You are not logged in.
Skylon's SABRE is a liquid-air cycle engine. They think they can solve the heat transfer problem to liquify the air as it comes in, with the cold hydrogen. Heat transfer is the slowest of the physical processes, which is why liquid air cycle engines were never attempted before. But, I hope they can do it.
Gas turbine engines can be built that have pretty good T/W. Isp looks good on subsonic fanjets, not so good on supersonic-capable low-bypass ratio turbojets, and really bad when you light the afterburner. The combustor can has to run very lean on fuel to hold turbine inlet temperatures under about 1800 F nor normal flights, under 2000 F for temporary high-power operation. Really exotic materials can add about 200 F more to those figures. But you hit those conditions with turbine somewhere between about M 3.2 to 3.6, almost no matter what design you use.
Most of the combined cycle proposals I have seen compromise performance of each component just to build one device instead of two. (Maybe SABRE can get around that, we'll see.) Most of these combined-cycle things end up pretty complex and heavy, as a geometry change is almost invariably involved. The best of these is the ejector ramjet my old friend Joe Bendot did at Marquardt. It adds a big rocket inside the duct of the ramjet.
On takeoff, the rocket thrust induces some airflow through the ramjet to generate static thrust. Isp is closer to rocket than ramjet. Once you've reached the speed at which the inlet functions properly (usually M1.5+), you can fly on ramjet alone at high speed. This will work as an accelerator up to around 60,000 feet altitude, where frontal thrust density drops too low due to low air density.
At that point you have to turn the rocket back on, and use both together at blended Isp, which floats down toward rocket levels as you climb higher into unusably-thin air. Run the rocket alone, on into space, at rocket Isp. SSTO.
But rocket Isp won't be as high as you are used to, because being inside the duct interferes with free plume expansion, especially in near-vacuum. Typically, lowered Isp raises GLOW of an SSTO no matter what kind of propulsion or airframe design you choose.
I actually like the parallel-burn approach better, because neither device is compromised, and you can run any blend of the two without variable geometry. It actually works out no heavier, and averages higher performance than the combined cycle designs. Rocket-ramjet (maybe to M6, and rocket-turbojet (M3.4+/- on turbo) would be two good candidates.
The air turboramjet could beat the M3.4+/- limit in turbine, if 100% air bypass from ahead of the compressor face was used. No one has ever built one, but I think it could be done without too much fuss and bother. If you can shut down the turbine and cut off all its airflow, bypassing straight to the afterburner, you can run the afterburner as a true ramjet. Capability could be as high as M6. M5 anyway. Hard, but do-able.
As for the book, I have several topics roughed out now, but only about 30% or so of them. None are in final form. The science is not too hard to write down, it's the art that's hard. I did some of that the other day, on really how to size ramjet system geometries for several different ramjet systems and weapon/launch applications. Still sweating. Or swearing. (Both, maybe.) I never wrote that stuff down before, I just did it, and noticed I was one of very few around the country able to do it that fast and well. Neither has anyone else ever written sizing procedures down, as near as I can tell.
GW
Rune:
I know nothing about any "super Draco" thrusters for Dragon. The ones they have are supposed to be the launch escape system for any manned Dragon, so capsule T/W is substantially larger than 1. For the Mars mission paper I gave at last August's convention in Dallas, I re-engineered some data from the Spacex website for Dragon as they had it posted last summer. I was showing 0.9 km/sec delta-vee capability with 6 suited astronauts on board. If you put more propellants in the unpressurized module, and connected that to the Draco system, it adds around 1.4 km/sec more to the capsule's delta-vee capability. I looked at that to get 2+ km/sec total delta-vee for Dragon as an emergency escape crew return vehicle from a very high-speed Mars return in an emergency, somewhere well above 15 km/sec. Maybe as much as 25 km/sec.
Stock Dragon post reentry has no unpressurized module. So the capsule's delta-vee capability is around 0.9 km/sec utter max, post re-entry. That's enough to land from considerable height without a parachute, but it's nowhere near enough to make the entire landing without a chute. Not on Earth. Moon? Not even there. Any "pinpoint" to the landing is during that final burn, but it's not all that much maneuvering. You know advertising flacks, they do tend to oversell stuff.
On the other hand, Dragon's true significance is that it is the very first capsule in history intended to be re-used, heat shield and all. It does need a new unpressurized module and a new nose cap for each flight. But, that heat shield was designed for a free return from Mars at about 50,000 feet/sec [15 km/sec] entry velocity. You could re-use it even after a return from the moon (36,000 feet/sec [same as 11 km/sec] for comparison). Coming back from LEO at around 26,000 feet/sec [8 km/sec] is by comparison "easy", since the heating is proportional to velocity squared. I'd guess they could fly the same heat shield to LEO about 4 times before replacing the pieces that make it up. And it is segmented.
