sevenperforce

To the re-entry question, again.... IIRC, the Shuttle's external tank was so lightweight and yet so large that when it hit the atmosphere, it didn't immediately burn up. It would tumble, efficiently dissipating heat due to its high drag coefficient and low mass, until aerodynamic forces broke it up, and then the pieces would burn up. The larger the surface area you can expose, the better. Going in nose-first would produce far more heating and a much nastier deceleration. EDIT: I kept the thrusters behind the landing panels to protect them from re-entry -- I can't imagine that plasma is healthy for engines -- but you never know. There might be a way to have them exposed but not directly in the plasma stream, which would allow them to be used for RCS control as well. EDIT 2: Note that with the first stage, the landing thrusters can be used to give an extra kick to TWR off the pad. They won't have as good of specific impulse as the main engines, but for very large payloads it might be a good tradeoff. EDIT 3: You can also use a larger cluster of dev Raptors on the first stage in place of the two full-size SL Raptors and maybe be able to land on the center one, but they won't have as good of a TWR or isp. I like the idea of using a smaller version of the same engine on the upper stage. It should be noted that the dry mass of the first stage is actually significantly lower than the dry mass of the Falcon 9 first stage, due to the use of a slightly greater diameter to hold more propellant for less mass, and the use of composites. It would be very easy to move and transport and still within road-transport size.

Actually, the biconic entry is far, far more mass-efficient. A larger surface area will encounter drag at a greater altitude and will be able to use hypersonic lift, dramatically reducing g forces and peak heating. Fuel is cheaper, mass-wise.

Thanks! Right now it would cannibalize Falcon 9 terribly, but it might be something they'd look at in the future. I'm really a sucker for horizontal, Star Wars style VTVL, so someone else might not have come up with something like this. Here's an underside render so it makes a little more sense:

If SpaceX's plans for the Mars come to fruition, SpaceX would eventually want to transition Falcon 9 and Falcon Heavy payloads to a Raptor-based architecture, to enable access to space for payloads smaller than the full ITS/BFR/BFS capacity. On the flip side, if the Mars colonization plans never quite pan out, SpaceX will still need a use for their methalox engines. So either way, we need a Raptor-derived fully-reusable TSTO with payloads roughly equivalent to the Falcon family. It will need to be man-rated, too, since there will surely be a need for sending passengers to orbit in numbers lower than the 100+ capacity of the ITS/BFR/BFS system. One problem with this is that a single-engine upper stage has a TWR too high to use for propulsive landing, even if it wasn't overexpanded at sea level. Thus, it needs auxiliary landing thrusters. Another problem is re-entry; an unmanned stage can come back using a heat shield on its nose, but that's not much fun for passengers, and I have a strong preference for a "true" TSTO where the crew cabin is integrated. You need biconic re-entry a la ITS. But if you're already using auxiliary thrusters and biconic re-entry, you don't necessarily have to align the aux thrusters with the main engine vector. That's where things get...interesting. Here is a line drawing and a very very rough sketch of my concept: Looks a bit like the ITS, doesn't it? Same basic principle (composite monocoque tanks, etc), except the diameter is only four meters, making it roughly the same size as the Falcon 9 but slightly wider. The first stage has two full-size SL Raptor engines for launch and boostback and six methalox hot gas SL thrusters for landing, along with four landing legs: The first stage has a dry mass of 17 tonnes and a propellant capacity of 421 tonnes; it delivers the upper stage at a notional staging velocity between 1.5 and 2.5 km/s and executes a boostback RTLS landing. Minimum initial TWR for the boostback burn is 2.7:1 with both Raptors at minimum throttle; maximum landing TWR on thrusters alone is 3:1 but it can easily hover. I've factored in the masses of the thrusters and everything else. The upper stage is where it gets really interesting. Rather than using Raptor engines, which would be way oversized, it uses a pair of the Raptor Development engines (1,000 kN SL thrust) with vacuum nozzle extensions. I'm estimating their mass at 638 kg each. Total stage vacuum thrust is 2,292 kN. Dry mass is 6.6 tonnes and propellant capacity is 141 tonnes. Because the vacuum engines cannot be used at sea level, I gave the upper stage eight SL-expanded methalox thrusters in addition to its vacuum-optimized RCS thrusters, with a combined SL thrust of 688 kN. But I didn't want to cluster them around the devRaptors in the tail, both for space considerations and because of damage to the engine bells. See the wing extensions shown in the above line drawing? The landing thrusters are placed underneath the wing extensions, pointing down. For re-entry, the upper stage enters biconically, on its belly. It then glides/falls to the landing site before hydraulically-actuated panels open up underneath the wing extensions, both exposing the landing thrusters and providing rear "legs" for the vehicle to land on, so it lands vertically but in a horizontal attitude, eliminating the risk of tip-over. The landing would look like something out of Star Wars, because it drops, winglets open, and it lands on the wingtips with rocket propulsion. Based on my simulations using this calculator, the launch system could deliver up to 6.8 tonnes to GTO with full reuse or up to 24 tonnes to LEO with full reuse. For LEO launches, the upper stage can also recover up to 30 tonnes of downmass. This is obviously plenty of margin to have a crewed version, which would use the same tank and body as the rest of the orbiter but have a crew cabin in place of the cargo bay. Payload capacity is high enough that the crew cabin could carry at least a dozen crew members plus unpressurized cargo and still have independent LES and re-entry capability (lifeboat). Know what else is great? Due to the vertically-oriented thrusters, the upper stage could both land on and take off from the Moon or from the surface of Mars without needing a launch pad. On Mars, it would need to be refueled on the Martian surface; the lower gravity means that the thrusters have enough thrust to lift it off the ground so the main engines could be fired up. For lunar missions, simply being refueled once in LEO would give it ample dV to fly to the moon, land, SSTO, and return to LEO.

