Just an idea

[RoadRunnerAerospace]

soo.... I love to launch model rockets... and watch the failures. But anyways, I was thinking about some cool project I could try to do, and I thought of a falcon 9 esque rocket, launching with a "d" class motor (no ejection charge), and using A class motors, which are electronically ignited when certain conditions are met (i.e. alt. of 10ft, speed of 20mph, or something) to land. to keep it stable, it would be using "waffle maker" airbrakes on the way down (the grid type things) which deploy once it reaches periapsis.

 

just a thought!

[HebaruSan]

Here's the thread from the last time this was discussed:

 

[Ultimate Steve]

I've daydreamed about it as well. It would be really cool, but hard to get the exact Delta-V and timing, especially because total impulse and ignition time can vary between each individual motor. I considered attaching a computer controlled compressed air thruster which could throttle and make landing slightly easier, but then it got a bit too complicated for my tastes.

[RoadRunnerAerospace]

well, using a flight computer (ardruino or something) you could give it a set of perimeters which must be met in order for anything to happen, and once it gets those perimeters, then everything is "unlocked" and the landing sequence starts, I`ve been thinking about two pairs of "A" class engines, so one ignites at "X" ft above the ground, and the second ignites at "X" distance above the ground, and if the speed is above "Z". I was also kind of thinking about having the "waffle maker" airbrakes adjust their AoA to keep the rocket steady using two-four range finders it determine the rockets attitude, and then comparing it to whats on the program, and making the necessary adjustments. So it wont be unguided (what everybody was freaking out about on the other thread)

[sevenperforce]

Carrying a CO2 cartridge that pops just before touchdown could slow it down without the issue of igniting an unguided retropropulsive rocket engine.

[RoadRunnerAerospace]

true, but how would you pop the co2 cartridge? 

A quick little flow chart that I threw together for the program:

all velocity's and heights are theoretical, and will be changed

Edited July 28, 2017 by RoadRunnerAerospace

[sevenperforce]

44 minutes ago, RoadRunnerAerospace said:

true, but how would you pop the co2 cartridge? 

A quick little flow chart that I threw together for the program:

all velocity's and heights are theoretical, and will be changed

Well, it needs to be linear, but that's beside the point.

Problem here is needing a laser altimeter to determine altitude. 

For popping the CO2 cartridge, you can cannibalize a broken airsoft/paintball gun, twist the cartridge into the receiver, and use a small servo to open the valve.

[RoadRunnerAerospace]

Holy crap lazer altimeters are expensive

could I use a altimeter which uses the pressure? like an aircraft's altimeter?

something like this: https://www.amazon.com/JBtek-Barometric-Pressure-Temperature-Altitude/dp/B00UUS12PO/ref=pd_sbs_147_2?_encoding=UTF8&pd_rd_i=B00UUS12PO&pd_rd_r=YRC5PPGHQWT1D1V8NE64&pd_rd_w=LH56p&pd_rd_wg=Ku5sB&psc=1&refRID=YRC5PPGHQWT1D1V8NE64

[magnemoe]

47 minutes ago, RoadRunnerAerospace said:

Holy crap lazer altimeters are expensive

could I use a altimeter which uses the pressure? like an aircraft's altimeter?

something like this: https://www.amazon.com/JBtek-Barometric-Pressure-Temperature-Altitude/dp/B00UUS12PO/ref=pd_sbs_147_2?_encoding=UTF8&pd_rd_i=B00UUS12PO&pd_rd_r=YRC5PPGHQWT1D1V8NE64&pd_rd_w=LH56p&pd_rd_wg=Ku5sB&psc=1&refRID=YRC5PPGHQWT1D1V8NE64

GPS would be more accurate and simpler, for one the gps attitude of the area stay the  same
Main issue would be sideways speed this will be increased during the landing burn if rocket point the same way. 

[RoadRunnerAerospace]

true, a gps altimeter would probably be the best option (I totally forgot about those), but how would it give a reading to the ardruino? also, if the rocket is facing retrograde, which is the way it will be facing, thanks to the airbrakes, it will cancel out the horizontal velocity 

[sevenperforce]

Pressure altimeters are not nearly accurate for the kind of thing you need. Honestly, GPS might not be accurate enough either.

A short-range infrared altimeter might work; I've used those before (though obviously not for this).

One possibility I advanced in the other thread was that the legs could be spring-loaded and the landing engine burnout thrust could be just slightly less than the weight of the empty rocket. So the landing engine decreases the terminal velocity enough that the spring-loaded legs can catch it without breaking.

 

[wumpus]

Any idea how quickly a laser range finder can spit out range?  You need some sort of range finder pointed at the ground.  I'd also wimpout and use a fake "0 altitude" measurement fairly high off the ground (and let my rocket fall from that distance).  I'm guessing that trying to figure out exactly what time the range finders finding was accurate is going to be the big killer (you can calculate your position easily based on that, but getting those two numbers to agree with each other is probably the reason robotics is hard (I've never been in field*).

* In the dawn of time I had a small job doing the telemetry of a drone designed by Burt Rutan.  But I was so green it might as well been considered a student project.

