Build Your Own Thrust Vectored Rockets for Vertical Landings Like SpaceX | Hacker News Reader

Build Your Own Thrust Vectored Rockets for Vertical Landings Like SpaceX

makezine.com

457 pointsgscott72 comments

Joe Barnard, who looks a bit like the energy & space entrepreneur, earning him the title among his friends of “discount Elon Musk,” is one of the very best channels among tinkerer YouTube. Not annoying or fake enthusiasm like some of the other channels (but is enthusiastic!), innovative, informative, and fun-loving. A joy to watch his videos.

Also, I had ice cream with him one time after watching an Antares launch in Virginia.

Also good to see Make isn’t dead.

Totally agree. Very prolific with a mix of software and hardware engineering, not at all shy about showing failures and setbacks and puts a lot of effort into the production value of his content. One of my favorites on YouTube for sure. I do hope he gets into hybrid or liquid propellant to avail himself of some basic thrust control...his new test stand has me hopeful.

https://www.youtube.com/channel/UCILl8ozWuxnFYXIe2svjHhg

> not at all shy about showing failures and setbacks

This is the same reason why I like Alec Steele, a popular blacksmith channel! It's much more interesting to learn about all the mistakes, failures and bad assumptions than to see a perfect, finished product from end to end.

If you haven't found him already, you might like Wintergatan. Martin Molin of Wintergatan is currently in the process of designing and building a successor to a marble machine instrument he built a few years ago and is extensively documenting the process through YouTube. He doesn't shy away from showing mistakes either and all of his videos are beautifully shot and scored (often with music by him). If you are interested in engineering of any kind, I would highly recommend watching the Marble Machine X series from the start. He learns (and thus teaches) a lot of valuable lessons in the process and it is really satisfying to see it all finally coming together.

I binge-watch the MMX videos over the course of a couple of weeks a year or so ago in order to get up-to-date and have been following it since. It's been great to see the thing evolve through the challenges, setbacks, and breakthroughs. The final machine seems so close now!

He consistently does experimental things with model rockets that no one else has managed to pull off, or would even try at that scale. Reaction controls, launch tubes... he's really cutting edge, which is unusual for Youtube channels.

Second this it is been awsome to follow his journey.

Like SpaceX?

"If you love rockets, you can’t help but notice that real space launch vehicles lift off the pad slowly, but model rockets zip up like darts."

That reminds me of a common issue with modeling: when you divide sizes by N, you divide surfaces/light-reflexion/air-resistance/lift by NxN and volumes/weights by NxNxN.

A 1/10th model is 1/1000th of the original weight with identical materials. All the dynamics are different and easier at smaller scales. This makes the real thing an expert work while modeling is reachable by hobbyists - very good ones in this case.

> All the dynamics are different and easier at smaller scales.

Are you sure about this part? (not rhetoric). Non-linear scaling of surface and volume are simple to understand, but dynamics doesn't look so straight forward to me... in my short lived experience trying to fly very small model helicopters, it was clear that the smaller they are the more unstable they were. I wasn't sure how much of this was due to limitations in human reaction time and how much was inherent aerodynamic instability at smaller scales.

In these rocket models the human limitation is clearly removed, the remaining dynamics look faster at least which may or may not run up against higher frequency sensor data and processing requirements... are there other dynamics i'm missing? i guess materials don't bend much at this scale?

Different? Yes. Easier? Depends. One thing you gain with size is intertia, which grows with N³, and depending on the design, can be helpful for stability.

Yes, thanks, this is what I meant but the relationship was not clear to me before, it seems obvious in retrospect: n^3 * density = mass scales cubed, which as far as control is concerned is both good (lower impulse requirements), and bad (lower relative mass requires much finer control)... happy to be told how to express the later formally :)

The think that makes this scaling "bad" is that moment of inertia is going down faster than everything else and this means the required control frequency goes up. A Falcon-sized rocket may only need to correct its course 50 times a second while a 3-foot scale model would need to do it 10x faster. This means your sensors need to be faster, your guidance computer has to run faster, and the actuators have to be able to respond faster.

The hard things on the other side tends to be power requirements, for exactly the same reason. Mass scales up as the cube of size, including the mass of the thnigs you have to move. Although you don't have to move them as fast, the net effect is still that your power requirements become very large for large-scale vehicles.

I think in terms of "less mass = bad", the formal answer will be found in control theory.

You are certainly right on helicopters' dynamics...

I was more thinking of planes and rockets when writing on dynamics: they are shaped to break through and be tunneled by a non-moving air for stability, and their weight decreased faster than their wing surfaces, making lift easier at smaller scales.

Helicopter are a different realm: they need to survive in the middle of the wind (and turbulence) they create to lift, and they are not tunneled (their body is not aligned on the vertical flow).