Yes, that's entirely possible. But my point was that collecting propellant on Mars via ISRU and transferring it back to Earth orbit in a tanker might be more efficient than simply launching it on Earth.

Then you can say that the KORD's system settings were wrong. Which is what I meant by programming. Or they could have set up the gyros to NOT send the LES activation signal under those circumstances.

Random thought. Once the ITS system is up and running, it may actually prove cheaper (dV-wise) to send a fleet of tankers to Mars and use ISRU there to provide transfer fuel. You only need a fuel fraction of 65.7% to SSTO from Mars (including enough residuals for landing), meaning that a single, full tanker can lift 888 tonnes of fuel into Martian orbit via SSTO, more than twice as much as an Earth-based tanker launched by the ITS booster.

Methinks that was the joke.

While I have no argument with the rest of your analysis, I'm pretty sure this is incorrect. Falcon 9 v1.0 could only launch 4.5 tonnes to GTO, expendable. Falcon 9 FT sent SES-10 to GTO, with recovery, and that bird was over five tonnes.

"Prepare for worp speed!" "Warp speed?" "No, not warp; worp. We need to change the argument of our periapse."

I get the idea of why you'd want to use "posigrade", since "prograde" and "retrograde" do double duty in identifying orbits which match or oppose the angular momentum of basically everything else in the system. Ideally, we would use "posigrade" and "antigrade" for thrust vectors while leaving "prograde" and "retrograde" to identify orbital directions. But "retrothrusters" would then make no sense, so it's probably better to just use context and be done with it.

Even with all their resources, the Soviets didn't have nearly enough capacity to do multiple back-to-back launches. Neither did we, for that matter. Nowadays, EOR would be much simpler. Or, my favorite: double lunar orbit rendezvous. Launch the lunar lander into lunar orbit unmanned with one HLV launch, then send the crew and return capsule into lunar orbit with a second HLV launch. They can dock once, go through the same Apollo-style mission, then come home in the original capsule just like Apollo did. Somewhat amusingly, the repeated failures of the N-1 program were almost entirely the fault of bad programming. The computer control system for the engines on the first stage had poor sensors but poorer programming, so it kept shutting down before it needed to. If they had devoted more time to failure tree modeling before launching, they would have caught the errors, and N-1 might have reached orbit. The Soviets have a poor track record in that regard. My favorite programming failure (is it nerdy to have a favorite programming failure?) was Soyuz 7K-OK No.1, where they aborted launch before liftoff but were taken by surprise half an hour later, when the LES spontaneously triggered, sending the capsule flying and setting the rocket stack on fire. Turns out the programmers had forgotten to account for the rotation of the Earth; in half an hour, the Earth rotated about eight degrees, triggering the gyros that are supposed to activate LES if the rocket tilts too far on liftoff.

The OP specified total reuse, which means the payload fairing either needs to be incorporated into the second stage or independently recoverable. One consideration is that for a man-rated RLV, you have a lot of additional constraints. If you want commonality between your manned and unmanned launches, you'll need to factor at least some of your man-rating designs into your unmanned design. Note that SuperDracos are not necessary; if you have Raptor engines, you can use the pressurized intertanks to run a few of the hot-gas oxy+methane thrusters intended for the ITS Spaceship RCS/OMS system. They come in at around 86 kN SL thrust and 93 kN vacuum thrust; we can ballpark their TWR at around 190:1 for a dry mass of 45 kg each. Specific impulse is probably a little lower than Raptor.

