The Trippy, High-Speed World of Drone Racing | Hacker News Reader

The Trippy, High-Speed World of Drone Racing

newyorker.com

68 pointssmoyer28 comments

I design and build my own quads, though I'm not a particularly great flier. The thing I love about the hobby is that it combines three key things:

* The control joy/mastery feeling you get from activities like car racing, riding a bike on technical terrain, playing a game, or flying a plane[1].

* A sufficient risk/reward profile; you risk smashing a $100 device, which is expensive enough to be interesting, but low enough not to be paralyzing

* The gearhead thrill of ordering, rebuilding, designing, and discovering new combinations of equipment

In my case, I want to optimize for quietness and unobtrusiveness, which means aiming for lightness, which is its own set of fun design problems - I'm currently down to 25 grams for a brushless FPV setup[2]. Now I'm on to redesigning the RC transmitter[3], since I want one with nice gimbals that takes up as little space in my bag as possible. The set of things you can do is endless.

[1] I hooked up an RC controller and went FPV with my home flightsim setup. It was fun, but there's a reason RC sticks are tiny - you need to move them around a lot more frequently and quickly than in a plane: https://twitter.com/gmurphy/status/832698732361551872

[2] https://www.instagram.com/p/BaSymwPH7XJ/

[3] https://www.instagram.com/p/BdnapxGltfw/

For the video in [1] are you continually adjusting the throttle to maintain altitude or does that quad have some sort of altitude hold?

Slight and continuous adjustments, though with a proper RC controller you can make faster and more accurate adjustments, so maintaining altitude would be much more accurate than all the wobbling you saw in the video :)

Nice clean build, what frame is that?

Just bought a pair of fly egg 100’s frames (not kits) for a tiny build but I was planning on deleting the “cage”.

The FrSky TX is quite a project too, very interesting. You got a write up or project page somewhere?

It's a 2g frame and camera mount I designed - you can download the plans here (the button next to the QTY selector): https://armattanproductions.com/pages/product_details/12158

The TX is a Jumper T8SG + FrSky M9 gimbals and recased into a 3D printed case I made. Haven't finished it yet (waiting on a third iteration to be printed), so writeup when I'm done :)

Loved the video. Was actually funny as well as being cool.

What's up with the old VHS-style interference? Is it an analog radio link?

EDIT: yep it is, I know the answer now I've read the article.

"Maybe someday we’ll all be flying in commercial airplanes directed by calm, professional, reassuring pilots who are sitting on the ground."

Color me old-fashioned, but I like it when my pilot has some skin in the game.

I recently saw a very interesting CFD study by NASA recently on drone configurations, where they found that a hybrid version with front rotors below the frame and rear rotors above was significantly faster, since the wakes interact less. Wonder if it's been tried on racing drones?

https://www.nas.nasa.gov/publications/articles/feature_hybri...

(The flow simulation and visualization in the above link is so insanely high resolution it's borderline pornographic.)

I'm not completely sure whether this is as relevant for racing drones as it is for surveillance/video drones. Racing drones often fly at a 45+ degree angle, so the thrust is angled more backwards than downwards. Also, one of the benefits of the NASA research is longer flights. But racers often don't fly more than 1-2 minutes, after which they swap batteries. A slightly lighter battery could give the benefit of slightly quicker acceleration though and that's pretty important. Lastly, in the pro leagues, they usually fly the same drone, so there's no need to have the fastest drone.

Racing drones typically fly at extreme tilt angles such that the wake of the front props isn't interacting very much with the wake of the rear props.

That said, I have seen that configuration on a frame.

I’ve not seen it on racing drones (likely for the reason mentioned by others), but have seen it in mid-range pre-built ones quite a bit. For example, the DJI Mavic has a similar configuration to what you mentioned.

I’ve built or bought well over two dozen drones ranging from palm sized to my beast, an octocopter with a 5’ “wing span” that draws hundreds of amps at 25v at max throttle, but never got into racing. I’m much more into the building and photography aspects. I’m tempted to build my own software from scratch next (there’s great open-sourced platforms, but I like the challenge of doing it myself).