None of Dragon's capsule-type competitors are intended to be reusable. Not Orbital's Cygnus, not Boeing's scaled-back Orion. While my contention is that Spacex achieved its lower costs per unit of payload to LEO by smaller logistical support tail, not actual reusability of anything, having a capsule you can fly several times is still going to help significantly. Plus, it's simply a technical capability that we need. And we have needed it for a long time. It is a major accomplishment.
I'd have to go find my original notes for that re-engineering I did on Dragon to find the Isp I used for the Dracos. But I think it was likely closer to 270 sec than 300 sec. There's quite a gap between what is theoretically-possible under perfect expansion, and what can actually be had with a real engine bell, even in vacuum.
GW
Plastics used outdoors on Mars are going to be susceptible to rapid UV damage, without some sort of opaque coating. Indoors, just like here, pressurized or not. TiO2-white or C-black paints work pretty good here, on kit-built glass-polyester/vinyl ester and glass-epoxy airplanes.
The cold is bad for plastic items, too. Most of the materials I know get rather brittle below glass transition temperatures that are not very far below room temperatures here. Sort of like our winter temperatures. Mars is a lot colder. Reinforced plastics would do better in the cold, but that's not extrudable or 3-D printable. You're talking hand layup in a mold for that, like building canoes out of fiberglass. Whether you bag-compress it or not depends on the materials. Vacuum-bagging makes really nice carbon-epoxy panels.
GW
Landings:
I quite agree with RobertDyck. I'd add these things to the discussion:
The main problem with lifting bodies, aside from that speed where control is "iffy", is the high landing speed due to very high wing loading or broadside ballistic coefficient. A vehicle designed to fly into space and back is inherently still a bit heavy on landing because of the heat shield and the control propellants. Most of the those lifting body designs had touchdown speeds at Edwards in the 200-300 mph range. Stability on the ground during roll-out (skids or wheels) is a serious issue, unless the vehicle is rather long, like X-15, which landed fairly well at 200 mph.
Gemini was originally a Rogallo wing-type parasail design, with skids for dry lake bed landing. My memory may be off, but I think the landing speed was near 200 mph like X-15. The Rogallo was steerable, but the short "wheelbase" of the skid system was unconditionally unstable on the ground at any speed. Test articles always flipped over and went bouncing down the dry lake. Good thing they weren't manned. It just never worked, so they went with a parachute and water landing. The capsule was already in production, so that's why the chute came out the side instead of the nose on Gemini.
I've got no real problem with MMH N2O4 residuals after landing. Just don't go sniffing the rocket nozzles for a while. I remember NASA did have concerns, when the shuttle first flew, but they got over being nervous about it after a while.
The Draco thrusters on Dragon are rather powerful, and it has a lot of delta-vee built into it. If one reserves a kitty of propellants, one could fire them last second to slow a chute landing on land to survivable levels for the crew. It would feel like a car crash, just like Soyuz. If it had shock-absorbing legs, the heat shield would survive for re-use, too. Of course, they're extra weight. Reduces payload or crew size a tad. But the water landing version is said to be able to carry a crew of 7.
GW
Rune:
Yeah, I kinda forgot about land landings. Old guy. Sorry. Most of the launch sites currently operating launch out over the oceans, except the Russians. I am more familiar with what we did than I am the Russians.
The Russians do land landings all the time with Soyuz capsules, and the Vostok/Voshkod series before Soyuz. They also developed a way to deliver tanks to the battlefield by air. It combines parachutes with last-second retro-rockets, and it has worked for decades.
That kind of thing is steerable by cg shift during re-entry hypersonics, not after. You're pretty well ballistic, coming down on a drogue, then some sort of main chute cluster, and those tend not to be steerable. Pinpoint landings are impossible, you will always be faced with recovery operations, and that's not cheap.
The sudden stop in the water at 20-40 mph is bad enough, on land it was unsurvivable, unless you did one of three things: (1) energy absorbing one-shot crush seats, (2) last second rocket braking, or (3) have the crew bail out on their own chutes before impact. Vostok and Voshkod did item (3). Mercury and Gemini simply ruled out land landings. Apollo and Soyuz now do item (1). I think some of the early Soyuz's might have used item (2), but I may not be remembering correctly after all these years.
But rocket braking works if you combine it properly with parachutes. The trick is to use the highest thrust you can tolerate for the shortest possible burn at the last seconds before impact. That's a serious control and altitude/descent rate measurement problem. Otherwise, you have to carry a huge kitty of heavy landing propellant, so you can just sort of "slop" your way through it. The Russians did this for parachute/rocket landing of heavy tanks to the battlefield from high-altitude transport aircraft.
