Create an Inexpensive Flying Wing / Drone (2019) | Hacker News Reader

Create an Inexpensive Flying Wing / Drone (2019)

robotshop.com

137 pointsghgr51 comments

People in the model airplane world have been building precisely these kinds of planes for probably at least two decades. In that sense, nothing new. That said, good to see the idea reaching a different audience.

If you want to save some weight you can buy small diameter carbon fiber rods for the pushrods (rather than using steel). You can also stiffen the wing by using a carbon fiber spar (either epoxy or cyanoacrylate to the bottom surface. If you know what you are doing you can even lighten it by cutting holes in the wings (think: punching out a bunch of circles) and then use packing tape to cover top and bottom surfaces.

That said, a light plane isn't always better. Some of my best-flying RC planes are not what anyone would characterize as light. For example, an electric-powered glider that weighs eight pounds (3.6 kg), can go straight up like a rocket, glide (power off) at over 120 mph and thermal with the birds with great agility and control. The problem with very light RC planes is that they can be very difficult to fly in windy conditions, this is particularly true of inexperienced pilots.

Here are some examples (no affiliation):

https://www.wmparkflyers.com/EPP-Foam-Airplanes

The point here appears to be very inexpensive and requiring very little skill to build. Basically only the electric equipment and the aileron hooks are not made from common cheap household items or flat TV packaging trash.

Certainly you can improve this design in many ways, while making it lighter, sturdier, and higher-performing at the same time. It would take quite a bit of skill, more tools, and a variety of more interesting materials.

"Flite Test" have been designing ridiculously cheap, low-skill DIY flying things for many years now - at the same level of "turning trash to an RC plane." Definitely worth checking out if you're interested in the space.

Agreed about Flite Test. They have kits using water resistant foam board that will perform and look better. Also this delta wing has options like being able to purchase the motor kit needed:

https://store.flitetest.com/ft-spear-wing-wr-1041mm/

It is a good way to get started.

Sure, that's why people have been doing this design for 20 years. Homebuilding included ;)

I agree with the grandparent, it's of course interesting and it's a different audience, a lot of the "drone" crowd nowadays is purely quadrocopters, but like, this was pretty standard stuff back 15 years or so ago, for the electric parkflyer crowd.

But much like every piece of software nowadays is an "app"... every model airplane is a "drone" now, and drone people aren't really familiar with the variety of aircraft that are possible because of the dominance of the quadrocopter. Like yeah, welcome to what the rest of the hobby has been doing while the drone community has farted around with quads for 15 years ;)

My grandfather built model planes out of balsa wood kits for at least two decades. Some of them had wingspans of 5 feet. :)

It has become so, so much easier now.

Thanks for sharing your knowledge!

> You can also stiffen the wing by using a carbon fiber spar

this isn't really optional imo, this is a "should" and it should be at least one continuous spar going across the majority of the wing (don't put two ending in the middle, that's missing the point here). If you want, you can even do two - a shorter one in front and a longer one in back, or do a triangle-shape to strengthen the wingtips.

This is because EPP/EPS doesn't have enough structural strength on its own and with brushless motors you can definitely go fast enough to cause the wing to "flap", and that will result in the elevons flexing in the opposite direction (as the servo pushrod is attached to the wing) causing a "vertical fishtailing" effect in midair.

https://www.rcgroups.com/forums/showthread.php?1500999-The-d...

https://www.youtube.com/watch?v=yYfT0tQm668

(you can see this design is problematic because of the high sweep of the wings... there is frequently a little bit of wingtip that you can't reinforce, but, most of that wing is just hanging in midair, and those types are VERY prone to flapping.)

You can dremel out a spar channel and put the spar in that. The filler material of choice back in the day was Elmers' ProBond Polyurethene glue - mix it with some water and it'll foam up like crazy and you just pour it into the channel and wipe off any excess as it dries. Not to be confused with Elmers' ProBond wood glue though - that won't foam like that.

https://www.homedepot.com/p/Elmer-s-ProBond-4-oz-Advanced-Mu...

Also don't use EPS. EPP is far far more durable. And DO use a shrink-down covering like monokote or ultracote - the shrink-down covering will provide additional rigidity in addition to the spar. I tried using packing tape at one point instead, and it was definitely noticeably less rigid than ultracote and I ending up removing and replacing it.

https://www.horizonhobby.com/airplanes/supplies-and-bench-ac...

As a practical matter, I highly recommend contrasting color schemes - like red on bottom, blue with a wide reflective stripe on top, or something like that. Because from the ground it's super easy to lose perspective if you can't distinguish which side is up.

