New Mars Forums

I kinda remember that. My main point of contention is that you are trying to marry an air launch with a SSTO and still get SSTO structural fractions. I imagine it would be much, much easier (perhaps even doable) to keep the two components separated, and do a straight air-launch. Kind of like the new concept that has been flying around by Scaled Composites and SpaceX (and a few strange names thrown in for fun). As to why, here are my main disagreements:

The reason for that is rockets don't have turbines placed right after the combustion chamber. Which, by the way, is the main limitation in jet engines, the fact that the first stage of the turbine can't melt. Neither do ramjets and scramjets have them, btw, but don't plan on having any turbojet combustion chamber double-duty as a rocket combustion chamber.

You want to add not only jet engines, but wings, undercarriage and the works, and still get SSTO mass ratios? Airplanes have structural fractions upwards of 30%, and there is a reason for that. No way you can do all that at 10% structural fraction.

You list every advantage that this "M3 airlaunch carried to orbit" gives you, and they are indeed great, but you gloss over the myriad of "inconveniences" that getting that boost entails, which is almost like carrying a SR-71 to orbit with you. Why not letting a lower stage carry the inefficiencies away at staging and get only the benefits in a true air-launched second stage? That would make much more sense than this. Get all the hardware necessary to do all the things you say, and instead of sticking it to the SSTO, build a plane our of it.

Rune. BTW, before you point to skylon's SABRE, that's different. Not the same cycle as a turbojet, and separate combustion chambers for the rocket part going into the same nozzle.

Yup, it is very cool indeed. And, pardon me for being a bit devious, perfectly released to compensate the bad press of the latest delay for C2/3. But, you know, I did enjoy it a lot, in both ways.

Rune. Maybe even more in the devious way, because we need good PR about spaceflight.

About recovery operations, parachutes, and Dragon:

First off, I agree with most of what you two say, let that be clear! We are so much in agreement, actually, that I feel compelled to point out something, if only to keep the discussion going.

So the first thing to point out is that you both assume Dragon is going to use the parachutes when/if the manned version flies (I'd say when, but you know). That doesn't really seems to be the stated intention of the company, if you go by their CEO's public claims. IIRC, of course, Mr. Musk has been saying in a couple of different webcasted press conferences (just google for one that mentions the reusable Falcon) that the manned version is going to have "pinpoint accuracy, on the order of meters". Now we all now that if you go in by chute alone, you get a dispersion parabola because of it, and no "last-second" engine firing is going to reduce that by any measurable fraction. That kind of just rules out the pad landing shown in the famous video, too.

Also, I distinctly remember him pointing out that a parachute, singular, would be retained to provide a backup and be used in the case of an aborted ascent. I remember, because that's when I found out the Dragon of today is designed for a double failure in the chute system. If that is the case and you read him literally, then that "backup" means the chute won't open during a nominal landing. Full Heinlein-like magic, which would cut the recovery cost to basically nothing. But a nasty environment for plants to grow near the pad, for certain.

Now I imagine I should explore a bit the practicality of it all, so there I go. First off, how much fuel can Dragon take, in excess of what it carries now, which I will assume is enough for a nominal mission plus reserves up to landing? Well, the downmass of the uncrewed one is 3,000kg. We should take away the weight of up to the seven stated crew members, so 700kg? Cal it one ton to account for life support equipment. The additional engine weight of the super-dracos that are under development is a full mystery right now, since I can find nothing on their T/W ratio. But 1mT seems enough, right? Even considering tankage, that's not asking much more than 30 out of it, since a fully loaded Dragon can't exceed ~10mT. First order of magnitude work here, expect it to be more than a bit off, of course, but I'm taking the pessimist side. For Isp, who the hell knows, but similar engines give about 300 vacuum, and data on sea-level isp is pretty much impossible to find, so I'll just pull 275 out of my... let's call it hat (Encyclopedia astronautica, you have failed me!). Straight good old Tsiolkovsky gives about 285m/s out of that, or a shade over 1000km/hour. So... yeah, if Dragon goes subsonic on it's own without chutes, then I call it plausible.

But then again, I have no freaking idea about that, so I take it all with a big grain of salt. And invoke the knowledge of my elders to fill the blanks I left, too.

Rune. I would kill for a look at those super-Draco designs, my curiosity is killing me on this. It would look sooo cool and sci-fi-ish.

