sevenperforce

30% of the payload is not much in comparison to the rocket. Falcon 9v1.1 massed 506 tonnes on the launch pad with an LEO payload of 13 tonnes. 30% of the payload is just 0.77% of the mass of the loaded rocket. Good luck figuring out a way to put wings on a booster for under 1% of launch weight. Landing legs have the support the same weight regardless of whether the booster is vertical or horizontal; the booster doesn't magically lose mass in the horizontal position. Landing gear for a horizontal landing requires the entire booster body to be strengthened to support radial stresses, while landing on the tail distributes stress axially, which is the direction the rocket is already designed to handle stress in. Finally, SpaceX uses no "mechanism" to balance the rocket; rather, they simply use the existing RCS cold gas thrusters required for attitude control during launch.

Untested crossfeed? Odd that you would cite the SSMEs as better than alternatives while claiming crossfeed is untested when the Shuttle used crossfeed. A large truncated annular aerospike offers a large surface area for a PICA-X heat shield to protect the booster during re-entry. There is no need for jet engines. Use an engine cluster and RTLS for a tail-first landing on the central engine. Make the central engine underexpanded if you desperately need hovering capability. The core booster will have a rougher re-entry due to its high velocity but staging gives us a wide fuel margin. Wings require landing gear, which weighs more than landing legs. Wings require a runway, which requires additional fuel for maneuvering.

I'm not suggesting we develop carbon nanotube weave merely to build vacuum airships. Carbon nanotube weave will be developed for numerous applications as soon as it can be. Everybody needs that stuff. Okay, so on certain scales, it is not possible to determine the position and momentum of a particle simultaneously. The uncertainty in the product of position and momentum is approximately equal to the Planck constant. This is because particles are actually wave functions and exist across all space with a defined expectation value. Or were you talking about Walter White?

Well, existed, more than exist. But yes, you can compensate for these things in an SSTO. I was just pointing out that these are automatically provided for with staging, but need to be specifically provided for in SSTO designs. Oh, I had no intention of actually burning LH2 in the strap-on boosters. Those tanks are there solely to feed the core rocket, allowing it to have a full tank at staging. Hydrolox has such high tankage volume that dropping tanks is extremely advantageous; using the side boosters to carry the drop tanks back for reuse should offer a substantial advantage. I'm not sure whether an SSME nozzle or an annular aerospike would be better for powered tail-first landings. An annular aerospike is more easily throttleable. Technically, winged autonomous landings of spaceplanes are 1 for 1 (Buran), just like tail-first autonomous RTLS landings of orbital-class boosters. And no, the core booster definitely wasn't expendable. If I was going to go expendable on anything, I would just say to go Falcon 9. The point was that rapid-turnaround 100% reuse could be readily realized using parallel staging, with almost all the same advantages as SSTO and none or few of the detriments. Assuming materials of equal strength are used, landing legs will always be substantially lighter than landing gear. Landing legs must bear the weight of the vehicle; landing gear must bear the weight of the vehicle AND transfer that weight to an axle AND to a set of wheels, all of which are heavy.

By adding wings that weigh twice as much? Sure....

No one suggested it was. But alleging that the suicide burns are the source of their difficulty when in fact stability was the source of the two normal-condition landing failures is quite incorrect.

Indeed. Though not quite the exact same. You don't need those useless fins, the center booster needs to be altitude-compensating with a truncated aerospike to act a heat shield, and you need to have LH2 and LOX crossfeed from the side boosters to the center booster (which themselves need to be kerolox). And everything needs landing legs. You could do it with a single booster, I suppose, but then you do end up with some COM/COT issues.

They could put a massively underexpanded Merlin 1D in the center of the cluster, but it would result in a performance hit. And honestly they haven't had trouble with the suicide burns. The first two barge landing failures were tip-over, indicating stability problems. The last attempt, a three-engine suicide burn, had a vanishingly low probability of success from the very beginning, made harder by poor landing conditions. They haven't actually attempted a normal-condition barge landing with the F9FT yet; both tip-overs were on the F9v1.1 model.

