Armchair Rocket Scientist

I've been having something very strange happen: While attempting to reenter with the first stage of a reusable launcher in RSS (tall skinny rocket, tail-heavy due to engines and LOX tank being at the back), the vehicle has attempted to fly sideways. Not nose-first, as I would expect if the vehicle was unstable, sideways. It refuses to remain pointing retrograde, and instead falls at an angle of attack of up to around 30 degrees! With Laztek's Falcon 9 landing legs attached and folded, it falls at an extreme angle of attack even at hypersonic speeds (and as a consequence breaks up due to a combination of heating, aerodynamic failures, and G-forces. With the landing legs removed, it falls at a survivable but still significant AoA at hypersonic speeds, and then increases to around 30 degrees at subsonic speeds. At less than 150 m/s, the rocket is approximately horizontal, and the prograde vector is about 30 degrees below the horizon. It stays in this orientation all the way to the ground. There are no asymmetric protrusions or anything strange other than a Mechjeb unit which is shielded from the airflow by an interstage fairing. Obviously the rocket is generating quite a bit of body lift which prevents its descent angle from approaching vertical, but that shouldn't affect its aerodynamic stability, right? Do you have any idea what might be causing this behavior?

Don't alcoholics frequently drink to the point that they throw up anyway?

I absolutely concur with Budgie's post, and I have a couple other things to add. You may want to discuss major choice with academic or career advisers, especially if you're considering a post-graduate degree in aerospace engineering. The closest major to Aerospace is probably Mechanical Engineering, and there are a lot of mechanical engineers in the aerospace field. Do this now, while you're still a freshman and it's relatively easy to change majors. Also, go to a dedicated model and high-power rocketry forum like this one: http://www.rocketryforum.com/ The KSP forums have a lot of rocketeers, but for shear number of knowledgeable members we can't beat a rocketry-specific community. If currently you know next to nothing it may be best not to mention the goal of reaching space, though. They get a LOT of "I've never built or flown anything before, now how exactly do I go to space?" type posts. People will take you a lot more seriously if you mention a smaller goal like Level 1 certification. As far as access to materials goes, keep in mind that while up to a point rockets can be built cheaply with components like scrap mailing tubes and plywood, some components like motor cases, propellant reloads, and electronics, are very expensive. At certification level 3, costs will be several hundred dollars per flight, plus probably over a thousand for the rocket. An actual space shot will most likely have a budget well into the tens of thousands of dollars. It would be hard to get a college to pay for that - which is irrelevant because you'll almost certainly be out of college by the time you're ready for a project that size. It might be a good idea to consider various means of funding.

These three are good choices: all are DD-capable, fairly cheap and popular, and will fit inside most rockets. However, all of these are barometric-based. http://www.adeptrocketry.com/ADEPT22go.htm http://www.missileworks.com/store/#!/RRC3-Sport-Altimeter/p/25239688/category=5760485 http://www.apogeerockets.com/Electronics_Payloads/Altimeters/PerfectFlite_StratoLogger_Altimeter?cPath=52_192&zenid=ljs9o3ncrk6vrkjarpls2ibii0 This one is extremely compact, but a bit pricier. http://www.apogeerockets.com/Electronics_Payloads/Dual-Deployment/EasyMini?zenid=ljs9o3ncrk6vrkjarpls2ibii0 This one is very expensive at over $300, BUT it has integrated GPS tracking. http://www.apogeerockets.com/Electronics_Payloads/Altimeters/TeleMetrum?cPath=52_192& Finally, there's the Raven. This altimeter is absolutely tiny including the battery, and is very capable, including an accelerometer and four pyro outputs. This means that among other things it can be used to ignite the motor in a rocket's upper stage and still control the deployment. It also allows the user to set up backup deployment charges. However, it is fairly expensive. Unfortunately, I don't know of any accelerometer-based altimeters that are inexpensive. However, it's possible you could build one yourself with an Arduino or similar product. However, it would most likely not be as compact as any of the commercial options, so that's only a good idea if you have a relatively large rocket that will fit it. http://www.featherweightaltimeters.com/The_Raven.php

For an interstellar mission, there isn't a viable alternative to a sentient AI. There's no way non-sentient computer programs would be flexible enough to conduct a comprehensive exploration about an alien planet we initially know very little about with no human input, and waiting 8+ years for a reply from mission control if there's a problem isn't feasible.

