sevenperforce

There was no probability of the orbiter surviving an SRB failure during the SRB burn.

Liquid rocket stages do a surprisingly poor job of exploding in flight, and particularly out of the atmosphere. Even in Challenger, when the entire external tank ruptured, mixed, and then ignited, the fireball produced only 4-5 psi of overpressure, which did no significant damage to the orbiter. The orbiter broke apart from aerodynamic loads. In order to have an explosion with overpressure, you need to thoroughly mix the reactants, then contain the ignition long enough to form a supersonic propagating wavefront, which results in a detonation. CRS-7, which was more or less the worst-case scenario for an in-flight failure of a liquid-propellant two-stage booster, had an extremely benign failure with respect to the capsule. Super Heavy may be problematic in this regard. If a leak formed in the common bulkhead on the pad, the liquid methane and LOX could mix, forming an explosive gel. Its inevitable ignition would be constrained by the cryosteel walls long enough to produce a massive detonation, larger than some tactical nukes.

In a propulsion-integrated, reusable capsule, this can be trickier. You need a heat shield on the back end, so the spacecraft's orbital engines (which would be on a service module if they were expendable) do not necessarily point backward. Dragon 2's major burns (deorbit particularly) use four Draco thrusters grouped around the docking port, pointing forward, and are covered by the nose cone during ascent. If you use the same propellant for abort as you use for orbital maneuvers and deorbit, as is the case with Dragon 2 and Starliner, then the only "extra" thing you're carrying to orbit is the weight of the abort engines themselves, and those are not terribly heavy to begin with.

Yeah -- KSP can automatically use gimbal to adjust pitch, roll, and tilt, but not to correct for CoM. The Shuttle's OMS engines are a good example of this -- they could gimbal parallel if they were being fired simultaneously, or they could gimbal outward to thrust through the CoM if they were being fired one at a time. The only way to model this in KSP is to position them gimbaled-out by default and add angled docking ports so you can "control from here" when using only one engine. And even then you get cosine losses when firing both. To the ongoing horizontal-lander structure debate...it's possible (and perhaps easier) to do the horizontal landing module with only a single engine, but having two allows the combustion chambers to be higher than the base of the capsule, which decreases ground clearance. Having the two offset allows the entire lander module to be substantially narrower than twice the engine bell diameter, which is especially helpful when you need to fit frag shields somewhere.

If the engines are fixed it is very challenging. If they have some gimbal, it's a little easier, because you can compensate through both gimbal and differential throttling. This is sort of where I'm thinking.... Top is the capsule. Roughly the same size as extended Cygnus. Internal tankage in bulkheads; airlock on one side, docking port on the other, berthing and propellant transfer port on the bottom. A few windows. At bottom is the capsule with the ascent module attached. Depicted is the ascent module structure, engines, and ascent propellant tankage. Drop tanks and landing platform not yet pictured (the drop tanks affix to the landing platform; the landing platform attaches to the central thrust column between the ascent tankage). Independent RCS for the ascent module is also not pictured. The ascent module forms a "cradle" under the cylindrical crew capsule and the engines and ascent propellant tanks are all offset but parallel to the CoM<->CoT vector. With drop tanks installed, the descent module almost exactly fills out the Falcon 9 fairing, horizontally. The landing platform would have deployable legs and mate primarily to the central thrust column under the ascent tankage. In compression, the corners of the ascent frame would rest on arms that would distribute the capsule's weight directly to the legs. It would also have frag shields for the engines, mounting points for the drop tanks, a ladder up to the airlock, and room for flatpacked cargo. EDIT: If needed, the airlock could be moved to the top of the capsule and the capsule itself could be shortened so as to allow the drop tanks to protrude upward on either side. The capsule itself doesn't even need to have the vertical-launch/horizontal-integration approach. But the enabling feature is to use vertical launch and horizontal integration for the lander module.

If you balance the engines correctly to begin with it's not a problem.

Correct. This is the biggest advantage, tbh.

I am considering whether a dual-engine design is a better solution. The descent burn is extremely long in most of my evaluations...almost to the point that you'd be burning all the way from LLO continuously. Oh, okay. No worries. I thought you were working in a whole plane change. I think this is the best solution, although 6m long is not ideal for craft stability when landing. It is if you're landing horizontally. I wish I had KSP on this computer because I think I have a really, really cool solution.

At least they knew the LES worked.

I am always shocked to see how far up the first stage the exhaust plume was entrained.

It appears to be a Fibonacci spiral.