Oh, tremendous. See, I've been meaning to make precisely this same thread for a while now. I've got a pretty neat design I've been kicking around, so I'll run all the numbers and do a lineart render later today...still excited to see what everyone else comes up with, though!

It can land upright in a static cradle with vertical arms. Pinpoint landing isn't a problem, never has been. The cradle would have wide open spaces for the exhaust plumes to exit through. EDIT: Wait a minute. If it can RTLS and control its descent with grid fins enough to pinpoint a landing, it can literally just land in a net, suspended above the ground. No SuperDracos required. Terminal velocity would not be very high. Case in point: https://www.youtube.com/watch?v=6qF_fzEI4wU

Honestly, you wouldn't even need attachment points. The second stage is short enough that as long as it slides in, it's okay if it lists a little bit.

The RoombaX is intended to secure the first stage on the droneship, not catch it. It wouldn't survive the engine wash. Might be different for a second stage, particularly if it had hover capability with a set of SuperDracos, but it would be tricky. Would potentially work for a RTLS but not for a droneship. A cradle with fixed vertical arms over a blast trench, maybe.

Another idea that is particularly promising... Autorotation. Terminal velocity on the first stage is (at present) slightly subsonic. I'm guessing it's somewhere around 300 m/s. The first stage masses four times the second stage, but since grid fin drag and pressure drag are the majority, this suggests terminal velocity that's roughly a quarter of that, or 75 m/s. It would actually be slightly higher because of the square term in the drag equation, but the new grid fins are larger so that's probably not much of a difference. The new grid fins are supposed to be able to provide a 1:1 lift-to-drag ratio on the first stage. And that's where things get interesting. L/D on the second stage with the same size grid fins would be on the order of 4:1 or maybe even 5:1. So if the second stage was allowed to autorotate, it could conceivably get enough lift to reduce impact speed to something on the order of 15 m/s, which might be bouncy-castle-survivable.

The consumables tanks could probably be much smaller, but getting fold-out landing legs requires a lot of space.

Glad we could help! Keep in mind that even though a higher orbit is slower than a lower orbit, you have to accelerate twice in order to reach it -- once in order to move into an oval-shaped transfer orbit, and second to circularize at the higher orbit.

Following up on this, consider what has to go into a "full" recovery module adapter: Heat shield Payload attachment points on/around/under heat shield Thrusters Propellant tanks Pressurant tank(s) Landing legs There's a chance you could pressurize the legs and the thruster prop tanks from the main-stage pressurant COPVs, but that would require re-plumbing of the main tank, and I'm not sure the Falcon COPVs have enough helium pressure to successfully run pressure-fed thrusters. The SuperDracos have a chamber pressure of 6.9 MPa. Consider also that the thrusters and the landing legs both need to be outside of the re-entry plasma impingement region. For the thrusters, this probably means being set significantly back inside the module with associated cosine losses. The landing legs are also problematic, since space is really limited. Using the Dragon 2's pop-out landing struts is not ideal due to the much greater height of the second stage; even a slight puff of wind would tip it over. The recovery module is also going to increase the effective height of the stage, significantly. Geometry is a limiting factor, here. EDIT: For example, here's what the difference might look like... First from left is the current Stage 2. Second from left would be the second stage with recovery adapter and flaps added. Third shows recovery adapter and feathered flaps on re-entry; fourth shows landing burn (high cosine losses) with legs extended. Obviously it would be balanced; this only shows one thruster and one landing leg for the sake of simplicity. The tanks shown are helium pressurant and hypergolic bipropellant.

It's possible, but the second-stage engine is a LOT more fragile than the fairings are, so I'm unsure about that. As everyone else said, the SuperDracos are heavier and require toxic fuel. They also aren't really needed. The Falcon 9 first stage uses its engine gimbals for pitch, yaw, and roll control on ascent; the cold gas thrusters are only used to adjust attitude/roll when the engines are turned off, and this requires only a very minimal amount of thrust since they are operating in a vacuum. Splashdown wouldn't destroy the stage, but the saltwater getting to the engine would ruin it. Even if there were airbags or flotation bags, waves and spray could still saturate the engine before it could be recovered. The engine bell would not be able to act as a shuttlecock; aerodynamic loads would rip it apart. There's a reason it launches inside the interstage. However, there's a SpaceX rendering that shows panels around the engine which could conceivably pop out and act as a shuttlecock: A "recovery module" would be straightforward enough, but fitting all the required parts into the form factor seems like it would be a pretty big challenge.