The Shuttle is actually not too far from the optimum layout for getting to orbit and back with some degree of fast and simple reusability. It was just way, way too big to be reused quickly, and the requirement of an internal cargo bay meant that an external disposable drop tank was unavoidable. If your goal is 100% reuse with short turnaround time, then your best bet is to build a VERY downscaled version of the Shuttle. Replace the SRBs with Falcon-Heavy style kerolox boosters, but add auxiliary LH2 tanks. Replace the Shuttle with a core hydrolox booster and give it a truncated aerospike nozzle for altitude compensation. Put a Dragon V2 style capsule on top (or a payload with a fairing, if you like). At launch, crossfeed the core booster from the side-booster LOX tanks and the auxiliary LH2 tanks up to side booster separation. Side boosters RTLS with vertical landing; the core carries the capsule into orbit, circles once or twice, and then deorbits, using its truncated aerospike as a heat shield to aerobrake to a vertical landing RTLS. Capsule (if applicable) re-enters separately and lands propulsively. On the one hand, that's four separate vehicles instead of just one. But any problems or extended repair time is now relegated to a single component, so that re-use of the other components is unhindered. With the extremely good specific impulse of hydrolox and the very high thrust of kerolox, the propellant mass fraction would be extremely low, meaning the launch vehicle could be quite small and thus easier to recertify and reuse.

Staging simply offers a LOT of advantages, not all of which are readily apparent. You can use sea-level-optimized engines at sea level and vacuum-optimized engines in space. Perhaps even more importantly, you can have a lot of thrust at launch and less thrust up high. Hydrolox offers by far the highest specific impulse of any reasonably fuel combination, and mass ratio is insanely important for an SSTO, so let's use hydrolox. You need about 7,500 m/s of dV to get into orbit, so that's a propellant mass fraction of 84%. Totally manageable, right? Unfortunately, it's not that simple -- you need to account for the lower SL specific impulse; even if you have an altitude-compensating nozzle of some kind, you're only going to get around 400s max at sea level compared to hydrolox's 450+ second vacuum specific impulse. So that tacks on the equivalent of at least 500 m/s, bringing your dV up to 8 km/s and your propellant mass fraction of 86%. Still feasible, right? But there's something else to account for: drag. Gravity drag and aerodynamic drag. And here's where the lack of staging really kills you. Hydrogen engines have high specific impulse but crappy thrust, which means they are going to have far worse susceptibility to gravity drag than a beefier fuel combination like kerolox. A kerolox engine might need an extra 1.5 km/s to fight drag going up, but a hydrolox engine will need closer to 2 or 2.5 km/s to fight drag, bringing the propellant mass fraction up above 90%. And your hydrogen engines will need to be heavier than a kerolox engine of equivalent thrust, which will probably end up cutting your payload in half or worse. Couple that with hydrogen's low density and high tankage volume, and you'll REALLY be wishing you had a kerolox engine and a kerolox tank strung alongside to help lift you the first half of the way. But if you have a kerolox engine and a kerolox tank strapped alongside, then it really only makes sense to drop them at some point so they can RTLS for reuse while your hydrolox stage rockets into orbit. 1.5 stages to orbit is SO much simpler than single-stage-to-orbit.

Yeah, he's a nice guy and all but I like Nye a bit better. Nye had the balls to take on Ham and his ilk, after all.

Of course, the kinds of advances which we would need to make in order to produce force fields would almost certainly be preceded by the development of macroscopic graphene structures and other materials capable of supporting a vacuum airship under tension. Wrap a geodesic sphere in fabric made from woven carbon nanotubes and you will have no difficulty pumping the air out of it entirely.

Open-cell vs closed-cell forcefield problem.

Yeah, I know, right? It really sucks when you can know the location of your airship but not its velocity, or the velocity but not the location. If we ever figured out lightweight forcefields then building a vacuum airship would be a snap.

The flexible skin probably resists puncture even better than a rigid one.

It's not so different from any other vehicle. The larger it is, the more opportunity there is for stuff to break, and the more refurbishing and repairing has to be done between flights. But since large rockets are more fuel-efficient than small rockets, a reusable SSTO must ride the knife-edge between "too small to deliver a meaningful payload" and "too large to allow reasonable turnaround time".

