You are not logged in.
I'm not familiar with the M2P2 concept, myself. I had in mind either ion or plasma jet thrusters, combined with the lighter-weight solar power being discussed here. Although, this M2P2 thing is another candidate.
The biggest problem with any of the electric schemes is super-low vehicle acceleration. It takes "forever" to leave a parking orbit for a transfer orbit, and the gravity losses associated with the very long burn times are enormous.
So why not do the burn from parking orbit to transfer orbit impulsively with chemical rocket, and then shorten the transfer by using the electric to accelerate on the transfer to its midpoint? Then decelerate electrically the same way during the second half of the transfer.
There's still gravity losses with the electrics, but, at least the transit time is much shorter. Your vehicle will size out somewhere in between the big all-chemical system and the small all-electric system.
Reductions in both size and transit time. Is that not the win-win that we all want?
GW
A logical question: how is using only circumlunar-capable manned technology a "steppingstone along the way to Mars"?
Another logical question: how is $6M to be construed as significant support for efforts really aimed at doing anything? This sounds more like buying PR with small business set-aside monies.
GW
Quaoar's "enflatable" in post 346 just above is a fuel bladder. That is an old technology long used in aircraft fuel tanks and some missiles. It works with room-temperature-storable-type propellants.
There are always possibilities, of course, but I know of no elastomerics suitable for bladder service with cryogenics. That means there are no such technologies available "off-the-shelf" and ready-to-use.
If you burden a new vehicle development program with new technology development, it never actually flies. That's been the history of such things. For a vehicle you really want to fly, pick only off-the-shelf technologies. Leave the new technology development for programs specifically so intended.
Just advice from an old development engineer.
GW
Multiple mile-long unit trains (coal and ag products) come through Central Texas every single day, plus an Amtrak (the Chicago-Laredo route). There's also more mixed freight than I have seen in prior years (a good thing, considering how badly clogged I-35 is with big rig trucks: the stretches near Waco are the most dangerous stretches of highway in the entire US).
I suggest that freight traffic and pre-existing Amtrak justify the kind of service I propose, even here in wide-open Texas. More folks would ride that Amtrak if it were an auto train (like the NYC-Orlando train). The tracks here in Texas as they are today are good for 80 mph at least. That's the intercity speed of Amtrak right now through Texas. The speed limit on I-35 is 75 mph outside the cities and far away from the DFW pollution nonattainment area, where it is 65 mph.
What is needed is merely to add an Amtrak option to the mixed freights and some of the unit trains, an organizational/conceptual thing, not a hardware acquisition or upgrade. Flat cars already exist. Auto carrier cars already exist. Dining cars already exist. The tracks are good enough in Texas to get started. It's political preconceptions and dogmas that prevent doing what's smart.
Track maintenance is not that big an investment to raise speeds to 100+ mph. As I said above somewhere, we used to go that fast with passenger service in the days of steam locomotives. 4-6-4 steamers on the New York Central line were pulling around 6-12 passenger cars at around 110 mph as early as 1900, and as late as about 1950. You can tell a real high-speed passenger steamer by the large-diameter drive wheels (72 inch and larger). Freighters had smaller drive wheels (under 50 inches) and pulled much longer, heavier trains. Simple torque vs linear speed trade at constant rpm and steam piston speed.
There's an old 1909-vintage steamer on the Texas State Railway tourist line that typically pulls at 30 mph. Its speedometer reads to 120 mph. The limit isn't the engine and its train, it's the tourist-line track: too distorted with age and wear, not a smooth enough ride to be safe at high speeds. I rode in its cab a few years ago. Given some instruction in start-up and shutdown, I think I could drive or fire it today.
That's nostalgia talking. We can do this with the existing diesel engines. They're easier on the track. Just stack up enough units to pull your mile-long freight at whatever speed you want. Distributing engines through the train helps with derail resistance in turns. Simple as that.
GW
RP-1, Jet-A, JP-5, and K-1 are all just about the same petroleum product. The real differences are just filtering cleanliness and additive packages. They stink, but not quite as bad as diesel. The residues persist, and continue to stink. I doubt a kerosene tank could ever be vacuum-cleaned well enough to use as a habitat. The heavier fractions cling to things, in spite of vapor pressure-driven cleaning into vacuum. I don't know for sure, but pure single substances like liquid hydrogen, oxygen, and methane might clean-up far better. But the caution is porosity: propellant soaked into small pores won't vacuum clean very well. Even metal tank walls have this porosity. That's why you never weld on an empty gasoline tank, no matter how many years it has been dry.