Spacex is going for the same scheme with land landings of Dragon, I believe. Not sure whether the Dragon would have legs to protect the heat shield, it might. But it comes down on a chute, then fires the Draco thrusters to slow for touchdown, last second or so. You just have to plan on landing with enough thruster propellants to cover your needs with a margin of safety. Of course, those propellants are a crew hazard after landing. Hydrazine and some oxide of nitrogen, I believe.
GW
I may be the victim of past thinking, but I think the main decision to make is whether you want lifting flight or ballistic flight post re-entry. Putting the weight off-center a bit in a traditional capsule gets you side forces of some significant fraction of the drag force during re-entry hypersonics. This allows you to shift the location of the landing ellipse by a couple of ellipse dimensions L-R, or downrange-uprange. But it's worthless as a directional control once the hypersonics are over. From that point, you're ballistic. Control there requires real wings.
Wings are pretty useless during re-entry because they're fragile and vulnerable to overheat, as in the shuttle design, the X-20 we abandoned, and the X-15 which was considered for orbital flight using a Titan-3 booster, but which plan was abandoned because it could not survive re-entry. In the X-15's case, this was because of wings burning away really fast at nose-first attitudes, but inadequate structural strength to re-enter more nose-high ("semi-broadside") like the shuttle. The X-20 was to have addressed that weakness with experimental ablatives, and/or transpiration cooling, and/or heat-sinking. Shuttle did it with slow ablatives (carbon-carbon LE's)and refractories (low density ceramics).
There are a couple of things we haven't tried yet: some sort of ablative-protected inflatable structures, and protection by an engine plume between vehicle and hypersonic shock layer. There is great potential in both ideas, but I see nothing at all going on around me. Not a space agency anywhere is exploring this, and private concerns “typically” do not invest their own money in exploring radically new technologies (that’s why there have to be government agencies, no one else will do that job).
The aerospike nozzle is indeed what you might want for engines of any type that must function across varying backpressure going up. Coming down, I fear the sharp aerospike might be rather vulnerable, if you used the engine plume as your heat shield. But I don't know. No one does, not yet. Aerospike would have a lot of application to first stages going up, except our current methods for building them end up being heavier and inconveniently-packaged for a cylindrical shape, and the nozzle efficiencies typically run 2 or more % lower than a traditional bell. It only looks better at off-design backpressure. It's a non-trivial trade-off, and unfamiliar territory, so no one has really done it. X-33 was supposed to do it, but never flew for a whole plethora of reasons, some technical, some not.
Recovering a upper rocket stage in its entirety would probably be best done with a blunt ablative on the front end, and some ablative-protected inflatables around the engine(s) to double as floats, and as water-impact shields for the engine bells. Take the main sea impact on the heat shield, we already know that works fine. You just need to add a series of drogues and chutes that deploy out the rear, around the engine(s). Biggest risk I see is shroud line fouling on the engine(s). But it's doable.
That's a lot of extra gear and equipment. Plus, the tankage has to take the "whack" when it hits the sea at dozens to a hundred gees. You're just not going to build a thing like that at 5-8% inert fractions. Steel, aluminum, and titanium are "only so strong". I really don't see organic composites as a tankage option for something intended to survive Mach 25-class re-entry, either. It's just too thermally harsh. Something that could be re-flown a lot of times is just going to be heavier. 10+% inert, probably 15+%. Just guessing.
Lower stages are far easier. Hypersonic heating under Mach 10 or 12 is a whole lot less intense. Organic composites in some areas become feasible. You could even put flyback wings on it, perhaps. But whatever you do, it won't be 5-8% inert. The way around higher inerts at fixed rocket performance is more stages. That, too, is well-known and well-proven. More stages does not necessarily mean more cost. But you do have to keep it simple so the logistical support tail can stay small. The rocket / system designs need to look more like battlefield weapons than our traditional launch rockets. Trying the first steps into that model is the real secret of Spacex's lower costs. They've re-used nothing yet. And at 5% inerts, they won't. You can bet any flyback first stage with landing legs or whatever, will be a lot heavier than the stage they're flying now.
As for the suborbital rocket planes, those typically leave the air and return at Mach 3 or so. That’s actually very little transient heating, and it is a short transient. Not at all the same problem as sustained flight at Mach 3. Even with organic composites as exposed structure, it is possible to heat-sink your way through that short re-entry. There is no radical new heat protection or structural technology there with Spaceship Two or Lynx. The “radical” item in Spaceship Two is the folding tail. Lynx’s “radical” technology is the long life / low maintenance engine.
GW
How to test gas core?
Short form: same as the solid core program. Put it on a stable thrust somewhere and and fire it. Same as all rockets. The first tests simply cannot be flight tests, that's way too much to bite off all at once.