Thus, as Temporal mentions it, inertia is important for helicopter stability and it is reduced with weight at smaller scales.

> I was more thinking of planes and rockets when writing on dynamics: they are shaped to break through and be tunneled by a non-moving air for stability, and their weight decreased faster than their wing surfaces, making lift easier at smaller scales.

I suppose there is also relative viscosity to take into consideration? so even if smaller scales are going to be more fidgety and "unstable" in terms of inertia (as TeMPOraL more clearly expressed)... taking aerodynamics into consideration (depending on the design) may provide more significant benefits to dynamic stability at small scales anyway.

Without being very scientific, it feels like the small scale dynamics are not merely easier, but significantly different. It's intuitive to see how insignificant aerodynamics are at take off in full scale rockets are compared to models, and how models are going to be more sensitive to aerodynamics than inertia... i suspect the proportions to the problem of dynamic stability might even be flipped.

You also have to understand that small-scale models need the time dimension to be slowed down by X too in order to get the same dynamics.

Curiously, he still has 'center core' issues even at this scale.

His rocket is '3 tubes' each with their own motor. The motors produce large thrust, so much that the energy is more than enough to break the three tubes apart if not applied in a very coordinated manner. To make 3 rockets into 1 rocket is a central issue of this design! Just like the big ones.

"All the dynamics are different and easier at smaller scales."

I suppose that's why balancing a pencil upright on your fingertip is like 10 times easier than doing the same with a broomstick, right?

I think you are being sarcastic, because I can balance a broom a lot better than a pencil on my finger, so I'm going to discuss my thoughts behind why.

Based on what you are saying and tomxor in a sibling comment. Stability seems to increase with scale. Maybe it has something to do with weight, or center of mass, or a combination of the two.

I know that the longer the object you are balancing upright on your finger (or palm for something bigger than a broom), the easier it is to compensate the shifts in center of mass as it tilts. For a pencil, technically i guess it is "easier" to balance if you are a machine able to make the fine micro adjustments required.

The equation for the period of a pendulum has sqrt(l) in it. Therefore, I would guess that to get equal dynamics at a minimum, the feedback loop speed (the time for your hand to respond in the case of the broom) has to increase by the square root of the scale factor.

And that is why I'm a web developer ;)...

I chose to avoid all math after high school, how dare you bring formulas into my civilized, philosophical discussion!

It's inertia. Among other things, the much higher inertia of the broomstick has a damping effect on changes in its motion.

A broom is much heavier.

Your arm and hand are not great at doing many precise and gentle movements per second. With the acceleration they produce they can do many small corrections for a broom position. For a pencil, much faster and gentler corrections would be needed, and the arm + hand are too slow and imprecise to tackle that.

That's the parent commenter's point.

See the sibling comment on why helicopters differ from rockets/airplanes.

I think this is an issue of inertia/mass that doesn't exist so much for craft that gain and retain stability from their surroundings.

Of course, I could also be completely wrong.

I'm actually with you on the dynamics thing- the actual oscillations are slower and smaller, but they're cubically harder to deal with. A small oscillation in a full-scale orbital rocket can shred the skin like gossamer.

However, the issue is much worse for the general problem of rockets, as the term in the rocket equation[1] is logarithmic. The difficulty takes off so quickly that rockets go from club sport difficulty (eg copenhagen suborbitals) to national research programs when the rocket is just a few times taller. It gets said semi-regularly, but if the earth was just a little bit bigger (again- gravity rises approximately with radius^3) then we would never have visited space.

[1]: https://en.wikipedia.org/wiki/Tsiolkovsky_rocket_equation

It's a frustration of mine with model rollercoasters. The train zips around the track much faster than the real rides. I've been playing with ideas like magnets and DC motors to try and make the trains move more slowly.

You could add inertia by including a flywheel (or multiple) connected to the wheels of the coaster. Not sure how easy that would be at small scales though.

I think you raise a good point about scale. I don't think full sized rockets lifting off has as much to do with gimbals vs. fins as it does with sheer weight. The Saturn V had fins on it and it lifted off slow, and I'm willing to bet that it had a lot to do with the fact that a Saturn V on the launch pad weighed somewhere around 6.2 million pounds.

Scale it down a ways to something like a RIM-116 and you can see the scale come into play. This thing goes from launch to almost disappearing over the horizon in under three seconds and then hits the target.

https://www.youtube.com/watch?v=pVk9VnUkvaU

Granted, that RIM-116 does have fins. So let's take the fins away and scale up to something like a Polaris SLBM

https://www.youtube.com/watch?v=sUlXty69-Y8

Still unbelievably fast, same with the Tridents. These get popped up out of the water before they fire their main engines. It's basically free standing in the air when it fires the main and it still shoots off super fast.

https://www.youtube.com/watch?v=h5KejRbD5s0

It has nothing to do with sheer weight and everything to do with the thrust-to-weight ratio.