[StupidAndy]

is this relevant?

https://shop.spacex.com/accessories/falcon-9-and-fairing-flying-model-rocket-kit.html

[YNM]

I'm guessing that if winds are not too strong and the lands are level you could get away with a timer instead. But to do it properly from various trajectories you need a range finder like everyone else says, and a custom-made rocket motor (I assume you need to vary the thrust somehow).

1 hour ago, StupidAndy said:

is this relevant?

https://shop.spacex.com/accessories/falcon-9-and-fairing-flying-model-rocket-kit.html

I don't know they have that !

[Shpaget]

It's probably been about 20 years since I launched my last model rocket, but if I remember correctly, those engines had a large spike in force at the start of the burn and then a longer period of lower thrust. That may prove to be difficult to manage.

I'm also not entirely sure about the consistency and repeatability of their performance.

[RoadRunnerAerospace]

In fact, that spike is what i`m relying on, so it starts to slow down kinda close to the ground, then goes "nuts" when its really close

[Snark]

If you're looking for "tech solutions to determine height-above-ground-level to prevent crashing", perhaps may be worth taking a page from the drone industry.  Drones these days (at least, some of them) have collision sensors on them to aid in landing (both as a tool for completely automated landing, and to save a human pilot from themselves if they descend too quickly).

One example I've seen is a drone that has two collision-detection systems facing downward:  an optically-based one (uses binocular vision between a couple of small cameras) and a sonar system (like those old ultrasonic rangefinders on cameras like the Polaroid).  It has the two systems because one can backstop the other-- they have different "trouble spots".  The optical sensor doesn't work if you're in low light levels, or if you're over an optically-uniform surface.  The sonar sensor doesn't work if you're over a very sound-absorbent surface.  Put them together and they're a bit more reliable.

One issue with those two technologies is that they don't work at big distances above ground level.  Not sure what the range limit is, but I'd guess it's under 10 meters.  That may be fine for a drone whose top descent speed is only a few meters per second; might not work as well for a descending rocket, if it's coming down so fast that it needs to fire retro-rockets to slow down when it's still high above the ground.

And, of course, there are practical constraints of form factor and such (i.e. drones and rockets aren't shaped alike), which could limit the applicability of any particular piece of drone hardware to a model rocket.  I only mention it as an area perhaps worth exploring-- i.e. "something small that flies, which needs a small / lightweight / reliable way of knowing height above surface."

[wumpus]

I've looked into optical range finding (one job involved focusing a camera quickly), but it didn't appear to be something you could quickly use (mostly due to speed/latency issues), and we had a sonar system that worked better in house (note that cost wasn't a big deal at this place).  In any event, at least one of the wikis has a sign wrong in the equation (that messed me up for awhile).

The CO₂ idea sounds ideal, assuming you can throttle it.  Especially since that the more "real" methods threaten to set anyplace they land on fire.  I'd assume anyone wanting to use "landing on a flame" should at least have several successful CO₂ landings.

[RoadRunnerAerospace]

It could be throttled, in theory with a remotely operated valve that opens a different amount at different speeds and altitudes. Another idea (thanks ultimate steve!) would be his compressed air thruster idea...

[kurja]

Range finder, valve & actuator & plumbing & co2 cartridge etc etc sounds like it's going to be bulky; can you reasonably assemble all that on a small hobby rocket?

[RoadRunnerAerospace]

who said it was going to be small?  I`m thinking of a size that is around 4ft high.... Maybe larger

[K^2]

Model rocket engines aren't nearly consistent enough for a guided landing. Take it from somebody who've spent a lot of time trying to build a multi-engine model rocket. Getting it to fly straight was a task I eventually gave up on. There is just too much variation in thrust profile from one motor to the next, even in the same batch.

Any home-built solid motor won't do any better. You have to build a hybrid with very sophisticated control system to even have a chance.

So lets see. You need to understand how a hybrid motor operates, including the relationships between pressure, oxidizer flow rate, fuel burn rate, and thrust. All of these things mutually influence each other, and all you really get the control over is oxy flow. You don't have to worry about complex nozzle, so you're off the hook on fluid dynamics, but you still need to understand combustion, thermodynamics, and non-ideal gas laws at rocket science levels.

Now, if you manage to not mess up the engineering part of it, we're next moving into building the control system. Throttling a real rocket is nothing like throttling one in KSP. The delay between change in valve position and thrust output is significant. Meanwhile, you have multiple sensors with very noisy input. So we get into control theory and filters here. There are fairly standard solutions that ought to work here, like PID and Kalman Filters. But at a minimum, you need to know how to implement them on your hardware of choice.

 

I would recommend starting much, much simpler. If you want to build a self-landing rocket, try building one in KSP first. All of the parts necessary to build a Falcon-like booster core in KSP are stock. On top of it, install kOS mod and write a script from scratch that lands the rocket safely at predetermined location. Make sure it can do so even if you vary the initial payload and ascent trajectory. Once you have that done, you'll know that you can at least handle the control portion in near-realistic scenarios. You still need to adjust for delays and noise in the real rocket, but that's definitely a doable task if you understand how to make automated landing in KSP. If you get that far, you can start looking at what it'd take in terms of an engine build.