I have used carbon fiber tape in place of rods to good effect. I agree that a proper spar is critical for handling the loads. To be more precise, you need to deal with shear and torsional loads. These super-light cheap planes will easily turn into a taco in flight without care given to forces that will crush the foam. Even a layer of packing tape on top and bottom surfaces is better than just foamboard. I've had a couple of these planes over the years. I don't really enjoy them. I'm into high performance gliders (e-powered), ducted fan jets or helicopters.

> I'm into high performance gliders (e-powered)

that's neat, what is the sailplane community up to these days?

It really depends on where you look. Tastes in the US are different from what's popular in Europe. It also depends on the audience. As an engineer I tend to levitate (fly?) towards the more technically sophisticated and challenging aircraft.

For example, I have been building and flying what used to be known as "F5B" aircraft. Back when we were using NiCd batteries I had planes that used 27 cells to power very compact multi-horsepower motors in these gliders that often are not any larger than about 2.4 meters in span. Back then (I am talking '80's '90's) these planes could climb out of sight in 5 seconds. They are, aerodynamically speaking, highly efficient. After the climb the competition consisted in completing as many laps of a circuit as possible without power, gliding at high speed. Once done, you get to use the motor again, climb to altitude and thermal (in some cases they don't to this). Finally, there's a requirement for precision landing --the closest to land to the center of a target the more points you earn.

Aside from that, I am very much into what some call "full house gliders", this to mean having a full set of control surfaces: two ailerons, two flaps, sometimes two spoilers, rudder and elevator. These are not typically high speed monsters. These days you use electric motors for the climb. I still have a winch (which was the traditional climb-to-altitude method) but have not used it in decades. With a full house glider the objective is to find and exploit thermals and, if in competition, usually have a precision landing just as explained above. I have had the experience of successfully thermalling for over two hours laying on my back listening to music, landing only due to just being tired more so than anything else.

There are a lot of great options these days to get into the hobby. You don't have to have a full-house glider to experience thermalling like a bird. One can get started with modest kits that are pretty much ready to fly. The more technically minded will eventually want to learn how to build from scratch or seriously modify aircraft (or both). While I haven't done it in a while, I used to regularly make custom planes using a range of materials, from balsa wood for normal performance planes to designing molds to vacuum-bag fiberglass, kevlar or carbon fiber designs.

Here are a few examples of the above:

F5B (I have several planes that look and perform just like these). The motor is only used to climb.

https://www.youtube.com/watch?v=T0z9iy2SJA0

https://www.youtube.com/watch?v=59V9ICM75co

https://www.youtube.com/watch?v=d4J8VQ4iNwQ

https://www.youtube.com/watch?v=MNydX3aOTr8

This type of glider/event seems to be far more popular in the US:

https://www.youtube.com/watch?v=uoZ78c-tIDY

Adding to that, here's slope soaring, which is really cool. I haven't really done a lot of this due to having to travel far to have access to this type of geography. Beautiful silent flight, no motors required. They do add motors to some of the planes just so you have backup in case you lose lift. Here's an excellent intro to the sport:

https://www.youtube.com/watch?v=Ev5fSVi1PrA

And then there's the absolutely crazy/demented --in a good way-- dynamic slope soaring. I have yet to experience this. This is the domain of sophisticated carbon fiber and top-quality components. You can find videos of planes exploding in flight due to loads.

This is a long video with a detailed technical explanation of dynamic soaring:

https://www.youtube.com/watch?v=nv7-YM4wno8

548 mph (882 kph) run:

https://www.youtube.com/watch?v=0nyYaL0dGAA

This reminds me of a kickstarter project for motorizing a folded paper plane: https://www.kickstarter.com/projects/393053146/powerup-40-sm...

Still available: https://www.poweruptoys.com/products/powerup-4-0

At the expensive end, model flying wing aircraft now have very impressive performance.[1] Over 400 MPH. That one has a jet engine which costs several thousand dollars and has enough power to accelerate straight up.

There's some suspicion that the triangular aircraft seen buzzing U.S. Navy ships off the California coast are drones like these.

[1] https://www.youtube.com/watch?v=DPGDAZyQ44k

The most powerful engine I'm aware of is the JetCat P100 Pro - over 1 kN of thrust or enough power to lift off 111kg off the ground [1].

Madness. But it will cost you more than a car.

[1] https://rcmodelturbines.com/de/jetcat/27-jetcat-p1000-pro.ht...

That's a fundamental problem with jet engines. They can be made smaller, but they don't get proportionally cheaper. Many people have tried. But human-rated jet engines don't seem to get cheaper below 6-passenger bizjet size. This is why light aircraft are still mostly powered by reciprocating engines.