Saturn V was slow to liftoff because the launch configuration had a very low thrust-to-weight ratio, IIRC about 1.1

And you don't think that has anything to do with the scale of it? The size of the rocket necessitating a lower T-W ratio due to the impracticality of making a rocket engine large enough to increase it?

And yeah, sheer weight definitely has something to do with the thrust to weight ratio. It's half the equation.

Honestly, maybe partially scale, but not really. It's more a factor of what it was meant to do.

Going to the moon means extra hardware on top, which means more fuel, which means lower TWR at launch.

If you're just going to drop a 6000kg bomb on earth somewhere, you need a lot less rocket on top of that first stage, and it's going to take off a lot faster.

I think we're just going in circles now. The scale was a result of what it was meant to do. The extra hardware on top and extra fuel all add to the scale of it. It was a massive rocket because it needed to do a massive job.

Anyway, I think we've gone off track from the original point, which was that full size launch vehicles taking off slow isn't because of gimbals vs. fins, it's because they're big and heavy.

We can be pedantic and say it's specifically because of a low T-W ratio, but the T-W is low because the weight is high and it's not practical to build a massive rocket engine to increase it.

um, sure.

Yep, and it's all down to the squared/cubed law. The thrust that a rocket engine makes at a given operating pressure is proportional to the area of the bell, and so scales with the square of its linear size. The mass scales with the cube of the linear size, and so for a given shape of rocket, doubling the size will multiply the thrust/weight by a factor of 2^2 / 2^3 = 1/2.

> That reminds me of a common issue with modeling:

The scientific field that studies this is called "Dimensional Analysis", and its major is probably the Buckingham Pi-Theorem, which takes maybe 10 minutes to learn, and is one of those simple physics ideas that translate to a lot of fields (like conservation laws).

Obviously we need human ovum sized spacecraft! :P

Then maybe we can get to orbit on solar pressure alone?

Humans are not reproduced by body alone.

To produce a useful human, you need 15-20 years of upbringing by other humans, preferably in a thick layer of material culture.

Colonizing space by transporting just fertilized ova is not going to work.

Been a big fan and supporter of Joe's work for a long time, really amazing work he is doing and how open he is about it, sharing lots of his knowledge in the BPS.space discord. Great to see him on the front page!

He’s such an inspiration. Zero background in any of this and just went “This looks cool, I’m going to figure it out”.

His videos are incredibly entertaining and educational too.

The Falcon Heavy model launch is one of the funniest and inspiring things I’ve seen: https://m.youtube.com/watch?v=ff5245EBfis

Awesome you mean! Even when it doesn't work, it's doing something amazing.

Somebody buy this guy a Red Button, he deserves it.

That is innovation right there in our faces. All that stuff was theoretically possible ten or twenty years ago. But completely impractical for a single hobbyist.

Not to mention that propulsive rocket landing is a thing for -what? Six years?- in the real world now.

There were hobbyists making thrust-vectored rockets at this scale twenty-ish years ago.

e.g. http://michael.sdf-eu.org/Gyroc/

Gyroc was amazing at the time and is still spoken about in hushed tones in the UK HPR community. While it was using ADXL50 devices for acceleration, the main gyro was an entirely custom built mechanical rotating gyro, wish I still had some pictures of it. Now that's all on a single chip you hardly even notice on the board :) Some pictures of it are here http://ukrocketman.com/rocketry/gimbal.shtml

It's crazy how some things can rapidly go from breakthrough to commodity nowadays. It reminds me of early drones vs ArduPilot in a way.

More like 50+ years, if you count the LIM, and all other retro-rocket use.

The first falcon 9 landing was about 4 years ago if I remember correctly

If you want a quick video overview:

"BPS.Space - Channel Trailer - 2019" https://www.youtube.com/watch?v=OE0_-g7YV1M

Fun fact, the landing method described is called a "suicide burn"[1].

[1] https://space.stackexchange.com/questions/10307/what-is-a-su...

I highly suggest doing some digging into the rocketry hobby (especially high power rocketry). Or, as my wife would say, "go down the rabbit hole".

It's like software development, there's some heavy duty stuff being done by incredibly skilled and talented people outside of work hours.

For example, this guy put a rocket over 200k feet altitude on COTS motors https://mach5lowdown.com/2018/11/07/phx4-rocket-launch-to-20...

Also, check out the engine this guy is building http://www.watzlavick.com/robert/rocket/regenChamber3/photos...