Yep, the exhaust is radioactive, similar to the Phoebus and Kiwi predecessors to the final NERVA. True enough. A properly-working open-cycle gas core machine will be running at low concentration radioactivity: around 1000:1 hydrogen to uranium fission products by mass. But the total uranium mass fed to the burn will be expelled as fission product mass. The early ones will be much worse, in a concentration sense, until the containment flow scheme works right.
Closed cycle ("nuclear light bulb" designs) will have a clean exhaust, unless the physical containment fails. It will in early testing, occasionally. The problem with closed cycle designs is that the core fission products get retained. I like open cycle better. On shutdown, it's "an empty steel can". Thermally and radiologically, it's "cool" in minutes to hours. Retained cores are dangerous for decades to centuries.
Problem: our rules no longer allow us to free exhaust radioactive plumes. It is possible (in a very expensive facility that we do not currently have) to capture the plume and separate the hydrogen from the radioactive "dirt", and sequester the dirt for disposal.
Wild idea: do it on the moon instead, as free exhaust. Exhaust speeds far exceed lunar escape, and there are no air and water to pollute, or neighbors to bother. It might (!!!) actually be cheaper to do it that way on the moon, instead of plume capture here on Earth.
Even a resurrected NERVA tested here on Earth will have to be tested plume capture.
It's far faster, more effective testing as open plume. The program proceeds much faster and effectively. Resources get concentrated more on the rocket and less on the facility.
Can't do it here? Then do it there! On the moon. It's close enough to reach quickly and with relatively low-performing rocketry. Emergency help is but 3 days away.
BTW, the Th-232 to U-233 breeder cycle, once bootstrapped into operation, yields fission fuel with shorter-lived daughter products. It'll work as a reactor fuel, probably even in nuke rockets, but is not "concentrated" enough to be a bomb. No plutonium in the cycle, either.
I kinda like the concept of a U-233-fed open-cycle gas core engine. It's something easily abortable on launch, and not very dangerous in a crash. At lower power, Isp is near 1500-2500 sec with engine T/W's 10 to 30, or perhaps higher. No waste heat radiator, regenerative cooling is adequate. Higher power, you need the big, heavy radiator: They were going for 6000 sec Isp at engine T/W maybe .05 to 0.1.
I sure wish we'd already done it. The bench tests 40 years ago looked very promising for both gas core reaction controllability, and for open cycle containment by that 1000:1 ratio. That was the target for "perfect containment" at their residence time and burnup rates back then. But it was just some academic-institution bench tests.
GW
Myself, I would get started going to the moon and Mars with the chemical launchers we have, or soon will have, like Falcon-Heavy. I'd (in parallel) work on resurrecting the old NERVA, since it did everything but actually fly on Saturn-5, and use that type of engine to build single-stage landing craft capable of tail-sitter landings on Mars, and returning to Mars orbit, all in one propellant loadout. Any boat that can do that, can ferry very heavy payloads to and from lunar orbit, down to the lunar surface.
In parallel with all of that, I'd also be working the gas core nuclear thermal rocket ideas, as those could potentially solve the radioactive core problems at performances as far beyond NERVA and Timberwind as they are beyond kerosene-oxygen. I'd probably do the nuclear rocket work on the moon, as doing down here requires no free exhaust, these days. Good reason to go back to the moon, in my opinion. Safe place to do dangerous work of high payoff potential, yet close enough to reach easily with stuff we have right now.
T/W > 10, maybe > 30 at Isp 1500-2500 sec? Good single stage launcher. T/W near 0.1 at Isp near 6000 sec? Good orbit-to-orbit "hot-rod" engine. The bench tests ca. 1969 indicated these things are indeed feasible.
Of course, for really gigantic orbit-to-orbit colony boats, there's good old nuclear pulse propulsion. The bigger the ship, the higher the Isp, and the easier the shock absorber design is. That comes quite a bit later, though.
GW
Bob Clark:
What I had posted for the two-stage airplane was not the final form, just what I had done at that time, and where it pointed. I tried to climb really high before pulling over and accelerating the ramjet first stage airplane, and the air was too thin for thrust-minus-drag to accelerate the mass at a practical rate. That's why the first stage went too far downrange to fly back.
Numbers down closer to 60,000 feet for the staging altitude look better. I need to re-run that same study with 60,000 feet staging, and see if the booster flyback becomes practical. I simply haven't done it yet. But I'm pretty sure it would work.
From what I'm told, there are 3 variables of importance to selecting the staging of a HTOL 2-stage vehicle. They are, in order of importance, speed, path angle, and altitude. The most important is speed. That's why I picked ramjet: with external or mixed-compression inlets, it's capable of useful thrusts to M5.5 to 6, and can take over as low as M1.6-ish. I went with separate rocket and ramjet engines that can be burned in parallel, that's how I achieve about 45-degree path angle at staging in a sudden pull-up transient without deceleration (combined cycle probably won't be able to do that).