GW
Excelsior put his finger on acceptability of passenger rail in the US. You have to drive to the departure terminal, and pay to park your car there. Then you have to rent another car to get where you are going at the other end. In a big, spread-out country like this, that doesn't work for people. Fix that, and passenger rail becomes attractive.
I put a suggestion for this up on http://exrocketman.blogspot.com as "Rethinking High Speed Rail to Make It Work", dated 1-30-14. You don't need really "high speed", just faster than the interstate highways. Do combined freight-passenger service, and run it at about 100-120 mph on the tracks we have. All you need to do that is good track maintenance. Yes, it can be done, we already did it: passenger service was typically 90-110 mph with steam a century ago in this country.
Passenger service gets added by a dining/club car at the tail of the freight. You drive to the station, and onto a flat car to be chained down, a flat car tagged for your destination. You go wine/dine in the club car for the trip. Taking your car with you completely fixes the inconvenience problem, and also gives trains a way to compete with air travel that the airlines cannot match. You do the load and unload of automobiles at the station, decoupled from the train. That way the train stop is short, and divorced from automobile load/unload operations.
Doing it as mixed freight-passenger "hides" the cost of hauling the automobiles as a tiny trivial item amongst the mass of all the freight. Trains already run 70-80 mph in Texas, even freights, and about 10 mph faster than that in Illinois (saw it riding Amtrak to Chicago a few years back). In some states, the maintenance is too poor to travel more than 20-30 mph. Fix THAT, and my out-of-the-box passenger service concept could be implemented far cheaper than anything I have ever heard proposed.
GW
There are good technical reasons why you want dense propellants in (at least) the first stage of a launch rocket. Both kerosene-LOX and liquid methane-LOX qualify, LH2-LOX does not. Kerosene-LOX is older and more developed. Liquid methane-LOX offers higher ISP at not much of a density disadvantage. Storables such as NTO-MMH are more-or-less almost comparable to kerosene-LOX.
Atlas-5 uses kerosene-LOX, then a LH2-LOX upper stage (the venerable old Centaur). It's only real problem is a Russian-made first stage engine in a time when we and the Russians are no longer cooperating. For heavier-lift, you add solid strap-ons. The higher Isp of LH2-LOX overcomes its density disdavantages in the smaller upper stage. It came from a single-stage-and-a-half Atlas ICBM, to which Centaur was added as an upper stage.
I'm not sure without researching it what Delta-4 uses. But I do know that for heavier-lift, you add solid strap-ons. It originally came from the single-stage Thor ICBM.
Falcon-9 uses kerosene-LOX in both stages. No strap-ons, for heavier lift, you stack up more first stage units in parallel and cross-feed their propellants (Falcon-Heavy). This one was never a missile design.
The SLS design is the LH2-LOX design, with solid strap-ons. It's based mostly on shuttle engine and SRB stuff, and made to resemble the old Saturn-5 in the way it is stacked together.
The old Saturn-5 was kerosene-Lox in stages 1 and 2, and LH2-LOX in the third stage. It originated not as a moon rocket, but a family of ICBM designs. The Apollo-Saturn moon mission architecture was made to fit around the biggest of the ICBM designs. It took an idea from outside NASA (lunar orbit rendezvous) to make it work at one launch-one mission. Before that was adopted, each moon mission required two Saturn-5 launches. The Apollo-CSM was the lander.
GW
Servicing Hubble was easier to accomplish with a maneuvering spacecraft (the shuttle) that could go to the Hubble, a structure to which to secure it (the "table" in the shuttle bay), and a manipulator arm by which to capture and position it.
The mistake was launching those three items for every mission, when the three critical items could have been launched just once, and manned by a much smaller rocket-and-capsule at those times it was needed. We could still do that.
As to other uses, like all tools, it can be used or abused. Of course, anything providing such a capability could be used for space warfare. Existence of a technology does not provide proof if its actual intended use, but it certainly provides opportunity for any such uses.
Which is an example of the old adage "pay attention to what they do, not what they say", or in shorter form, "actions speak louder than words".
GW
"You do meantion coal and at one point earth did rely heavily on it for powering until oil was found." --- even today, the majority of the electricity made on Earth is made using coal.