Drone racing was on like ESPN 8 one day and it was very interesting to watch. Honestly if E-Sprots/Nascar have a following this should as well.

I can't tell if this is serious. Do we really have "the ocho" now?

It was not really ESPN 8 but it was pretty far down there

ESPN did it for a single day last year.

ESPN: College was taken over for a single day as "The Ocho" last year:

https://www.sbnation.com/lookit/2017/8/3/16091446/espn8-the-...

https://www.youtube.com/watch?v=jow5hZPF3zY

I would love to see this evolve as a competitive sport.

Because no pilots would be injured, it would possibly be okay to allow the drones more contact with each other bringing in exciting aggressive elements.

The problems I see are just how fast it moves. There would need to be so much live slow motion replay to catch everything that was happening. It would be very difficult for commentators.

Also the drones need to light up if they miss a ring, or have a better way to express that.

I've seen it peak much higher on the list than 'the ocho'.

love this quote "An immeasurable amount of scientific and technological progress, like a huge invisible inverted pyramid, converges on this small, toylike point. "

A friend of mine in the drone racing/stunt hobby said something to me the other day that really blew me away, "the real improvement in drones in the last 2 years isn't batteries, it's that the output of the motors has increased so much that we've all dropped a frame size and kept the same output." He then said that it might happen again soon.

I don't know if that means that the real output of the motor is truly increasing that fast, or the effective output taking into consideration the reduced weight, but those little motors putting out a ton of power, very efficiently.

The combo of brushless DC motors and lithium batteries more or less revolutionizes everything it touches.

Max output of a motor is constrained by weight, or more specifically, its ability to dissipate heat. Racing quads generally use 30-40 gram motors. The top racing motors these days produce 1.6kg of thrust at max output and consume 40 amps of power in the process. It's not difficult to produce a quad under 500 grams.

4 motors would produce 6.4kg of total thrust for a 500 gram quad for a nearly 12.8:1 thrust/weight ratio at a 160 amp current draw. A few years ago a 4:1 thrust/weight ratio was considered pretty good.

160 amps might be a stretch, but it's not uncommon to get >100A out of a 1300mah battery.

All that is pretty crazy, but I think the more interesting bits are in the open source efforts around flight controllers and speed controllers.

With those kinds of currents, it'll be interesting to see if point of load DC-DC converters become integrated into the motors themselves so as to reduce the wiring requirements for distributing the power to the motors. The batteries will need to stack more cells in series to support this but I wonder at what currents the wire gauge, and hence weight of the wires, become a big concern.

Maybe modern quads already do this?

Wire runs are so short and peak current is so brief that it effectively doesn't matter.

Most folks run 14AWG between the battery and the power distribution board. PDBs are generally made with 2oz copper and wide traces to accommodate the current, from there, the total current load is divided by 4, assuming that's how many motors or running.

> Wire runs are so short

This! The length of wire makes a huge difference in rated capacity.

All of the flight controller development I see is just circuit optimization for LC filters (cleaner video), integrated OSD, smaller 20mmx20mm footprint for micro-builds, hardware-wise anyway.

There is a good amount of competition in the cleanflight/betaflight/raceflight firmware arena and there is outstanding hardware support for an amazing number of FC boards.

Speed controllers are playing songs now, months after my bro did some ESC assembly hacking to get his playing the first few bars of the doom theme, it’s just a feature now.

Frankly I think the powerful open-source FrSky gear, the taranis TX and the OpenTX movement has made the biggest impact on the RC scene overall raising the feature bar and lowering the price point significantly over spektrum and futaba gear.[0]

[0]: https://www.frsky-rc.com/product/taranis-x9d-plus-2/

160amps would be a real stretch for the weight requirements. I’ve pulled 400amps consistently before off a 6C battery under 4lbs, but getting nearly half that in under 1/4 of the weight seems a bit out of reach for existing tech unless you don’t care about exploding batteries.

Late reply, I know.

The folks pulling down 160 amps are racing and don't care about exploding batteries. In fact, they expect to.

The embedded video is well worth of watching! Trippy indeed.

Reply on news.ycombinator.com