I just couldn't make it work right at 100,000 feet. Frontal thrust densities are an order of magnitude higher at 60,000 feet. Altitude is the least important of the three variables, so I feel pretty good about the HTOL 2-stage approach with a rocket+ramjet airplane 1st stage. The 2nd stage can be a rocket ballistic pod, or a rocket airplane, whatever the mission needs.
Configuration design for drag reduction and for impinging-shock avoidance are the truly critical issues. Shock impingement heating can cut through structures in a second or two at M6+. We saw that on the X-15 flight that carried the scramjet test article.
GW
I suppose you take your time before dismounting, and then move away quickly, once you clamber down. I envisioned a crane arm that makes a nice personnel and cargo elevator. Proximity is bad (inverse square), but really it accumulates over time. What you don't want to do is stand near the thing any longer than necessary. (One of the reasons I like gas core concepts better is faster thermal and radiational cooldown: essentially an empty chamber.)
A few dozen meters away is a pretty good distance temporarily for unload trips. Pitch camp maybe a km or two away. You're pretty safe up in the lander cabin with the propellant and structure for a shield. Nice shelter for solar flares if it's built tough as an old boot, and you have a couple of water and/or wastewater tanks to hide beneath.
The Mars Society archives on-line have my original paper. A version of it is posted on "exrocketman" dated 7-25-11. Some second thoughts about using NERVA vs electric as the backup scheme is posted 9-6-11. That site is http://exrocketman.blogpot.com If you click on the identifier "space program", then it shows only those articles with that identifier.
GW
What RobertDyck said is exactly correct. Decide what has to go to the moon and land there, and what has to return, first. Then get it there from LEO. Then launch it. That is the correct design sequence. And that's why what you intend to do on the moon so entirely drives the design.
Designing the launch vehicle first (as with NASA SLS, designed by Congressional politics, not engineers) is a wasted exercise.
Getting what you need to LEO need not take one launcher. We now know how to dock things together and assemble very large items in LEO. It's not so much the number of launchers that drives cost, it's cost/mass delivered, and what payload sizes are already flying. Why build a bigger rocket and have to amortize its development costs, if you have a smaller rocket that is "big enough" and already "cheap enough".
Spacex Falcon-Heavy is 53 metric tons to LEO from Canaveral, at roughly $600-1000 /pound (same as roughly $1200-2000/kg). Closest rival is Atlas-5-Heavy at 20-25 tons and more than twice the price/mass.
SLS will be built from retreaded shuttle components by the same entities that built the shuttle, working in the same ways they always did for shuttle. It will never be as cheap as Falcon-Heavy, or even Atlas-5. Shuttle was $1.5B for each 25 ton payload.
GW
The mass ratio and the thrust/weight has to be there for surface launch. Adding engines usually drives the mass ratio down, unless you scale up the tankage a bit. That's why most first stages are so large.
First stage engines are usually of different design than upper stage engines: shorter bells. It really needs to operate perfectly expanded at launch level (usually sea level). She'll be underexpanded as you climb, which costs performance, but then so does over-expansion, and it's far worse. Usually first stages leave the sensible atmosphere, so it's in vacuum by burnout. Upper stage engines can be designed to be optimal in vacuum-only, with very long bells.
So, it's not exactly the same engines. Even for the same propellants, first stage Isp is a lot lower than "typical" vacuum Isp designs. Be careful trying to scale from one application to another.
GW
Josh:
What I had in the paper at last August's Dallas convention in part called for a single stage NERVA-propelled lander, a big rover with a drill rig on it, a swarm of small robots to assist 3 persons on the surface, and an inflatable Quonset hut to live in. You erect the hut and base your activities at a safe distance away from the NERVA, not in the lander. But the lander makes a better shelter if a solar flare occur. Surface time was a week or two. You leave a transponder at each site, to enable future precision landings.
My paper said send 3 down while 3 monitored and did science from orbit. Then alternate crews and landers. I put enough delta-vee into the 20% inert 10% payload landing boat to land 25 degrees out of plane and return, with rocket braking all the way down. That's no aerobraking credit, but a gravity and drag penalty on the way up, so its a very conservative size-out. Having at least one lander ready in orbit at all times for a rescue trip was one of my mission requirements.
You could do that with any mission making multiple landings. My design was 6 persons, and 3 landers so that the loss of one did not terminate the mission. The three landers pushed all the landing propellant supply unmanned to Mars orbit. I sent the manned vehicle with enough propellant to return, in case rendezvous failed for any reason. Suspenders-and-belt.
The prime design called for fast trip zero-gee manned ship, but the backup was a "slowboat" NERVA, which would be spun end-over -end for artificial gravity. No more than one year at microgravity, and set right at 1 gee for artificial, again suspenders-and-belt, based on what has already been done. That's probably what we really want to do.