What made wood, coal, whale oil, and petroleum oil feasible and economic to use on Earth was our oxidizing atmosphere. By far, the largest massflow through any sort of combustion device is the oxidizer. All we had to do was find a fuel.
It's not like that on Mars: the atmosphere is inert. You have to find or create both fuel, and the far-more-massive oxidizer. I have my doubts about combustion engines ever proving to be truly practical there, precisely because of that problem.
GW
Bob:
What about combining solar-electric with conventional propulsion? Do the main burns conventional to leave orbit. Then apply solar-electric during the transfer, to vastly shorten travel time? Might that be the best way to avoid the slow spiral-out problem?
GW
Quaoar put his finger on the needs for going to Mars or anywhere else further than the moon. That's the habitat and protection technologies for long-duration flight. Up to a 3 years or so, we can use packed supplies as expendables for life support; beyond that, we'll need a functioning closed-cycle ecology (something we don't yet have). We'll need artificial gravity (not only for health protection, but also to make many life support functions easier, and for basic comfortable living). We have it with spin. We'll need radiation protection from solar flares, we have it with 20 cm of water/wastewater. Beyond about 3 years, we'll also need to counteract cosmic rays (something we don't really yet know how to do) in solar min years when the flux is maximum.
Problem is, I don't see any of those things being worked on in any significant way by any of the space agencies. NASA is focused on trying to find a mission for its politically-mandated reprise-on-steroids of Apollo-Saturn: Orion-SLS. "Is it Mars, or the moon, or is it something to do with asteroids?" Some years ago they cancelled the medical centrifuge module for ISS: so much for finding out "how much gee is adequate?" And instead of spending what little money there is developing the ways for men to travel long distances, NASA would rather develop an itty-bitty asteroid-capturing robot without the science to support credible design requirements, so that it need only use its moon rocket technology with the men.
Moon rocket technology. All that cash to reprise a technology we had, and then frittered away, 40+ years ago. What a waste!
And even then, the Saturn family was originally a suite of Army ICBM designs, not one of them a real moon rocket design. The big Saturn just happened to be barely big enough to use as a moon rocket, but only after some real cleverness got adopted from outside NASA (the lunar orbit rendezvous concept). Von Braun did the Saturns for the US Army, before NASA grabbed him. Apollo-Saturn mission architecture was made to fit an existing family of ICBM designs. People too often do forget how we got there.
GW
Hi Quaoar:
Well, in the mission plan those illustrations came from, no empty tank was discarded, at the cost of bigger-than-minimum ships. Most would be left in orbit about Mars for future but unspecified use. Some would be in LEO upon the ship's return. Those could be either refueled or re-purposed. It's so expensive to launch stuff, I figured why throw stuff away?
I suppose that rearranging the internal structure of such a module might make it more appealing as habitat space, once vented. There would have to be included some sort of access door through which to load the interior equipment one might want. The propellant transfer piping would have to be designed for easy removal, as well.
And, to utilize the space inside both tanks, there would have to be some sort of connecting doorway designed-in. I'd guess that most of this could be designed-in with low (but non-zero) impact on the inert structure mass fraction of the module-as-a-propellant-tank in its initial mission.
GW
Skylab was not launched with any propellants inside it, nor did it have any engines at all (which is in fact why it fell on western Australia). The Saturn 5 that launched it was two stages: first and second. Skylab was just dead-head payload made out of third stage hardware. The lost shield tangled tore off one solar wing and tangled the other into non-deployment. The first crew had to fix that. The freed the remaining wing and poked the sunshade "parasol" out through an airlock passage (don't remember which). Between the one wing and the Apollo telescope mount panels, there was barely enough electric power.
When the astronauts ran around the inside of those tanks, there was "artificial gravity", but only for so long as they continued to run. Use the astronaut's forward speed and the radius to those tanks, and you have the angular rate = forward speed divided by radius. The effective gravity is then radius x angular rate squared. You'll have to use consistent units, of course.
Skylab still had several tons of septic tank waste on board at entry. Best estimates of mass at entry were around 90 US (short) tons. I heard they picked up 75 tons of debris off the desert and some roofs in Australia. One big chunk was a 2-ton lead-lined film vault. Eyewitnesses said the main entering chunk didn't break up finally until it was down near Mach 1 at about 20,000 feet. That's about the conditions where Columbia's crew cabin finally crushed, by the way.