NASA and Williams tried to crack that problem in the 1990s, without success.

the return of the Zagi flying wing!

https://radiocontrol.fandom.com/wiki/Zagi

there's a huge variety of kits available from a variety of manufacturers. They're also super easy to make with just a hot-wire cutter and there's lots of derivative designs.

https://flightpoint.co/rc-flying-wing-plans-free-download/

Generally speaking it's one of the easiest to build, and most durable designs. Those "kits" aren't even grandpa building a plane from plans and balsa wood, they're wing sections that you cover with shrink-film and attach some hinged elevons and a motor to.

SweepWings has some great build videos (and kits, for those interested). Not affiliated - just a fan.

https://sweepwings.com/collections/frontpage/products/flinch...

I've been looking for instructions for a drone this simple and cheap for a while now, guess I have a project to do now.

This uses the same control surfaces as a v-tail RC airplane ;-) . I built one like this, fairly simple controls using an aileron mix (as elevons) to combine the traditional aileron, elevator, and rudder controls. Note that you can find a lot of RC projects like this online under keywords like "RC flying wing plans".

somewhat relevant, saw it today. Create an inexpensive bomber copter from scrap metal: https://pbs.twimg.com/media/Fb6A-NxWQAMjTC7?format=jpg&name=...

Looks a bit more complex than "scrap metal" -- it might be rough-looking, but it's a quadcopter drone with payload management, and some pretty large motors.

Enough of these drones, who else here remember "model planes"?

It's only a drone if it can demonstrate stable flight by itself.

Sadly, the remote control chapter is not (yet) there, and the self-flying drone capabilities are only briefly mentioned in the intro.

I suppose that for a real self-flying capability, a model plane would need at least a 3-axis accelerometer, a speed sensor, a height sensor (or at least a ground proximity sensor), and a controller to make sense of all of the data. That part is going to be a bit less cheap and trivial.

They're quite cheap and "trivial," honestly - iNav, ArduPilot, and PX4 are all strong in this space, and controllers start in the $50-$60 range (look at Matek hardware). Of course, for some definition of "trivial," but we do live in an age where a fairly accurate MEMS gyro/accelerometer, a barometer, and a fast STM32 microcontroller can all be had at a low price, and the control algorithms are mostly tried-and-true PID / feedforward PID attached to some form of input filtering.

One can definitely build a fully autonomous flying wing for under $200, especially if the foam is self-cut. There are also a wealth of cheap foam pre-molded models available these days. We're really in a golden age for DIY flying models right now, pending regulatory crackdowns in most countries.

I wonder if pitot tube input is a thing, yet, for inertial-platform controllers. Most you can buy were designed for quadcopters.

It is astonishing to me how you can run fantastically good Free Software control software on them all, interchangeably.

Yes, pitot tubes have been supported for a long time by all of the "mainstream" flight control software (PX4, Ardupilot, iNav although in iNav it doesn't do much besides provide an on screen display readout). Both analog and digital (I2C) pressure transducers are generally supported by direct interface to either an ADC pin or the I2C bus, which is exposed on most boards, even those designed for quadcopters.

Many Ardupilot and PX4 boards are starting to come with CAN ports as well for UAVCAN peripherals.

The wings are two sandwiched pieces of foam with only the slightest bit of trimming on the leading edge. Certainly not anything like the smooth upper surface I’d assume was necessary for a wing to be aerodynamic. I guess it doesn’t matter too much for a plane this size, but would a smoother wing surface actually perform better?

I’m also fascinated by the variety of shapes of RC planes, compared to your traditional passenger aircraft. Is there a good primer on model plane design?

Even a flat piece of foam will generate lift [1]. The only requirement per se is that the trailing edge of the wing is somewhat sharp. A more optimized shape would allow the plane to fly at a lower angle of attack and would reduce drag, but drag isn’t exactly your problem if your thrust to weight ratio is 2:1.

1- https://www.av8n.com/how/htm/airfoils.html#fig-barn20

Yeah, a proper airfoil would absolutely perform better, and people use hot wire cutters on foam like this to produce airfoils [1][2], but that's more of an advanced craft and not necessary for getting something which flies well.

  1: https://www.youtube.com/watch?v=yIxyyTnT-hM
  2: https://www.youtube.com/watch?v=JIRzQO4F8ck

The power to weight ratio on these machines is very high indeed. They can accelerate straight up. In horizontal flight you can get plenty of lift from angle of attack without careful wing shaping. You might get more efficiency with better shaping, but these things are already so outrageously flyable that the demand on the pilot is the limiting factor.

Normally the foam core board is cunningly bent over to form a pretty darned good airfoil.

It's a bummer that the only ESC/motor they recommend has been discontinued. Choosing the right ESC/motor/prop has always been the most daunting step for me when considering parts for a plane.

ESCs have improved absolutely phenomenally over the last 10 years. You can hardly go wrong anymore.

I have an RC plane that is without remote control. How can i bind the random remote control I have with the plane?