Or a reaction control system. I think you might be a little too much focused on either runway horizontal landing or sea landing. What's wrong with vertical powered landing? Slowing and landing a subsonic, mostly empty, stage propulsively is a piece of cake, both thrust- and fuel-wise. And, you don't have to add either wings and undercarriage (10% empty weight? Plus heavier TPS on the sharp leading edges), or floats and the structure and corrosion treatment to survive the ocean (bad company to reusability, ocean immersion). You just need lots of engines or deep throttling, and bigger tanks. Yet I don't recall to read you ever suggesting it, except for Mars landers.

Rune. We pretty much agree on everything else.

That, at first glance, seems to be a very complicated way of achieving nuclear fission. And it's got to create radioactive material, of course, if it is (slow neutrons, decay... yeah, radiation for sure). Is it expected to have any benefits at all? I imagine the only point is that all of that can happen at sub-critical concentrations of non-nuclear (at the start) materials, right? Oh, and the nice science, of course.

Rune. I don't think that is what that Rossi guy is doing, at all.

Correction: a video on youtube says that. Big difference.

By whom?

Rune. Really?

I took my sweet time replying, but I hope this is worth the wait for you guys. Here are my very rough numbers for a nuclear MTV, as you suggested, for a hypothetical mars mission. Or, as it is single stage and starts and ends its missions at LEO, with 20 restarts on its core before the engine needs replacing, regular martian ferry trips. Enjoy, and don't expect great accuracy, just a rough first-order-of-magnitude sketch.

I have divided the work in, roughly, 6 areas, namely payload (1), engine (2), delta-v and therefore mission profile (3), which gives our mass ratio (4), structural and tankage fractions (5) and then the resulting numbers that come up from all that at the end. Plus, I'll throw out some of the conclusions I get from this grossly unscientific thought experiment at the end. Without further introductions:

1.What's the payload? I'm going to go with a 50ton module. Enough to fit a Transhab, Dragon, and tether system to rotate the whole thing for artificial gravity (with room to spare, Transhab is only 37mT), if it is a crewed flight. Or a decent-sized habitat or lander/ascent vehicle. If you go with the optimist view of things, I think you could use MD's numbers and fit a hab and a lander there, for a certain crew size. Or you can't, and therefore two flights (or two ships) are required to land both a hab and leave a lander on orbit. Also, 50mT is a nice round number, and it fits the lowest cost heavy rocket expected to be in service in the near future. If I cheat a little and look at the end of the post now it's fully written, I can also tell you that if we don't have to bring the payload back (as is the case with surface supplies and landers, and pretty much everything but the hab), the hydrogen-fueled ship can put a little over 115mT ton in mars orbit in this "cargo" version, while the methane-fueled one can deliver over 166mT. In any case, more than enough to deliver both a surface hab and a lander in a single flight, even if the lander is fully fueled.

Note: I will admit it's weird the methane-fueled ship, with it's lower payload fraction, gets a bigger benefit from returning empty, so if someone wants to check my math (did it twice myself), I just worked out two different mass ratios for the two legs of the trip, backwards and without counting the payload on the first one, I'm sure you get the idea. Might just be the mass ratio margin is just bigger for that rocket.

2.Ok, so for the engine I'm going to pick a good old solid NERVA design, with the fuel elements substituted with modern-day CERMET materials and a few overall tweaks bringing the isp to 950, like you said, and maintaining the overall weight and T/W ratio (I think this is conservative enough the resulting engine is reliable as hell). So I get to it, and what's my surprise, DoD did all the hard work fro me. Turns out there is a NERVA gamma, a design study based on the alpha engine that was actually built and tested, with some material upgrades added, including the new fuel elements and an electrical power conversion system. It's also nicely tiny at 3.4mT. The only problem, I am making them run for x4 longer firings. Obviously someone was thinking 0.5G when they designed this engine. Well, I hope they are really well cooled and the core has energy to spare. Here it is:

Engine: 3,410 kg (7,510 lb). Area Ratio: 100. Propellant Formulation: Nuclear/Slush Hydrogen. Restarts: 20.

Status: Study 1972.
Height: 4.46 m (14.64 ft).
Diameter: 1.24 m (4.06 ft).
Thrust: 81.00 kN (18,209 lbf).
Specific impulse: 975 s.
Burn time: 1,200 s.
First Launch: Designed 1972.