Agencies around the world lied about these things "burning up harmlessly", for years before, and years after, Skylab fell. The most egregious lie was told by the Russians about Cosmos 954, just before it fell on Canada. They said that a graphite nuclear reactor core "was designed to burn up on reentry". Total BS, that was.
GW
Actually I agree with all of you. The disjoint here is looking at NASA's plans and proposals, and expecting something that makes common sense. NASA is is entirely managed by congressional politics. That's why nothing making any real sense has been done for 4 decades now. Expecting otherwise is insanity (by definition), until the politics is ousted from control over NASA's objectives. The other countries' agencies suffer from their versions of the same problem, so expect no better there.
As to our discussions about what to do with asteroids and whether-or not to go to Mars first, consider this. If you don't limit yourself to a minimalist mission design, the Mars mission quite naturally takes the form of an orbit-to-orbit manned transport, with some provision for landers. It gets built and launched from Earth orbit. This is a very old idea, at least 6 decades. But it's a good one, still. One good thing from the ISS experience is that we learned how to dock-together large structures from launchable modules. That's the key to the orbit-to-orbit transport construction. On-orbit refueling techniques still need development further, but that's at least being done with small quantities of storables now.
Now, here's the thing. Any such orbit-to-orbit manned transport must be capable of supporting its crew on long-duration flights: months to years in space. Any such vehicle is equally suitable for manned visits to Mars, the main asteroid belt, NEO's, Venus, and even Mercury! So why not build just one (or maybe two), build it tough and reusable, and over time do all of those things with it? As time and money permit, and opportunities suggest? Further, as "hotter propulsion" becomes available, refit it, and go further yet: moons of Jupiter and Saturn.
If you have a long-range plan like that, you can amortize the cost of building the ship over multiple missions. Then each mission just needs to pay further for developing the specific landers and surface equipment peculiar to that destination. Landers for Venus and the asteroids are unnecessary. Landers for Mercury will look more like those we used for the moon, just bigger. Landers for Mars will look like some sort of propelled space capsule, because of atmospheric entry needs.
But you have to have a long range plan, and the freedom to do things that make good common sense (instead of politics). No government agency on the planet has those things. And THAT'S why men haven't left LEO in 4 decades.
Final comment: this can be done with commercial launch rockets we already have (counting Falcon-Heavy, and that looks like a good bet). A super-heavy like SLS could reduce the cost somewhat, but only if were a commercially-driven low-cost system. NASA won't do that with a one-use design, they never have thought about how to launch cheaper in any significant way.
Atlas and Delta were commercialized, which drastically lowered their costs to the neighborhood of $2500/pound in 15-20 ton payload ranges. Simplified systems and logistics did this. Titan never was commercialized, and was about 4 times more expensive than what was finally achieved with the other two, compared at the same payloads.
If SLS were commercialized, it would be under $800/pound, following the curves evident from payloads and launch prices exhibited by commercial launchers from around the world. Nothing I have seen from NASA about SLS suggests they will ever be that cheap. Expect no savings there. You might as well fly smaller modules with the rockets we have. Same or lower price. And no waiting, they are flying now. Makes good common sense.
GW
RobertDyck:
I was talking about the issues with NASA's asteroid retrieval mission, whether or not they ever really do it. What they had planned was not to visit with men a small NEO in-situ, they planned to use a robot ship with a can or bag, capture an itty bitty one, and drag it near the moon. Near the moon is a feasible mission duration for a crew to endure, crammed inside nothing but an Orion capsule.
If you're going to capture that itty bitty one, you'll have to de-spin it first, or it'll tear up the bag or can that you want to put it in. That requires pushing on it more-or-less tangent to the surface somewhere, to despin it. Either with a physical force or the force of a concentrated beam of light. That second option might (or might not) rely on thermal spalling to create a reaction force. Be that as it may, whatever force is used, it had better be weaker than the very weak gravity force binding the object together, or the thing will just fly apart in your face.
As for "killer" asteroid deflection, I first saw credible plans for this at the IAA conference on the subject in Granada, Spain in 2009. These plans revolve around either "little push" or "big push" methods. Little push is gravity tractors, small lasers, and similar. Big push is impactors and bombs.
There is no blast wave in space: nuclear bombs work from alongside as nothing more than super-intense flashes of light. That causes surface spalling, the reaction to which is your "push". Same thing would happen with a really powerful laser, just nowhere near as large an effect.