You need to find the type of RC receiver on the plane, and the protocols it speaks, and the type of radio/transmitter you have, and the protocols it speaks. A common controller/radio/transmitter brand is FrSky with the ACCST protocol (apparently now ACCESS), Flysky with AFHDS2A, and Spektrum with DSMX. Each of these brands and protocols has their own binding procedure described in the transmitter and receiver manuals. Since this article is about inexpensive, the least expensive, well-known brand I listed here is Flysky. Frsky is more about open source and modularity, and Spektrum is the most American but not cheap or open. Radiolink appears to be another large brand but I don't know as much about them.

Generally, all of these are about sending a PWM signal, with the pulse width being 1.0 ms - 2.0 ms, where 1.5 ms is the "center" position, at an update rate of 50 hz, to represent the controller stick positions, with 4 channels, two for each axis of two sticks, and perhaps additional channels for communicating more data from a fancier controller to a fancier receiver for things like flaps, landing gear, flight modes, etc.

Frsky used to be good but they pushed too many shady things and caused too headaches to their users.

Everyone now moved onto ExpressLRS (an open source, very high quality rc link protocol) that multiple hardware vendors support.

Radiomaster, Jumper, Mateksys, Betafpv and more vendors support Elrs.

My next radio probably would be radiomaster once my frsky hardware dies.

I don't really see what FrSky did as shady - they introduced a new system with DRM that was not compatible with their old system, which was widely cloned. They did a really poor job at communicating these changes and explaining the implications, but I don't feel that there was anything malicious about it - just a matter of poor documentation. Their main competitor at the time (TBS) also has extensive DRM on their system, they just haven't yet had to deal with 1:1 clones.

Anyway, you're right that ExpressLRS appeared to eat their lunch, especially now that it's approaching legality in many regulatory regions. ELRS is really a very strong open source story - a hack project appeared out of nowhere and managed to disrupt a whole cottage industry by virtue of being community built and supporting a wide range of implementations.

The shady part was not them releasing their new ACCESS system. But releasing stupid "critical" (not really) firmware updates to their ACCST 1.x systems, that were backwards incompatible with their own systems. At one point users couldn't even find the older firmware for their own hardware, and had to download it from 3rd party backups on Google drive etc.

It seemed like they always took that one extra step to make life more miserable to their own users. Inverted ports for sbus (that were a pain for telemetry on F4 processors). Kept releasing so many protocols (D8, ACCST 1.x, ACCST 2.x, R9, R9 2019) and new hardware that didn't work well with their own older hardware, just to stick it to "1:1 clones"?

I used to help out a lot of newbies with this hobby and half the issues I have had to help them with were just the Frsky compatibility issues.

You'd need a really strong reason to not use ExpressLRS, especially when coming in new. The Wifi flashing, binding phrase, "power pellet" receivers, and (in ELRS 3) Wifi-connected Betaflight (no more USB!), are space age technology compared to the previous generation.

The best support for random consumer rc protocols is the open source https://www.multi-module.org/

It can be plugged into various TXs that support modules, and many TXs use it as their built in primary TX component.

There's a lot of details missing from your question I think. I tried doing some brief research for you because the question sounds interesting, but without knowing information that would narrow down the quality of the remote or the plane, it's too difficult to guess what your next steps are.

Look at the docs for the radio on the plane and then acquire a remote control that is compatible with that radio. Or, if you don't even have a radio, buy a remote control and radio, read the manuals, install the radio (and battery), connect the servos, check that it works.

Obligatory link whenever people are interested in flying wings: https://m.youtube.com/watch?v=hTYuj7hzKeI (building an rc flying wing with a pizza box) . It's actually a really cool DIY channel imo.

In general, I think these flying rc toys (especially the fpv side) are a great hobby to get into for the Hacker News crowd. Almost all of the software is open source. Hardware is mostly open too. Lots of things to learn and tweak and tinker.

I wonder you were to start such a project and desired a video stream, how feasible it would be.

Very, very, very.

You can't sell a house anymore without a fly around video.

So, I'd you don't want to build one yourself... What is good ready-to-fly option?

A good plug n play model is Zohd dart? https://www.getfpv.com/zohd-dart-250g-570mm-epp-fpv-wing-pnp...

There are even more "ready to fly" kits but in this hobby, you typically want to buy a good transmitter separately. That way you can reuse it across all your models, especially when you crash and get a new one etc..

A good radio is maybe Radiomaster tx12 mk2 Elrs version and Matek makes decent elrs pwm receivers.

I mostly fly fpv quadcopters and don't need too many sliders and knobs on my radios. Hence the TX12. Someone else may have a stronger opinion about Radiomaster Tx12 vs. Radiomaster Tx16s for fixed wings.

This thing looks like it would get obliterated by event the slightest amount of wind.

Amazingly, not. These things are very sturdy.

Reply on news.ycombinator.com