The thing eats hydrogen, but the original NERVA was tested with a variety of propellants, including methane. This is important, since cryogenic propellant management technology is in its infancy, while the rest of this work relies on nothing more advanced than what we had in the 70's. Extrapolating the numbers from the old NERVA gives you an isp of 761s in this version of the engine running on methane, and I will keep it in mind as a desirable option. I really don't know what is the lower limit on T/W to call this a "fast burner" that benefits from the Oberth effect significantly (or doesn't incur in excessive penalties), but I reckon at the very least it's got to do significantly more than 0.1G. So for every ~100mT of ship, put a couple of them to have a nice 0.16G acceleration at full weight and plenty of in-built redundancy at <7% weight fraction. Oh, and 20kw in electric power to "run things". Keep in mind that if we make it run on methane, we are blowing considerably heavier gas in our exhaust. If the reactor power was increased, for the same turbopump and related machinery, we would get considerably more thrust, but I won't get into it (even if it makes methane a much less attractive option, T/W-wise).

3.So if we want to get this thing to mars and back, preferably single stage and without needing pesky heatshields so we can aerocapture, what would the delta-v budget look like? From LEO to a Mars capture orbit (big ellipse with great eccentricity that comes within a hair's breath of the atmosphere at the closest point), all propulsively, about 5.2km/s according to wikipedia. You can later, over an extended period of time, slowly aerobrake your way into LMO proper for landing, with perhaps extended stays at selected orbits to study the martian moon's through flybys. Or not, but it increases the lander's reentry speed.

Form LMO to a suitable LEO, 6.12km/s, and you brake propulsively all the way, even though you could just capture propulsively and aerobrake over a few orbits just like you did at Mars, again without any heatshields or extreme precision involved. But I want some margin somewhere, and the astronauts might be in a hurry to get home (and this increases the mission time, at mars you are just cutting on surface time, and there's plenty of that). By the way, I have just pinched numbers from here, if someone wants to work out the real numbers, assume a launch date at least 5 launch opportunities away and work out the actual orbits yourself. But since I used such an approximate approach to everything else, I wouldn't bother.

Incidentally, since the payload requirements I laid out require either two different cargo flights in separate windows for the same ship to be reused (and flight-tested) or building two separate ships (or more, each costs a bit less than the last one), you could go with the expensive safe option and send two manned ships with half the crew each to add more redundancy, and make the lander/ascent vehicle the critical choke point in mission safety (I should work out lander weights one of this days, but it is not that day).

4.Anyhow, adding both up gives you 11,32km/s, which could be a nightmare to fulfill with a single stage chemically, but is a piece of cake four our stupendous nuclear engine. Working the rocket equation, the mass ratio comes up to a shade less than 3.3. Call it 3.5 to have plenty of margin, and fuel to park our ship around a martian moon if the fancy strikes us. That is 2.5kg of propellant for each kg of cargo. If we go with a methane-fueled engine, like I have advocated for a long time, the mass ratio ends up being 4.55, call it 5 to have some margin. 4kg of fuel for each one of cargo, not such a bad trade, considering the fuel is a gazillion times easier to store without no losses, and considerably denser, too, so it would require much lighter tanks. Potential for future ISRU is also greater, but I am sure that thought crossed your mind already.

5.We would ideally like this to be reused a bunch of times, and sturdy as hell to cover any design flaw behind walls of redundancy. So give it a 20% structural fraction. But wait, you say, if you just said the ship has a mass ratio of 5 if methane fueled, that means you just killed that option, right? Well, yes if we built our fuel tanks as sturdy as our main pressure hull. But I am not going to do that. The way I see it, the tanks could be dropped after each mission, be modular and 50mT in weight when fully fueled, and have a structural fraction by themselves of perhaps 10% for the heavily-insulated hydrogen ones. So I'll only count towards structural fraction the unfueled weight of the ship: the payload and the engines, plus the various communication/attitude control/navigation gizmos (a couple of tons at best? I have still budgeted around 6-7mT for them, around 1% of the fully fueled ship, just to be on the safe side and provide both a decent margin and nice round numbers). To put things harder for the methane ship, I'll use the same 10% fraction for the methane tanks, just in case that makes them refuelable at a Phobos station eventually or something. Or good habs to reuse.