My paper's design made 16 widely-separated landings in the one trip, plus a visit or two to Phobos. A real planetary ground truth survey. There were 3 unmanned vehicles sent one-way to Mars plus one manned vehicle that goes two-ways and is recovered in LEO to be used again. All 4 were in the 600 metric ton class as assembled by docking 34-ton modules in LEO. (That was based on Spacex's projections for Falcon-Heavy before they settled on 53 tons.) It true exploration based from orbit (what all is there? where exactly is it?)
The return from that trip would be the critical ground truth information for selecting one or two experimental base sites. These would be the places where you set up the real ISRU that you tried out on the first mission. Some of those first examples will have worked, most won't. But that experience takes "theoretical" designs and turns them into machinery you can rely on, for the second trip. Again, suspenders-and-belt. This second trip is surface-based work, and "looks" more like what people are proposing in these forums and most of the papers I see. But the chances of success are much smaller, if you don't do what I suggest for a first survey mission.
The experience and supplies generated by those couple of well-sited experimental bases is what enables a more permanent settlement, that could blossom into a real colony if some trade commodity could be identified. Done right, this could happen fairly fast. Done wrong, the inevitable failures and fatalities might well kill the process.
Suspenders-and-belt. And an armored codpiece!
GW
I did exactly what I suggested in my previous posting, and created the launch costs plot. I did it for the 3 Spacex Falcon birds, 3 of the Atlas-5 family, and for Delta-4 heavy. I plotted the data in metric units as $/kg vs metric tons delivered to LEO from Canaveral. I also replotted in US customary, for folks who know those units better: $/lb vs US tons delivered to LEO from Canaveral. Those data are public view over at http://exrocketman.blogspot.com
GW
I didn't make the meeting where Paul Webb spoke. I would like to have. My dad knew him decades ago as a crew escape expert, before all the pressure suit stuff was known as well as it is today. Dad and I were both aeronautical/aerospace engineers.
I did make the Dallas convention this last summer - gave a paper in the "advanced technology" session on doing a whole slew of landings in one trip to Mars. I got to meet 3 of the original NERVA guys at that meeting, and I bought the book that one of them was selling. It matches my memories of NERVA pretty well.
I quite agree that we really don't need to build another Saturn-5. But if we did, one could resurrect the old NERVA upper stage design for it, and do exactly what I wrote in my previous posting, making a bunch of landings in one trip. The new government heavy-lift launcher design is a new Saturn-5-like vehicle based on retreading shuttle technology. But I seriously doubt any government design will ever be more cost effective than they ever were, which is ineffective.
On the other hand, the new Spacex Falcon-Heavy that is supposed to fly next year is priced at around $1000/lb ($2000/kg) of payload, for 53 metric tons deliverable to LEO. There is not one single reason in the world why a nuclear transfer stage, a couple of big capsules, and a whole slew of modern LEM equivalents cannot be assembled in orbit from 2-5 Falcon-9 launches. One trip, a bunch of landings. Same as Mars.
There is also not one reason in the world why a smaller nuclear upper stage could not be fitted directly to Falcon-Heavy itself. The options are wide open.
GW
i
Why not use the bigger NERVA 3rd stage design, which is restartable (and was back then flight-ready), and push more than one C/SM and a whole slew of landers to the moon, and make several landings at different sites, all in one trip? Isn't that a better return for the launch cost and all the trouble of going there?
Trouble with reviving NERVA or Timberwind or Dumbo or any of them is the lost engineering art as just about all of those guys died or retired. Rocket science ain't all science. It's about 50% art that was never written, just carried in the minds of the practitioners. It's about 40% science, all written down somewhere. And it's about 10% blind dumb luck. And that's in production work. It's worse in development - the art factor is a lot higher. If you lack it, you will think think the blind dumb luck factor is huge, and mostly bad luck.
Been there and done it....
GW
Try graphing Spacex's costs as $/kg payload to LEO vs kg payload to LEO. You get a decreasing curve of costs, not linear. Then spot Atlas-5 20-25 tons at about same per-launch cost as Falcon-9. Bigger tends to be lower cost but Atlas-5 falls well above the Spacex curve. This in spite of Atlas flying in one form or another since 1956. That will illustrate best what I have been saying about the importance of small logistical tail and a not-gigantic (bloated) company. If Spacex were to close up shop today, Atlas's prices would at least double tomorrow.
GW
Thinking like suspenders-and-belt is how you survive in space. I would definitely try some fuel production on the first landing. I would not count on it in any way for the return from that first landing.
Question: why is everybody still focused on one trip-one landing? Why not make one trip and several landings? It's a lot of trouble to go there. Why not make it really worthwhile?