At that meeting, the concern back then was the unpredictability of reaction force to spallation. That's still unresolved, but since then, some more of the deflection community seem to have become aware of just how fragile some of these rubble piles are. I personally doubt that big push methods will ever work on an unconsolidated, non-ice bound, rubble pile. Disrupting one too close to home can turn a single billet hit into a shotgun blast, actually the worse outcome.
Underlying all of this worry is a complete lack of the subsurface "ground truth" of these things. What they are made of, how put together, any binding forces or agents, typical mechanical properties, that sort of stuff. As variable as they appear to be, I have grave doubts we could learn very much truly useful info by looking at just one.
That last is a really good argument for making the long voyage out there, and investigating as many as possible. What we learn for deflection also supports mining. That's a win-win. Plus we can look at all sizes, too. There's some reason to believe that size is the difference: why some have ice and others are dry. But, we just don't know, yet.
GW
"500 ton rock" -- my point was that the best data we have says these are NOT rocks, they are rubble piles. The "space rock" concept is quite wrong.
Many of the smaller ones seem to have no binding ice: they are dry, and therefore quite loose. And, yes, they do reach a rotation speed where they fly apart. Lots of evidence for that already exists.
So, how do you push on a thing like that, for the robot asteroid diversion mission, or for a save-the-Earth deflection mission? Or for mining? That's what we have to learn how to do. With the variation from one body to the next in individual properties apparently enormous, how can investigating one be representative of them all? These are VERY serious questions with no firm answers yet.
We can always argue over whether it's smarter to defer (or not defer) developing the capability of long-distance manned travel in favor of a robot asteroid capture capability, so that we still only need the capability of reaching near-moon space with men.
Deferring that capability puts off sending men to Mars, but does allow you to recreate the capability of sending men to the moon, something we have so clearly lost. If instead, you have the capability of long-distance manned travel, then both asteroids in-situ and Mars become reachable with the same hardware (and the moon is "easy"). But, that's a tougher technological and proven-hardware "reach".
It's a ways-and-means thing. How best to go about answering those two fundamental questions about asteroids: (1) How do you push on one of these? (2) How variable are their properties, really? The answers are SO fundamental to so many things we might want to do. This is very important, we have to do this right.
You decide what's smarter for yourself. I did, I favor going straight after long-distance travel as enabling the bigger payoff.
GW
I quite agree with Robert Dyck.
It's time to "crash" the commercial manned launch program. I think if that takes more money, it would be well worth it. Besides the astronauts-riding-on-Soyuz problem, there are only 3 commercial launch vehicles (Atlas-5, Delta-4, and Falcon-9), soon a 4th (Falcon-Heavy).
None of those are man-rated yet. But, of the four, one (Atlas-5) uses Russian engines in its first stage. So we are vulnerable there, too.
Saving the extra money by choosing now from among manned Dragon, CST-100, or Dreamchaser would be penny-wise, pound-foolish management. And there's been way too much of that already, thank you very much.
So, fund them all to completion, at best-possible-speed. That way, we get redundant means of launching astronauts, as soon as possible.
By the way, I suspect that CST-100, manned Dragon, and Dreamchaser could be launched by any of 3 launchers: Atlas-5, Delta-4, and Falcon-Heavy. All that is needed is a set of adapters for each spacecraft on each rocket, and some wind tunnel data. That would be a very good investment, and not a very large one.
It wouldn't hurt to invest in another engine to use in Atlas-5, either.
GW
I think the "robot getting the asteroid" idea was limited to very small ones, around a meter or so in dimension. The problem, as I see it, is the de-spin operation. You can literally bag a thing that small, and drag it home. But it cannot be spinning when you bag it, or else it'll punch a hole in your bag.
The problem with de-spin is that you have to push on the thing somewhere, and more-or-less tangentially at that. Most of the data we have at the small asteroids visited so far by probes suggests that these things are dry, "fluffy", loose piles of all-different-sized rocks and mineral dust grains, bound together only by mutual gravity. That's very nearly no binding-together at all. They will exhibit very-near-zero bulk compression and bulk shear strength.