I don't think any of the barge-landed stages would actually be flown again. More likely they'd do 2-3 landings on a barge a few miles offshore as a final stage of testing before landing on to a landing pad near the coast. And SpaceX can't afford lawyers? These are both multi-billion dollar companies. I believe NASA is paying SpaceX something like $400 million per crewed dragon flight, when their actual cost is supposed to be something like $150 million. Even if the legal fees reached a hundred million dollars, SpaceX could absorb that with the profits from a single CCDEV flight. Patent trolling is mostly effective because individuals or small companies can't afford a year-long legal battle even if they're guaranteed to win, so they can be intimidated into out-of-court settlements. If someone actually stands up to them they won't do so well. And I can't imagine Blue Origin actually getting a patent on landing a rocket on a boat upheld. Also, Blue Origin suing SpaceX would be PR suicide. Currently, SpaceX is very well known outside the aerospace field, but Blue Origin isn't. If they sued in the next couple years, they'd become world famous as "that company that tried to sue SpaceX over a frivolous patent when they don't even have a working spacecraft to land on their barge." Jeff Bezos might as well eat a baby on national television.

I know. What I'm saying is that a launch vehicle based off a single F-1B engine might also be effective.

Speaking of the F-1B: I calculated that a rocket with an F-1A first stage and a J-2 second stage could put a little over 20 tons into LEO. With an optional Centaur-D upper stage I imagine the payload to GTO or interplanetary trajectories would probably beat rockets like the Titan IV, D4H, Atlas V 551, Ariane 5, Proton-M, etc. Since we're already reviving the F-1 and J-2 for the SLS, why not use the F-1B and J-2X to develop a new heavy lifter? In addition to being an excellent launcher for interplanetary probes, high-orbit satellites, and space station modules, it would be perfect for the Orion on LEO missions (IIRC right now the only rocket that fits the bill is the Delta IV Heavy).

Yep. Pseudoscience from someone who doesn't understand how gravity works. It's about as legitimate as timecube.com.

Okay, all the force equations look mostly accurate, but thrust is actually a function of height, and the mass of the vehicle is Minitial - t * mass flow rate. I think you'll get something absurd like d2h/dt2 = (thrust as a function of altitude)/(Minitial-propellant mass flow rate * t) - (some constant) * (dh/dt)2 * e-h/H / (Minitial-propellant mass flow rate * t) - G*Mplanet/(r+h)2. This is a nightmarish differential equation, including one term with an awkward function of h(t) and t, one term with a nonlinear function of h(t), and one term with an awkward function of h(t) and t multiplied by a nonlinear function of dh/dt. I kind of doubt this is analytically solvable.

Holy cow! I haven't been following the development of Universe Sandbox the last few months, and look what happens! What's the Earth's axis tilt after the impact, by the way?

Here's what a 500 km asteroid would do: Ceres is about 8x the mass, so the damage would be even worse.

I can't remember, does the Roll Your Own configs spreadsheet use some sort of improved gimballing mod? I ask because I have a vague memory of using one in previous versions, and I can't seem to get the thrust vectoring working on any of my engines (tip: even without RO, a single 2.5m reaction wheel is not sufficient to control a 400+ ton rocket).

The computing would probably be possible. The issue is that it's going to be a VERY complicated and difficult mission. Most likely we'd start with a Venus flyby. Venus can actually give a decent gravity assist, and with repeated flybys can get a probe to Mercury or Jupiter (as was done with Cassini and MESSENGER). We'd probably want to encounter Mercury and Mars early in the mission, before heading for the outer solar system, which would make the trajectory more complicated and take longer. This phase will probably take 5-10 years, with low dV requirements. At some point, we get Venus and Earth to throw the probe out to Jupiter. At this point, a gravity assist from Jupiter can send the probe to any of the other outer planets provided they're lined up properly. With a Voyager style trajectory, we could complete the mission very quickly, but launch windows would be more than a century apart. Otherwise, we'd need to do a much slower outward transfer, because if we need to loop back around to encounter more planets we can't really do a super-fast solar escape trajectory. This also means the mission will take a VERY long time. I have no idea how long, but it could very well take 50-100 years. Even conventional RTGs will have trouble lasting that long, let alone other hardware. That will be quite the engineering challenge.