We talk often about "orbital velocity", usually in reference to the velocity of a spacecraft orbiting in a circle just above the atmosphere (about 7.8 km/s or 17,000 mph), but there is a different velocity for every possible orbit. In order to move from orbit to orbit, you need to either increase or decrease your velocity, which requires you to fire your engines. The greek letter "delta" is used in physics and mathematics to refer to a change in some physical quantity, so "delta-velocity" or "delta-V" is used to represent the change in velocity (e.g., between two orbits). But since we don't have infinite fuel, we can't just fire our engines whenever we want for as a long as we want. Just like in a car, you need to keep an eye on your fuel tank. But how far will a given amount of fuel get you? Fortunately, you can use an equation (the Tsiolkovsky ideal rocket equation, or rocket equation for short) to figure it out. Plug in the mass of your spaceship, the total amount of fuel, and the exhaust velocity (also called specific impulse) of your fuel, and the rocket equation will tell you that if you burned all your fuel, it would change your velocity by some specific amount. And remember, a change in velocity is also called...delta-V. So if your fuel tanks give you the ability to change your velocity by 1,000 m/s, then you can execute any maneuver that requires you to change your velocity by 1,000 m/s or less. The nice thing about this is that you don't need any complex math to figure out how far you can get. If you need to add 50 m/s to your orbital velocity now, and 100 m/s to your orbital velocity later, then just make you have at least 150 m/s of total dV in your tanks. Just add it up. Note, however, that you can't subtract. If you need to add 50 m/s now and then subtract 100 m/s later, then you'll still need 150 m/s of total velocity change; your engines don't care whether you're speeding up or slowing down. Because the delta-V value is so useful, it is used for other things as well. If you are taking off from the ground and going to space, then you will experience drag due to friction with the air. Rather than having to calculate the forces on your rocket and subtract them from your engine thrust for every point along your ascent, you can just determine that the total air resistance will rob you of about 750 m/s of delta-V. In other words, you burn 750 m/s of dV in fuel just fighting air resistance on the way up. Well, the change in velocity you need to move between two different orbits doesn't depend on the size of your ship at all; it's just based on orbital mechanics. The mass of a satellite or spacecraft doesn't affect its orbital velocity; and good thing too. That's how a small capsule is able to dock with a much larger space station: they both have the same orbital velocity and orbital location, even though their masses are different. So a big ship needs the same change-in-velocity to move to a different orbit than a small ship does. However, the bigger ship will need more fuel. If the two ships have similar engines, it's fairly simple: both ships will burn approximately the same proportion of fuel to achieve the same change in velocity. For example, if the small ship masses 10 tonnes and needs to burn 2 tonnes of fuel to achieve a certain change-in-velocity, and the bigger ship masses 100 tonnes, then the bigger ship will need to burn 20 tonnes of fuel to achieve the same thing. Note that if one ship has better fuel/engines (e.g., liquid kerosene on one ship and liquid hydrogen on the other ship), the ship with more energetic fuel won't have to burn as much of it for the same difference. We often think of gravity as a force, but gravity is actually a field which permeates and bends space. Just like a surfer can ride a wave in the ocean, a rocket's exhaust can "push against" a gravitational field in order to increase the change in velocity. The closer you are to a planet or other massive body, the more gravity there is, and the more of a push you can get. As a result, if you're in orbit around a planet, it's always better to make your biggest maneuvers as close to the planet as possible. It gets better -- if you are coming from far away from a planet and passing very close to it, you can actually increase or decrease your velocity without burning your engines at all, just like a surfer can get a "kick" from the wake of a passing speedboat. However, this only works if you are passing by the planet; it doesn't work if you are already orbiting the planet. It's as simple as that!

I've seen no indication that SpaceX has considered air capture. Splashdown will give more than scrap metal if chuted; the second stage has plenty of buoyancy. Would probably hose the engine, but it could be a first step toward guided recovery. Going with a "recovery payload adapter" makes a lot of sense if they are willing to do all the design work...but it is a LOT of design work.

My table uses fuel fraction and fixed vehicle TWR to estimate component dry masses as a percentage of GLOW, so the actual mass of a given engine is not really factored into the equation. There's a discrepancy between what is listed on astronautix and what is listed on Wikipedia; I was going by the latter.