So, how do you push on a thing like that? In any direction at all? I'm not even sure that a stake driven into a thing like that will hold any force at all, in any direction, either: the particles will just "flow" around your stake. I'm guessing that you'll have to shine a laser on it, but not so much as to push the particles apart: you must use a weaker light pressure than the pull of gravity, and I don't yet know how to quantify that. But, I'd bet that it'll take a long time to have any effect.
GW
Russia partitioned and annexed part of Georgia a few years ago. The same thing is happening to Ukraine right now. There's a few others the west doesn't care enough about to go to war over. But, there's the 3 small countries on the Baltic Sea, and there's Poland. All are part of NATO, which is sworn to defend them. All have direct borders with Russia, and to one extent or another were either part of, or vassal to, the old Soviet Union.
If Putin messes with a member of NATO, war will ensue. The question is, how strong is his dream to rebuild the old Soviet Union? How strong is his desire to confront and win over the west, particularly the US? How willing is he to sacrifice other people's lives to reach his dreams? The parallel with Hitler's Nazi expansion in the 1930's is unnerving. We all know how that ended up.
Once again, the personality characteristics of a single dictator could determine war versus peace. And I do not like what I see in Putin.
GW
In effect, using a closed cycle, so that the low pressure location is no longer the local atmosphere, but something closer to the atmosphere here. That makes existing compression machinery quite feasible, but at the expense of carrying around the low pressure reservoir, which is going to be huge compared to the high pressure reservoir.
I honestly don't know, but it's a trade-off certainly worth investigating.
GW
One of the tacit assumptions in all of these SLS and Orion projects is that only a capsule with service module and a propulsion stage is needed to fly beyond the moon. That's demonstrably not true. The key phrase is right in the pasted article: "The SLS program does not include the crew quarters. However, NASA is separately developing the Orion Crew and Service Module which will be integrated with the SLS at the launch site."
"does not include the crew quarters"!!!!! So, who's working on crew quarters for very long trips? In any significant way, I mean? Not NASA.
Flying beyond the moon requires months to years, not days. We have known since Gemini 7 in the 1960's that a crew in a cramped capsule is good only for a few weeks, maximum. Sufficient living space properly distributed among functions is required for long trips, period. And, you need to worry about solar flare radiation shielding, and for trips over about a year, microgravity disease. Otherwise, your crew comes home dead, nearly dead, or not at all.
I believe that the lack of effort toward long-duration flight is why "going to an asteroid" got subverted into "let the robot bring the asteroid to the moon, where we can go". Under such conditions, a private outfit has a better chance of sending men to Mars successfully in the next half-century than any of the government agencies. At least they have some motivation. The government quite clearly does not.
GW
Quaoar is exactly right about using dry gases for compressed gas systems. Ice from humidity is a very serious problem, even with ordinary "shop air".
The Indians already market a compressed gas car. Runs on compressed air. Tata Motors, I believe, and I don't know how they solved the moisture problem, except that they did or the car wouldn't work at all. It fits a very special niche market in the way they do things in-country over there.
As a car, it doesn't meet safety standards over here in the US or Canada, which lack is a part of what makes it feasible for their niche market over there (weight reduction for stuff left off). The range is very short, just a few miles. The compression pressure is 4 or 5 thousand psi, if memory serves. That's most definitely not the "shop air" (85-100 psi) you find at filling stations. Nor is it what you can buy retail in a compressor at the hardware store. It is comparable to the military compressors used in submarines. Those are not cheap, nor are they energy efficient.
Compressed gas propulsion will face exactly the same problems on Mars, compounded by the near-vacuum atmosphere pressure available to start your compression process. Here, with a 1 atm source, good-quality real-world compressors are around no more than 65% energy efficiency. With 0.6% of an atm source, that inefficiency gets far worse! It compounds exponentially as the number of stages increases, which it must.
The non-compression phase-change methods of supplying the initial compression to near 1 atm, will all be afflicted by very low throughput for the machine mass and power demand. Further, they are easier to implement as batch processes, not continuous processes. That's things like trapping dry ice inside a fixed small volume, and then subliming it, for self-compression of the vapor.
GW
Hi Josh:
I honestly don't know which engine might be better. The piston types handle varying load better than turbine, and by far. You can get around that issue, by using the turbine to drive a generator, and go electric drive. Add in some batteries to handle surges, and you can reduce the size of the engine to a minimum, and get better efficiency by a factor approaching 2. That's the secret of the series hybrid here. Not just cars, but locomotives and submarines, too.