Well, the gravitational force between two point masses Fg= G*M*m/R^2. Divide that by the mass of the smaller object (say, a Minecraft character) to get the gravitational acceleration, and we get g=GM/R^2. Translating that into vector form, the gravitational field is g®= -G*M/|R^3| * R. (bold refers to vectors). This field is a vector with a vertical component and two horizontal components. Since the plane is infinite, the horizontal components end up cancelling each other out. Let's solve for the y component of g (the y axis is vertical in the image you posted). We'll assume the point where we're solving for the gravitational acceleration is a height h above the object and a horizontal distance r away from it. Therefore, R = sqrt(h^2+r^2), and Ry (the y component of R) is equal to h. So, we get gy = -G*M*h / (h^2+r^2)^(3/2) Now things get tricky. Each of those point masses we'll assume is a differential element with a volume of dV and a density of p. This calculation will be easier with cylindrical coordinates. We'll let each of our differential elements be between radii dr apart, across an angle dtheta, with a height of dy. Its volume dV will therefore be r*dr*dtheta*dy, and its mass dM will be p*r*dtheta*dy. At this point, we'll put the surface of the world at y=0, and our object a distance of y1 above it. Our variable "h" is now y1-y where y is the y position of our differential mass. We plug that into the gravitational field expression to get: dgy = -G*dM*(y1-y)/((y1-y)^2+r^2)^(3/2) dgy = -G*p*(y1-y)*r/((y1-y)^2+r^2)^(3/2) * dr * dtheta * dy. (don't you just love typing calculus?) To make our calculations a little easier, we'll assume y1=0, i.e. we're just solving for the surface gravity. Now: dgy = G*p*y*r/(y^2+r^2)^(3/2)*dr*dy*dtheta. (the negative ys cancel out). To make things easier, we'll use cylindrical coordinates, and model the Minecraft world as an infinite plane with a thickness of 256 m. We're going to integrate this from theta=0 to theta=2pi, from y=0 to y=-256, and from r=0 to r= infinity. This will not be easy to type, but basically we get an answer of gy = -2pi*G*p*256m^3. As it turns out, depending on the density of the material, this is probably between 10^-5 and 10^-4 gs - basically zero. With a density of 5000 kg/m^3, to get a gravity of 1 g, the infinite plane would have to be about 4700 km thick.

Yep. The moons of some of the giant planets originally had roman numeral designations (e.g. Jupiter III is Ganymede) but those aren't really used anymore. The named moons are just referred to by their names, and the unnamed ones have numerical designations based on the date of discovery among other things. Earth, incidentally, would probably be called Sol b. The others were known since antiquity, but Earth was known about long before humans ever thought of systematically looking at the sky and realized some of the stars moved. As far as recognition of a planet goes, the other planets weren't really recognized as planets by a modern definition until about the same time the Earth was, i.e. when the geocentric cosmology was falsified. Pretty much, but stars are referred to with capital letters (which aren't normally used for a single star). For example, Alpha Centauri consists of Alpha Centauri A and Alpha Centauri B. I think the designation is in order of mass, but it might be in order of discovery (those usually match up anyway). Planets extend the classification. For instance, Kepler 16 includes Kepler-16A, Kepler-16B, and Kepler-16b. This seems a little odd; I'd think they'd call it Kepler-16c to avoid confusion. With planets orbiting one of the stars in a binary, the planets have the star's name, but extended. For instance, Gliese 667 Cc is the second planet discovered orbiting Gliese 667 C. Presumably exomoons would continue the same classification scheme. For example, Earth's moon would be Sun bb.

Unfortunately, the first bit is almost completely wrong. Even the hotter and more compact blue supergiants and hypergiants are far less dense than air at sea level on Earth. In general, stars, including our sun, don't have a defined "surface" any more than gas giants do. According to wikipedia: Basically, the plasma just gets denser and denser until we define it arbitrarily as opaque. It's not really a surface.

Something I'm curious about: just how big are the explosive charges used in a flight termination system? I know that the rocket completely disintegrates, and in vehicles with SRBs fragments of propellant are flung away from the explosion at fairly high velocity, but how much of that is from the charges themselves and how much is from the fuel all burning at once / being flung away by its own combustion pressure when the tanks/casing are opened? To put it another way, if a rocket was sitting on the ground with its tanks full of something inert like water and the FTS fired, how much damage would it actually do?