So how does it all look like when put together? It is mostly a trial and error process from here, so I'll just give you the results:

With Hydrogen fuel:

Total fueled weight:  525mT
Empty weight:         150mT

Mass breakdown:

Payload:                  50mT
Engines (10):            34mT (T/W:~0.16)
Structure:              22.5mT
Comms/Nav/Margin:   6mT
Tankage:               37.5mT
Propellant:             375mT

With Methane fuel (and a relaxed T/W, 'cause maintaining the ratio increases the mass to absurdity):

Total fueled weight: 1.250mT
Empty weight:          250mT

Mass breakdown:   

Payload:                  50mT
Engines (15):            51mT (T/W:~0.1)
Structure:                42mT
Comms/Nav/Margin:     7mT
Tankage:                100mT
Propellant:           1.000mT

Well, that was that. I have the numpad smoking, that took some iterations to get round numbers (feel free to make up your own)... and correct the stupid mistakes, too. What do I get out of all of this, then? Several things.

First, T/W is important, especially as you get over your exhaust speed in delta-v requirements. The methane case exemplifies it perfectly. The increased mass ratio, though it doesn't seem like much at first glance (3.5 vs 5), ends up costing us a ship two times as heavy with half the acceleration (which will translate in performance losses, and significant ones), because we have to maintain the structural fractions in order for both ships to be as durable and safe. Ok, maybe I over-exemplified it since I used the same weight fraction for both tanks, and a methane tank designed for zero-g could have a structural fraction well below 7%. And the fact that even if it weights four times as much as the other ship, it will be more compact on account of hydrogen being six times less dense.

But you know, the point still stands, this is a mission that is right at the edge of what can be done with <800 isp. On the other hand, this is a mission that is perfectly doable with such an isp, because once you think about it, this ship could be assembled in about 24 flights (20 of them just dumb full fuel tanks to be picked up by the ship), for a launch cost of perhaps ~$2.5B using Falcons. Not bad, not bad at all, considering the R&D cost for the couple of hundred tons of really complicated, aerospace-grade, rest-of-the-ship. Long-term, when we are routinely performing ISRU refueling, this is clearly the way to go at least in the vicinity of mars.

Of course, if we can manage the cryogenic handling of propellants, then oh boy, do we have a ship indeed. Launched in just 11 FH's, refuelable with eight more. A very sustainable launch rate of four FH's each year, plus one every two years for payloads, could maintain a steady stream of missions, each leaving some hardware behind. Add another to loft new engines and spare parts now and then (the engines run out of restarts after 5 round trips at most), and it is still not much. Call it 5 launches a year, for a launch budget of ~$0.5B each year, or about a sixth of the shuttle program. If an engineer with experience on the subject told me the cryogenic problem is solvable within reasonable costs, this is the certainly the way to go in the initial stages of exploration. Later, in the far future, it may very well be that we have become masters at Hydrogen handling and we don't need to switch propellants, or that the physical laws that say "storing liquid hydrogen is difficult" will prove insurmountable and we just switch to something easier to produce and store everywhere (hello ammonia/methane/steam rockets).

Phew, quite a big post in the end. Hope you enjoyed it!

Rune. Or endured it till the end at least... ^^'

Edit: Spotted a couple of mistakes. Didn't change the conclusions on bit.

Ok, I was in the process of replying, then I realized the foolishness I was about to defend. Well corrected, Hop, no window trade. You can smile smugly with my official blessing.

Well, I kind of got what you wanted to say then, I just thought it was better to just jump to the conclusions and move on from there.

However, maybe I'm a little less turned on by the idea of an all-nuclear solid core launcher. It's not really the radiation, too, although maybe if we used nuclear rockets with the frequency I would like to see rockets used, we could have a problem there. But a nuclear rocket would only make sense to go full SSTO, and then you would run in the usual T/W and throttling problems. I know Timberwind makes it seem as if the T/W problem is surmountable (I think a switch to methane propellant would do wonders, too), but Timberwinds are really complicated things, what with centrifuging the fuel (the nuclear fuel very close to melting temperature) at thousands of RPM's to make the cooling system work, and not being restartables or have a significant design life or throttling capability. Kind of makes the whole reusability thing a bit moot. I'd like more a durable, sturdy design like the old NERVA. Small number of moving parts and active systems (mostly just the turbopump, valves, and the control drums), modest T/W, several restarts, long core life, can even accept several different fuels with no problems.

More sensible, IMO, is to provide a first stage to ease the T/W problem. Some dense liquid or even a solid (though I hate them on reusability grounds) reusable first stage would work better. Plus, it would let you expand the range of places we can launch to directly from the ground, with designs with up to 50% more delta-v in them than current 2-stage vehicles of the same size (and similar mass ratios). You could launch a Dragon-sized capsule, for example, with a Falcon 9-sized rocket, to L1, GEO, or lunar orbit. Or launch more to LEO, or use a smaller rocket, of course.

