All of the extra juice was going to keep that 4000mAh battery flying.
Regular props do give you a bit extra flight time by comparison to bullnose props. The props we are using are the DAL V2 props which are a nice blend between a chopped bullnose prop and a regular prop. They have just enough taper to keep the turbulence down, but still have a lot of blade area to really move some air.
AAC: Is the Integrated FrSky module D or X series compatible? Are there any extra channels available from the receiver for extra servos or fun hacks?
B: We are still working with FrSky on the receiver, but the final version will be X series-compatible, connected via S-Bus, with full telemetry available from the flight controller. So your battery, RSSI, accelerometer X, Y, and Z axis, and so on will be available. All of that stuff is also going to be available via the onboard OSD.
The receiver will not have a separate extra PWM output, but we may (space permitting) leave a few PWM pads free on the flight controller, so you may be able to use those to control extra servos and such. Tanky is a tiny platform, though, so I am not sure what could be done with it.
AAC: Why did you pick the TBS transmitter from all the available options?
B: We wanted a transmitter with integrated FPV screen and that was the only option available. But that’s selling the Tango a little short. Even if you forget about the integrated FPV system, it’s a wonderful transmitter with some of the best gimbals on a transmitter in this price range. It was very important for us to include in the RTF kit a high-quality transmitter, something that people would love using with their other drones as well.
The reason we wanted an integrated video system was because we do not want to include goggles. Choosing the right goggles is just such a personal thing. I, for instance, cannot fly with FatShark goggles because I have extremely wide interpupillary distance (73.5mm). Many other people have other issues, like prescription glasses, stigmatism, etc.
We included a controller with a screen so you can start flying right away, and then you can shop around and see which goggles work best for you.
For me, the answer was the HeadPlay goggles with a custom RHO lens, though I am also making a custom set of goggles for myself right now using FatShark Base screens and ski goggles frame.
AAC: The protective Motor Shroud is interesting, but I'd be worried about debris getting in and air flow through the motor. Have you had an issue with either of those?
B: We never had any issues with motor bell rubbing against the shroud or with debris getting caught in that space, and we sure crash a lot. So, if that was a possible issue, it would have happened by now!
We did have an issue with cooling and airflow in our very first design. Our first design had openings cut around the motor bell, but none under the motor. This trapped air under the motor, reducing airflow through the coils. In V1.2 of the plastics, we solved the issue by removing the side openings and creating undercuts to drain away the air from the bottom of the motor. Those undercuts perfectly lineup with the outlets on our motors, so the air has a clean an uninterrupted flow through the motor.
The main benefit of the plastics around the motor bell is shock absorption. In an impact with a hard object, like a rock or a curb, the plastic deflects and compresses, dissipating a ton of energy in the process. The plastic may deflect enough to actually hit the motor bell, but it protects the motor from sharp and hard impacts, impacts that could cause one of the magnets to dislodge or the bell to bend.
AAC: You're running Betaflight for the Firmware rather than alternatives like Cleanflight or Baseflight. What were the reasons for that decision?
B: We are working with the developers from Sirin FPV for our electronics. They are major contributors to the Betaflight project and are developing some really amazing hardware that we are going to be also using in Tanky. I can’t talk much about what they are cooking up next because they haven’t announced it yet, but it’s going to be pretty cool. All of their OSD code is based on Betaflight.
In general, Betaflight has really improved in the recent months. So much so that “Beta” is hardly an appropriate prefix nowadays. It’s a really solid and impeccably organized codebase. That’s not to say anything bad about Cleanflight or Raceflight, and I think some of the OSD features in Betaflight will be ported to those firmwares as well. So with Tanky you’ll have a choice of the firmware you want to run, but out of the box we will be running Betaflight.
AAC: Are all failsafes set up out of the box?
B: Yes, the failsafe will be setup via the Betaflight failsafe feature. If the receiver loses connection, Tanky will level out and throttle down to 10% power, and fall to the ground in a relatively controlled manner.
AAC: What's the most popular aesthetic so far in terms of colors and engraving?
B: We have had a few people choose the custom engraving option. We wish it was a little less complicated with Kickstarter though. Kickstarter does not allow you to combine rewards, so people just have to figure out what package they like, and then add 25$ on top for the engraving.
We haven’t picked the colors we are going to produce yet. We are going to do a backer survey at the end of the campaign ask people what colors they would like to see, then pick the three most popular ones and go from there.
If Tanky is super successful, and we have 500+ backers, we will be able to produce a greater variety of colors. It’s hard to do it with small batch orders because injection mold manufacturers do not want to setup the machine multiple times to accommodate many different colors.