There's some hardcore necromancy in here, but it is actually continuing the thread so maybe the mods will let us keep it open. The short answer is "yes, but it's unlikely." As I said in my previous posts, gas giants really need to be at least three Jupiter masses to support a moon that could sustain Earthlike surface conditions. A 13 jupiter mass object at the boundary between a planet and a brown dwarf would probably have at most 1 earth mass worth of moons. This means that in theory it could have three habitable moons, but in reality often you'd get one habitable-sized moon and a bunch of moons too small to support life. Another problem is the moons would have a limited habitable zone around the gas giant. For instance, young gas giants are extremely hot; the hypothetical planet Tyche was predicted to have a surface temperature of 200 K, or hotter than Jupiter, despite being in the Oort Cloud and receiving virtually no energy from the sun. That's a 4 MJ planet at 5 billion years old. A 10+ MJ planet could very well put out enough radiation in the first few hundred million years of its life to boil away the atmospheres of close-in moons. Brown dwarfs would be even worse; in addition to their greater mass, they'd get an extra burst of initial heat from deuterium fusion. You also may end up with a goldilocks zone of distance from the planet where the tidal heating and radiation levels are suitable for a moon to retain a habitable temperature and atmosphere. Finally, remember that just because a planet is of earthlike size and composition and located in the habitable zone it isn't necessarily actually habitable. However, it would be quite possible for a planetary system to have multiple habitable worlds. This article: http://www.newscientist.com/article/dn25653-ultimate-solar-system-could-contain-60-earths.html#.U_UT3GOHOhQ claims that a single star could support 36, but this is unfortunately bogus for a variety of reasons, namely that as mentioned five habitable moons around a single gas giant ain't happening, and if you packed four superjovian planets into a habitable zone they'd probably destabilize each other's trojan points, if not the planets themselves. Nevertheless, you could probably fit a total of four habitable moons into a system if you were very lucky, and maybe a couple trojan planets. With much smaller gas giants, too small to have habitable moons, you could probably still fit a few planets at the trojan points.

That, and the tidal forces would rip apart the crusts of both planets.

The issue is, every time the system flings something into orbit you take angular momentum from the tether and transfer it to the spacecraft, so the tether's rotation slows down. To spin it back up, you'd have to increase the speed of the reaction wheel every time you launched, which would quickly saturated. Reaction wheels are good for satellites that need to make tiny, very precise adjustments to their position without using propellant, not so much for something with a whole bunch of momentum. It's better to throw out the reaction wheel and use whatever system you're using to desaturate the wheel to directly boost the tether's speed. Fortunately, because the tether is really big and can spin back up over the course of several hours or days, it can use propellantless or high-efficiency drives that ordinarily would be too low-thrust for orbital launches. One thing to note: if the tether is changing the payload's velocity by 5 km/s (enough to reach LEO from roughly mach 10) the tether would need to be 1000 km long to get an acceleration of 5 Gs. With a dV of 8 km/s (good for Geostationary Transfer Orbit) it would have to be 2560 km long.

It's a textbook case of "Vehicle designed by business majors and graphic artists" syndrome.

Like this?

Well, basically the smaller the molecular mass of the exhaust gases, the faster it will move for a given temperature and therefore the more efficient your engine will be. With hydrogen, the exhaust is a mixture of H2 and atomic hydrogen, both of which move very fast. Likewise, methane and ammonia partially decompose, releasing some fast-moving hydrogen atoms that boost the ISP. Water doesn't work as well because oxygen forms stronger bonds with hydrogen than carbon or nitrogen do. In nitrogen, the two atoms are linked by a very strong triple bond, so it barely dissociates at all. This leaves you with an exhaust stream of nothing but diatomic nitrogen, which has a rather high molecular mass. That would actually be quite effective, particularly if the empty hydrogen tanks were either jettisoned or refilled with methane manufactured on Mars.

According to this: http://www.projectrho.com/public_html/rocket/enginelist.php#ntrsolidcore An NTR running on methane would have an ISP of around 640, which is about the same as for a LANTR (Lox-Augmented NTR) running on hydrogen. Carbon dioxide is a bad propellant for an NTR designed for hydrogen or methane, since at high temperatures it decomposes and produces oxygen, which as you mentioned reacts with everything. Nitrogen is a monumentally awful propellant, with an ISP of 270 - worse than a decent solid-fuel motor! Our LV-N has an ISP of 800, which is about right for a basic NTR design running on LH2.