Upon landing, thrust isn't really an issue, so here's when I appeal to your lack of nuclear fear and ask you to let me land these second stages (almost empty now, so T/W isn't an issue even though the engines run at close to 100% thrust)... "like God and Robert Heinlein intended them to".

Rune. How else?

It's kind of cute how you all forget the ground and communications infrastructure. In order to pull any of this off, you need constant two-way communication between all the pieces of your puzzle, for starters. So make it a few (3+spares) geosynchronous satellites, a couple of ground rely and tracking stations (and only a couple because of the sats), and a lunar orbit satellite constellation (quite extensive if you want to cover the poles, tens of birds at the very least). All of which need monitoring from real people on the ground, and eventual replacement (10 years is a good average for the life of a comsat, and the cost of a single engineer job for that amount of time is greater than the sat's, but the same order of magnitude).

Not to mention the couple of space stations you are suggesting here and the abundant traffic of fuel tugs (which do wear out). BTW, Terraformer, how many tons of propellant produced for each one delivered to LEO? And to LLO? And by what (I hope) single stage method? Just curious to see how the fuel economics would work, maybe you end up needing a huge powersource on the ground to support all of this activity (and more mining gear, and more crew to handle it, and so on), maybe you can make do with less. Also, am I correct in assuming you envision two fuel depots/transfer stations, one in LEO and the other in LLO, with all the fuel supplied from the moon? If so, consider the trade with a single station in L1, in terms of launch windows. Something like once every two weeks from a particular orbital plane in LEO to a particular orbital plane on the moon, IIRC? No idea off the top of my head, really, but I do know L1 is accessible once every orbit, which is 90-something minutes in LEO, or pretty much anytime in other words (once a day from the ground).

In any case, expect a ground support staff of the approximate size of ISS's, if a government is involved, way more if several are. Even a SpaceX-style with "eight guys on a trailer" will become huge quickly, regulatory hurdles aside. I won't get into revenue, 'cause that's frankly not my thing. Somebody else hunt for the contracts, I like the challenge of design.

Rune. The little inconvenient middle steps are a bitch.

I just saw something funny in the public manifest over at spacex.com, check it out:

http://www.spacex.com/launch_manifest.php

It's this part, specifically:

Never mind the fact they are planning to launch 14 rockets that same year, 13 of them Falcon 9's... they have one booked with Bigelow! And it's no Dragon flight, so they are taking hardware up. And here I though they didn't have anything sized for spacex's rockets. It seems we get an inflatable ~10mT space hab in three years. First piece of a commercial station, or further testing of the hardware after so many years without building one? (And firing half the company that built them, too).

Rune. Hope that manifest is actually accomplished. 13 10mT launches... that's a lot of payload. More than China did this year, or anyone else for that matter (though the entire US market comes close if you count the last shuttles, I just checked).

I'm with you on the logistics tail part, of course. And part of that success stems from a simple design (one fuel, one engine) that allows to simplify production. The thing is, they started out with a simpler design still, pressure-fed engines without turbopumps, but got away from it early on grounds of performance. Of course they came up with a pretty amazing engine instead, but I'd bet a Merlin without turbopump would be even cheaper to build, even if it doesn't break any records. Dunno, maybe the isp and weight fraction hit makes TSTO completely impractical, and they really wanted to keep the number of stages to recover controlled. But I "somehow" doubt it's because stronger tanks made of inexpensive steel, not those fancy lithium alloys, are more expensive...

Also, I'm a bit dubious of ocean recovery, since it requires both specialized equipment and stress on the hardware (read: corrosion, but also the water impact) leading to additional work to "process" the stage for a new flight. Every engineer hour spent refurbishing a rocket could be spent building a new one. So in order to have a truly effective reuse you have to minimize the processing effort. Ideally, that process would be as simple as hooking a computer so the internal systems verify everything is OK, doing a test fire, stacking, filling the tanks and charging the batteries, and nothing else, at least not for dozens of flights. Well, maybe the traditional visual inspection before flight and some ground transport by other means. I see that happening only if either the stage lands horizontally on a runway or vertically under rocket power, and since wings are for airplanes and not for rockets, vertical landing it is, IMO.

Rune. Short of a low-tech staged DC-X. As many stages as it takes to make it cheap and lasting.