GW
The thing the crew will spend the most time inside will be the habitation module of the orbit-to-orbit vehicle. That's the one that needs to be voluminous, and it does not ever need to land. If it's at one end of a long ship, just spin it end-over-end nd you have artificial gravity. That solves a whole host of life support problems if you have artificial gravity. That approach does require de-spin for maneuvers, but that's no real problem.
I like the submarine example, too. The old diesel electric boats, particularly German U-boats and the 1920-vintage US S-class, were very cramped, not intended for more than a month or so at sea. They were called pig boats for a very good reason. The larger fleet subs were still cramped, but livable for a nominal 3 month war patrol. 60-70 men inside a 300 foot ship that was chug full of machinery. Crawl through one sometime. There's several on display as memorials. The new nuclear attack boats are better still (see Nautilus, on display), and the missile boats are very spacious by submarine standards. 6-9 months possible in them.
Here is something to consider: it's not the gross volume but net, after subtracting off for machinery and equipment that occupies spaces. When you do that, the various space stations don't look so very spacious inside, excepting the old Skylab. You can see this effect in photos. It is very significant to psychology.
What you take to the surface of Mars, you really don't have to live inside-of for so long, plus you can go outside! An inflatable pitched near the descent vehicle makes a lot of sense.
I'd be very careful comparing a Mars lander design to the Apollo LEM design. The velocity requirements are very much higher at Mars, plus you have entry heating and ascent drag to deal with. Likely 3-4 stages if chemical. I'd go nuclear - single stage is possible with NERVA-type technology, which we could still resurrect (not quite all those guys are dead yet).
GW
At 2-6 mbar pressures, it is energetically very, very unfavorable to utilize the Martian atmosphere as a source of CO2 carbon, although it certainly CAN be done. Compressing gas from low density to high density is very, very, very, very energetically inefficient, and we are talking final processing pressures here measured in tens of atmospheres, or tens of thousands of mbar. That is one whale of a compression ratio. We've never built machines like that before. The ones here going from 1 atm to 10's of atm are AT MOST 70% efficient. Efficiency decreases SHARPLY as required pressure ratio rises.
Sources of solid-CO2 dry ice near or at the poles would be a whole lot easier to utilize: you just mine it. You do have to be careful of sublimation in uncovered deposits, that's all. Once indoors and warmed up, it's a gas at atmospheres to tens of atmospheres. Very little inefficient gas compression is required, IF you plan your process correctly. This sort of thing WILL NOT be possible at all the interesting landing sites, only those at or near the poles, where the dry ice is hidden amongst the water ice (the OTHER very thing we are looking for).
This is exactly what I keep writing about. No two sites are alike in resources. The point of "exploration" is to go and find out for sure (1) what all is there, and (2) where exactly is it? This is to support any possible future activities utilizing those very same ill-distributed resources. What if there once was life on Mars? There might be coal, oil, and gas deposits. Who knows? Nobody, yet!
It would be just about as easy and expensive to tote all your propellants for the first mission(s) from Earth, as it would be to try to make carbon-based chemical propellants (methane, etc) in-situ from such a thin atmosphere, local ice deposits notwithstanding. That's because of the 10^4 compression ratios required. Ridiculous prospect, technologically.
Now, hydrogen and oxygen from water only, that's a different picture. Ease and feasibility depends on ice deposit THICKNESS and PURITY. If favorable, then how easy and efficient that process might be! Even if restricted to solar PV efficiencies under 10%. You do not learn about thickness and purity from orbit, or by scratching the surface 10 cm deep.
GW
Now, now, guys! Civility!
I just posted on another thread some of the cost figures per pound of payload to LEO from Spacex's page a few months ago. We can do a lot better than $5000/kg I saw posted here a conversation or two back. That corresponds roughly to $2400-2500/pound. That's Atlas-5 at max capability 20-25 metric tons to LEO, and the same for Falcon-9 at 10 metric tons to LEO.
Spacex's projections just a couple of months ago for Falcon-Heavy are 53 metric tons to LEO, which works out to $800-1000/pound (roughly $1600-2000/kg). I doubt very seriously the new government design could ever even possibly approach that cost figure, in spite of being 100+ tons, and there is a scale effect. It's shuttle derived hardware and ways of operating, which derive from shuttle at $1.5 billion per launch of 25 tons max. Not carrying an orbiter will help, but not all that much. You work it out. Ridiculously expensive.
I'm for using Falcon-Heavy, not screwing around with some ridiculously expensive government design.
GW
Many months ago I looked over Spacex's page and investigated Falcon-Heavy. Back then, it was projected at 34 metric tons to LEO. More recently, they're projecting 53 metric tons, with an uprated Merlin engine variant. Propellant cross-feed helps squeeze out all the performance they can get. The price per launch they quote works out at $800-1000 per pound for delivering 50-53 tons to LEO, by far the cheapest in the industry.