Trick of SpaceX success wasn't in vertically landing rockets - with mass being concentrated in the bottom in form of engines, thrust structure, and legs, it's not all that hard.

What was extremely hard, is havigation and control to enable return back to launch site optimally, going through narrow corridors of acceptable parameters (go too shallow and you risk being unable to correct your position in the end as precision drops, go too steep and rocket breaks up due to aerodynamic forces, and so on), while burning minimum fuel. That was an insanely difficult GNC task which took so many iterations to perfect out...

Plus yes, on landing they didn't have a second chance as rocket couldn't hover - minimum engine thrust was bigger than it's mass so if it missed, it either crashed, or flew back up until running out of fuel and crashing, too.

If I understood correctly, the way to land when even minimum thrust is too much is both simple and clever. The goal is to reach zero velocity at zero altitude. But there's always some uncertainty in how rapidly an engine will fire up and come up to power, so you can't just precisely time the engine startup. So what you do is to time the startup so that the predicted zero velocity point is underground. Then once the engine has started and spooled up, you continuously measure altitude and velocity, update the predicted zero velocity altitude, and progressively increase thrust until the zero velocity altitude is at zero altitude.

I mean, seems pretty hard to be honest. They got the trajectory down pretty close for the vast majority of these (otherwise would be too far away to see with cameras). But so much extremely expensive space hardware destroyed in the pursuit of this goal. They earned it: https://youtu.be/bvim4rsNHkQ

Mass distribution isn’t part of it as long as things are symmetrical (https://en.m.wikipedia.org/wiki/Pendulum_rocket_fallacy) but yeah, threading the needle like SpaceX does is not easy.

Got to say, I don't understand how the argument in that article works. If Goddard's rocket that's on the photo had a canvas stretched between the wires/pipes that form its body (I assume it's a part of the rocket, and not a launch platform), then I'd expect aerodynamic forces to act on them, countering the tilt. Conversely, if such surfaces were above of the engine, I'd expect them to make the construction unstable.

You're right that any canvas stretched around the frame could act as aerodynamic stabilizer fins. The pendulum rocket fallacy is the assumption that a higher mounted engine will stabilize the rocket's direction due to it 'hanging' from the engine, by analogy with a pendulum (where the higher a point you 'hang it from' the more stable it is). The fallacy lies in the fact that the pendulum is hung from a fixed point which can apply forces to generate a stabilizing torque, whereas there are no such outside forces generated by the rocket engine (other than the thrust it produces).

Fins totally work and that's why almost all rockets use them to some degree for stabilisation while in the atmosphere.

The pendulum rocket fallacy refers to the placement of the engine itself, not center of mass. Mass distribution absolutely matters, because with vectored thrust, your engine can exert a higher torque on the rocket if the center of mass is farther away. You can try it yourself with a long pole and weight (e.g. broom with big brush?). Balancing the pole with the weight on top is easier than with the weight on the bottom.

But for landings, it's totally a part of it. If there is a mentionable lateral velocity component, rocket will tip over on landing unless it's center of gravity is very low. It happened just once with SpaceX rocket though.

Due to fuel slosh, things aren't symmetrical really.

I showed my wife the video, and she remarked "You know, we went to the moon with less than that".

So, puts it in context! So much possible now that was unthinkable back then.

After seeing the headline I thought "If this is by anyone other than Joe Barnard they've really missed an opportunity here". Glad to see it's him! His videos are always in-depth and don't hand-wave or gloss over the various technical problems he encounters/fixes. He also does detailed breakdowns on the successes and failures of his launches which I really enjoy.

In my eyes it’s amazing business using affordable hardware components. Respect for the creator! I dream finding similar niche.

I have been launching rockets with my son the last few years, but it was kind of boring after a while because you I thought you can't (legally) introduce any controls. I thought! This has taught me that you can introduce controls for stability. That sounds pretty exciting.

We've gone from Elon Musk doing cutting edge stuff as a billionaire to a guy doing the same thing in his backyard within a decade. What an amazing time to be alive!

I thought vectorable thrust could run you afoul of ITAR, and pushed you into a much higher realm of scrutiny, since you've basically designed a missile.

In the article he declares he is doing stability control as opposed to guidance control, and only the latter is regulated.

How do you land without doing guidance control?

He's not landing it back at the pad or anything like that, just wherever it happens to be coming down at. Then keep it vertical(ish), keep it from moving too fast sideways or it'll tip over (again, stability, not guidance), and wherever it touches down, it touches down.

So by design, the rocket motor is going to be active, and pointing at something he doesn't control?

After all the dire news in May I'm very happy so see good content on a Make domain :)

Cool project, but ... that website gave my browser cancer (if you scroll down too far, your history is messed up and back button does not work).

Reply on news.ycombinator.com