It would be a very bad day to be sure, but it could happen. If you were inside an orbital module and suffered damage from a micrometeoroid strike which punctured the module and damaged your re-entry capsule's heat shield, then it would be worth knowing whether de-orbiting with your capsule inside the shell of the dead orbital module would give you a better chance of survival or a worse chance of survival. Or, what if your RCS system was completely destroyed and you had no way of decoupling from the module in-flight? Would it be safe to let the module push you onto the right trajectory and then just allow re-entry to burn it off? Obviously that is an extreme case and we may already have the answer in that particular situation. But space is a dangerous place and a lot of stuff can happen. I wouldn't ever want to give up the opportunity for more data. I don't mean re-entry aerobraking; I mean skip aerobraking. If a flight anomaly or other emergency on an interplanetary flight (to or from Mars, for example) left you with excess velocity and not enough dV for a complete orbital insertion, how far into the atmosphere would it be safe to dip in an attempt to bleed off that excess velocity? We don't model re-entry disposal so much, but we definitely model re-entry in general. And being able to extend the modeling software's accuracy range to include extreme scenarios is a good test of your assumptions. EDIT: Sniped by fred; he said exactly what I was thinking.

Well, I don't have a source; that's why I said "I bet". But it makes sense. That cylindrical ring up at the top of the New Shepard with the deployable air brakes is just about the same size as the pusher motor shown underneath the capsule in the pad abort test. And you'd want control surfaces in an abort. If you are aborting away from the booster then you are no longer worried about landing the booster.

I would have thought so too, but apparently not. You know, I bet the pusher abort system remains inside the top of the New Shepard booster during normal flights.

Yeah, you're right. Watched the launch abort video. In the text at the start they expressly say it uses a "pusher escape motor" for launch, and then on landing there was the same characteristic retrorocket burst. Dunno if these recent tests have been with the pusher escape motor or not...nor whether the pusher escape motor is nominally attached to the booster or to the capsule.

Any sort of ablative or single-use or otherwise expendable re-entry system can benefit from the data that would be gleaned when the ISS burns up.

Does the capsule have landing legs? If so, BO could make them out of a steel honeycomb that would crumple at faster impact speeds; this way the capsule could use its retrorockets for launch abort and then smash its legs to pieces on the landing. You don't need to worry about reuse in a launch abort.

Did a bit more digging, and noticed that the BE-3 is going to be getting a vacuum-optimized bell projected at 670 kN some time next year. That's a great deal more than the 490 kN of the SL version. They can downthrottle to 18%, though, which is odd. How can they get that low without flow separation? Here's a pretty good view of the BE-3's nozzle: It seems more conical than the conventional bell shape of the Merlin 1D: So maybe they are using a differently-shaped bell to allow deeper throttling at the expense of lower specific impulse and some underexpansion? Then again the flow doesn't look terribly under-expanded in videos, so I'm not sure. If we could be sure that the current BE-3 was properly SL optimized then we could get a good idea of specific impulse by comparing the projected vacuum thrust to the SL thrust, but we aren't.

Are you really suggesting that there's absolutely no way we could glean useful data from a better understanding of how modules intended for on-orbit endurance handle unprotected re-entry? After all, it's not like space flight ever involves unplanned emergencies. Bare minimum, using a video feed from the deorbit of the ISS would provide useful data for determining how aggressively we could aerobrake an unprotected orbital module in an emergency situation. But it's bigger than that. We can model a lot of stuff, but we also need to be able to test our simulations. Having real data from the breakup of multiple ISS modules would provide a huge resource for testing our modeling software. If our software gives predictions which diverge wildly from the actual sequence, then that's something worth knowing. We may never need to exactly predict the breakup altitude of something the exact size of the ISS again, but that doesn't mean we won't be simulating breakups for other purposes. For example, what if ten years passes and we find ourselves with a need to controllably return sturdy payloads from orbit to Earth? Is it inconceivable that spent tanks or disposable modules, with their low mass and high surface area, might be used to "bleed off" the first few km/s of orbital speed, allowing the heat shield to be that much smaller? There are so many reasons why this sort of data might be useful.

They have tested them before -- not sure whether they used SpaceX for those launches or not -- but they have not tested them as add-on modules for the ISS before. And obviously they have not had people in space anywhere near them.