Atlas-5 at its max capability (20-25 metric tons) looks an awful lot like Falcon-9 at 10 tons, both near $2400-2500/pound delivered. There is a scale effect here: larger rockets deliver more, but don't cost that much more to launch. That means Spacex is already doing better down in the 10-ton class by around a factor of 2.
They did it by logistics, not reusability. Smaller, leaner company, and a design that requires a village, not a major city, to support each launch. (Reusability would help lower costs further, except I seriously doubt it can ever be achieved at the 4-5% inert mass fractions in Falcon-9 stages.) That can be taken further still, if the launch folks would talk more with the missile folks about simplifying and designing for very small support crews.
My crude investigations say it would be easier to achieve reusability in 3-stage to orbit than 2-stage to orbit. With 3 stages there is a lot more room for higher inert fractions. That's what has to "cover" the structural survivability "beef", and all the added recovery gear (whatever it is).
GW
Why not use the experience we have from Mir, Salyut, and Skylab, as well as ISS? The unique one was Skylab. Alone of all of them, there was a huge open space in which to live. That would be an upper design bound on volume per person. I'd use Salyut and/or Mir as a lower design bound on volume per person. ISS falls in-between, and seems a tad crowded sometimes, with 6 on board.
What that says is any crew module for the long ride to/from Mars is going to be a big one. Not necessarily really heavy, but voluminous. There needs to be some real elbow room inside, like Skylab, and some spaces where individuals can go to get away from everyone else. I's suggest a few Bigelow-type inflatables docked together would be the most practical way to launch it, using Falcon-Heavy at $800-1000/pound. I'd also suggest recovering it and using it on subsequent missions to Mars and elsewhere.
GW
If these paper airplane things really were released, I suspect most of any survivors went into the ocean. I never heard about any results though. Fascinating experiment!
As for Falcon tanks "exploding" during entry, I am not surprised. Heat shield cork or not, with inert mass fractions in the 4-5% range, these items are quite fragile. A tumbling cylinder is going to get crushed from the side by stagnation pressures as it tumbles broadside. High internal pressure could stave that off a bit, but stopping the tumble with a drogue to take the loads end-on is actually more effective, and the toughest, heaviest part of your heat shield can be smaller.
Spacex's cost reductions come from a smaller logistical support tail, not from reusability. Falcon-9 at 10 metric ton payloads has the same price per unit mass delivered as Atlas-5's max 20 ton payloads (both near $2400/pound, if memory serves). Falcon-Heavy will beat Atlas 5 by about a factor of 3 on unit price, and deliver more than twice the mass at 53 metric tons. And that's without effective reusability. Small logistical tail is the real driver for cheaper access to LEO, not reusability.
Reusability might help, though. (But, I really doubt it would be more dramatic than what Spacex achieved with its smaller logistical tail.) Maybe the first step is detach and save the engines only. Use the tankage sacrificially to protect the engines during reentry, then detach and parachute the engines to the sea. They'll have to be tough engines, sea water does bad things to hot metal. They'll need a float, too.
GW
Shuttle leading edges went to 3000 F and required carbon-carbon composite precisely because the wing loading (vehicle weight divided by wing planform area) was up around 1 or 2 hundred pounds per square foot, just like a fighter jet, and all the space capsules. The white tiles on the sides and upper wing surfaces, and the black ones on the belly and lower wing surface, were low-density alumino-silicate, with a solid phase change that causes cracking at 2300 F. Those were restricted to peak 2000 F skin temperatures on the shuttle. Carbon-carbon is weak enough structurally, to be sure. Those tiles were far more fragile yet.
If the vehicle has a much larger aerosurface for its weight (low wing loading, say 10-20 pounds per square foot), peak skin temperatures reduce to under 2000 F, although total heat to be absorbed and disposed of actually increases. Skin temperature drives the material selection problem, the other is handled fairly easily. You will be decelerating at higher gees to make this happen. Plus, it's all a transient.
Myself, I rather like the idea of enduring a bit rougher ride in order to make my re-entry vehicle out of simple aluminosilicates, perhaps even plain old fire curtain cloth on a steel tube frame. I think it is very funny that the better, less fragile reentry vehicle might actually be built similar to the venerable old Piper Cub of the 1930's.
And, low density ceramics should be fiber reinforced as ceramic-ceramic composites, not that fragile stuff on the shuttle. I have done this, a quarter century ago. They are still extremely low density, yet fairly tough structurally. Really tough compared to that fragile nonsense they flew on shuttle. I used mine as a ramjet liner, which survived hours of burn and hundreds of excursions into very violent rich-blowout instability. The only reason I quit then was the project was done. Could have gone on for many more hours.
As it turns out, the materials I used then are still available. I checked just the other day.
GW