There are reactants the turbine could use that fuel cells cannot. I don't know about actual cycle efficiencies, but I'd hazard the guess that a turbine electric hybrid would be competitive with a fuel cell electric. But, the turbine electric hybrid is a lot more versatile as to reactant identities. Here, turbines can be universal-fuel, if you just design it that way. It's not that hard to do.
Your biggest problem with turbines is the blading. It has to be restricted to survivable temperatures, and likely in the presence of some oxygen. Those aren't carbon steel. Typically, they're superalloys like Haynes, or Rene, or Hastelloy. The gas temperature approaching the blades cannot exceed about 2200 F in the very strongest military designs. Those alloys are difficult to make, even here, and quite expensive. A radial flow design might ease some of the structural demands, if you sacrifice a little efficiency.
For any sort of combustion engine, I'm worried about the gas feeds to it. Here, the biggest massflow coming in is the air stream, which needs compression from 1 to a few dozen atm. There, it will be your oxidant plus any diluent gases. A compressor for gas flows of similar magnitude on Mars, going from 0.6% of an atm to a few dozen atm, is going to be way too big and power-hungry to be practical. You'll have to generate these gas flows, likely from stored liquids, real-time and very rapidly in some sort of feed plenum chamber, at near 1 atm. It's a really tough problem, no matter how you look at it.
That's why I keep iterating back to fuel cell electric. It's the one thing we already know exactly how to do at Martian conditions.
GW
Josh asked something about higher-temperature alloys just above. I used one in a reducing environment that was quite strong. I think it would function in a neutral environment, too. That was TZM, a moly alloy doped with Ti and Zr. It oxidizes-away rapidly above 1300 F, but if you avoid oxygen exposure, you can take it a lot hotter with near-room temperature strength. I used it up to 2200 F like that, in thin tubes subject to lots of internal pressure. I think it melts close to 4000 F, but I'm no longer sure. That was 30 years ago. I think it's available in bar, tube, and plate. I doubt it's available in sheet. It's machinable with some difficulty. I don't believe it's formable at all (sort of like 6-4V Ti in that respect).
GW
As a matter of practical heat transfer, the kind of cooling "radiator" (a misnomer, it works primarily by convection, not radiation) and fan that works on vehicles here, will not work well at all on Mars. The air is 0.7% as dense there as it is here, and transfers 0.7% of the energy, all other things equal. Those things are flow speed, radiator area, temperature difference, and "air" heat capacity.
Temperature difference and heat capacity aren't free parameters under your control, they are artifacts of what you are doing. Flow speed and radiator area are inversely related, but their product will be a constant driven to enormous size by the low density, roughly 140 times larger than we are used to, designing radiator-fan systems here. Sorry, that's just heat transfer facts-of-life.
You'd have better luck on Mars with true radiative cooling, except that requires far higher radiator fluid temperatures (T^4 dependence of Boltzmann's law). If it weren't for being mobile, I'd say your best cooling bet is convection directly into the cold regolith as a heat sink. That takes a field of buried pipes. For mobile operation, at high power, you might get away with brief surges by using a sacrificial phase-change coolant. Dump your heat into ice and expel it as steam. Very wasteful, requires a lot of water. Tons not kg for a very few HP-hours.
The much-higher rejection "reservoir" (sink) temperature of a true thermal radiator operating in vacuum or near-vacuum conditions drastically lowers the Carnot efficiency of your associated heat engine cycle, which is a fundamental upper bound on what you can do with heat engines. Your real cycle efficiency is far lower than Carnot (by factor 2 or worse). Carnot efficiency is 1 - Tc/Th where Th is the hot source temperature and Tc is the cold sink temperature. Sorry, that's the facts-of-life from Classical Thermo 101.
The thermal radiator is the only thing that works closed-loop in space. We already know that. The thin air on Mars is a close cousin to the vacuum of space, that's why automotive-type radiator-fan combinations will be ineffective there.
That unfavorable outcome sort-of argues against using heat engines for mobile operation. Fuel cell electric is not a heat engine, and does not directly suffer the same thermo and heat transfer limits, operating in a near vacuum like that.
But for stationary operation, heat engines make a great deal of sense, using the regolith heat-sink approach. If the pipe field is large enough, convection through the regolith to outside the field makes the cold sink look "infinite" in size, allowing steady-state operation. That effect will likely size such sink fields on Mars; not something we usually run into here, so very much.
GW