The Boeing lander appears to fill out the whole volume of the SLS Block 1b payload fairing pretty thoroughly. Ah, see, this is a critical difference. I was going to have the capsule without any internal propellant reserves, fed only from the pressed drop tank. The trouble with pressure-fed thrusters is the pressurant. The SSVP docking port that Progress uses to mate with Zvezda has three different lines to allow transfer of fuel, oxidizer, and pressurant, because the propellant won't move without the pressurant to push it. But that makes things quite complex, because you have to have two liquid lines and one gas line, all at different pressures, and you have to have an internal system on the receiving tankage to vent existing pressurant in the propellant tanks and valve off the pressurant while it is being filled, and so forth. You probably need a separate nitrogen tank to purge. If you run your pressurant through a regenerative cooling loop, that adds more complication. I don't even know how ullage is handled. I thought it would be simpler, all things being equal, to dispense with the internal tank altogether and simply plumb the reaction control system thrusters to a two-line port for the drop tank to mate with. That way the pressurant never has to transfer; it's merely used to push the fuel and oxidizer out of the drop tank and into those lines through the port. One-way flow, no ullage problems. If we do trickle propellant transfer into internal tanks, however, a lot of things change. There is no inline drop tank. The crew capsule's fuel, oxidizer, and pressurant tanks would be refilled from excess capacity in the transfer drop tanks while still connected to LOP-G. Depressurizing the whole capsule is definitely not a good option. What we most need is extra volume. Volume is not always heavy. This is another reason why doing the crew capsule launch independent of the lander is helpful. You can launch a very lightweight but very large (4+ m diameter, 6+ m long) module vertically on Falcon 9 to TLI, but then have it mate to LOP-G and to the lander horizontally. I will try to run some numbers and see how it works. One note -- if the capsule is being used for significant dV, the R4-D is a better choice for the RCS thrusters, and they get 312 s. Also, I would propose giving 100-300 m/s of dV to the capsule so that it circularizes while the lander crashes; that will help with the overall efficiency as well. I don't think either of these are necessary. The departure burns from NRHO are optimized for proper planar insertion, and the ascent burn is timed to coincide with return to NRHO. That's the nice thing about polar landing sites.

Question is when do we start judging? I mean, we can go back to when R-7 was an ICBM in development, but that doesn't seem fair either.

I just realized that possibility. There are some really, really cool ways to try and do that. Particularly if two engines are used for descent but only one for ascent. The Boeing solution is to use a cargo SLS launch, which is inestimably stupid and generally regrettable. The Boeing lander could be easily sent to TLI using distributed launch -- send it up on Vulcan or Atlas V or New Glenn or Falcon Heavy; boost it to TLI with a naked FHe -- but it won't fit in any commercial fairing.

You can go here but it might take you a while. Also hard to know where you do the cut-off. If you use a contemporary time period with SpaceX then it's easier. The various Soyuz versions (all R-7 derivatives) have had 157 launches since June 4, 2010, six of which failed. The Bayesian success rate over this period is 95.6%. If you use the same time period during which Block 5 has been flying, then you have 25 launches and one failure, which gives you a success rate of 92.6%. Not really fair to Soyuz but them's the numbers.

For inspiration, here are a number of possible designs floated/discussed during Constellation: I'm trying to think outside the box. There are some really wild possibilities...like, what if you had two engines mounted in parallel to the CoM, but only used one for ascent? What if the sole docking port is angled? What if the docking port is in the side of the crew module?

I remember the first time I made a ramjet jetpack. It was glorious.

Pricey. About the same cost as renting out an entire vomcom flight, but for one seat.

Surface tension doesn't hold you underwater in normal conditions because gravity pulls the water away. That doesn't happen in microgravity. The surface tension won't hold you underwater so much as the surface tension will leave you coated in water and unable to clear the space around your face for breathing. Watch this video of Col. Hadfield wringing out a washcloth in microgravity: With a bunch of water in microgravity, imagine you are the washcloth trying to get free.

Progress can transfer nitrogen tetroxide, UMDH, and pressurant into Zvezda via connectors in its non-androgynous SSVP docking ring. The US's NDS is planned to someday have this capability, though it does not at present. In any case, this type of transfer capability is likely not high-flow enough to enable the kind of architecture we're discussing. On the other hand, what we need is simpler than the SSVP, in a sense, because we do not need to transfer pressurant independent of propellant. The tank that would couple to the crew module would contain its own pressurant and simply force fuel and oxidizer at high back pressure through one-way ports into the RCS piping of the crew module. It would only make sense to have the connector also serve as the structural mating point for the tank. The tank itself can be very "dumb" with all the valves controlled within the capsule's RCS. The Altair lander solved this by having a separate airlock module mounted offset, parallel to the ascent module, to be left on the surface. Of course that was only enabled by the prodigious 10-meter fairing on Ares V, which is out of the question here. Remember that the crew module does not necessarily need to go in the fairing. It's fine to launch it separately. We might have to eschew plans to include a Cygnus, but that's okay. What about using an off-axis docking port? I am envisioning a docking port that mates to the ascent stage, but also includes an access tunnel to a lower-stage airlock that sits next to the engine. That tunnel is severed with electric bolts (frangible as backup) before ascent. Balance the weight of the airlock on one side with downmass capability on the other, and mount the ascent tanks perpendicular. You don't have to do all the modeling. Just post the wet and dry masses of the ascent stage and the drop tanks and I will do the math to figure out what it could reasonably drop onto the surface from TLI.

Resistance-based exercise is lighter in weight and seems to work well:

Since when does ICPS have two RL-10s?

"But why doesn't Mercury cast a shadow on the sun? Shouldn't we see Mercury's shadow underneath it? SPACE IS FAKE!"

The tension rod cannot slide through the rings, though. Looks like I will get a view of the new constellation in my neck of the woods right at around 6 PM local time. 3.6 mag might be too low to be visible from the city though.

I asked over at NSF, and it was explained that each sat has three cylindrical mounting points that nest into the next one down, preventing any tilting or planar motion during flight. There are four tension rods which keep the two stacks in compression; when the tension rods are dropped (evidently without much of a spring because we saw a tension rod drifting away quite slowly) the sats release each other and drift apart. The three mounting points are shown here: I wonder if the tension rods fit over the topmost mounts? They certainly can't pass all the way through the rings or the sats would end up stuck on the rod. I wonder how they unload the reaction wheels. Maybe they can use the solar array in tidal tension?