New Mars Forums

There are a number of possible methods. Here's a good reference to start with.

http://www.niac.usra.edu/files/library/ … ngland.pdf

Personally I prefer to see advances in electrochemical processes such as..

http://research.jsc.nasa.gov/BiennialRe … F/EA-4.pdf

Edit: Something I didn't know before. Martian air is 0.13% molecular oxygen.

I'm afraid you're going to have to a) read it in more detail and b) if you're still struggling ask some specific questions.

Let me explain to you why low Mars orbit figures in this. The cost of getting something from Mars surface to low Mars orbit is best summed up by the mass ratio of 3.6. The cost of getting something from Mars surface to a high (near escape) Mars orbit is best summed up as a mass ratio of 5.3.

Now, whilst it is indeed quite feasible to build an ascent vehicle that can take you from the surface to a near escape orbit, that difference in mass ratios adds up to quite a few tonnes. Not just in the fuel, but in the vehicle itself - which will need to be landed in the first place.

When you do the math there is a very clear advantage to having the minimal mass ascent vehicle, and refuel it in low Mars orbit. To get a 5 tonne vehicle from low Mars orbit to high Mars orbit requires about 2.3 tonnes of fuel. The same vehicle launched direct from the Mars surface to high Mars orbit would require an additional 8.5 tonnes of fuel at Mars surface but then would also be a heavier vehicle. That extra mass then scales the fuel. So you end up with something like (roughly) a 7 tonne ascent vehicle needing 30 tonnes of fuel when originally you needed 13 tonnes of fuel on the Mars surface.

Fuel produced on the Mars surface does not come for free. It inevitably adds to the mass of your mission back at Earth.

Ok, that's one reason I'm considering having a habitable vehicle positioned at low Mars orbit. The other reasons are safety and flexibility.

What you have to remember here is that the bulk of what you're transporting in the process of getting a crew safely to Mars and back is the consumables and what I'm doing is keeping the bulk of that mass in high Mars orbit. Which has a very large fuel saving.

As for your second paragraph. I think you need to read more carefully. Manufacturing cost is not the killer. Its development cost. And I'm actually trying to avoid throwing things away. That's why my transit vehicles are reusable. The concession made to the fact that these things ultimately have a life span is that when they have been through two missions already they are then left in low Mars orbit. There is a variation on that theme and that is to design them with a life span that exceeds ten years and as the project progresses, simply reuse them more on actual crewed transfer missions.

As far as overall costs are concerned, the biggest cost is development. Which is why I'm proposing to limit the non-Mars-surface part of the mission to essentially two distinct vehicles and no more. Compare that to a lot of other architectures.

Now on the point about hydrogen. I'll go back into that in my next post.

And I wouldn't be too fussed about a supple suit. I mean yes it would be nice. But given an agile teleoperated robot I'd rather sit in a shirt sleeve environment. That would cover a surprising number of tasks that don't simply require something to be pushed, shoved or kicked.

I guess my first question is what is the anticipated mass of the ITV and which Mars orbit is it using? If I were to guess, what you're aiming for is a high Mars orbit, closer to escape than to low orbit in terms of energy?

And the basic idea is that the ITV cycles between low Earth orbit and a high Mars orbit?

What I'd probably disagree over is the use of a Mars ascent vehicle to carry the fuel needed for TEI. Obviously if your ITV is in a near escape Mars orbit and you're using aerocapture back at Earth, then the amount of fuel for TEI is kept to a minimum. So about 1.2 to 1.6Km/s depending on the original orbit and how long you wish to take. So the mass ratio on return to Earth is about 1.5.

Now, just throwing in some concrete numbers, a 20 tonne ITV would need 10 tonnes of fuel. Problem I kept running into at this point is that not only does the ascent vehicle need the extra delta-V to get to a high Mars orbit, it also needs to carry that extra fuel for TEI with it. From Mars surface to high Mars orbit is a delta-V of about 5.8Km/s and that's a mass ratio of 5.2. The notional 10 tonnes of TEI fuel becomes 50 tonnes of fuel back on Mars. And with that comes ascent vehicle bloat.

No matter how I came at that problem, I couldn't find a way to get fuel from Mars surface for the return journey that made sense. At one point I had a multiple-reuse lander/ascent vehicle that needed multiple trips to ferry enough fuel to orbit. And eventually I figured that would be asking too much of ISPP. Very large power source etc.

Its on the basis of that math that I settled on having a habitat in low Mars orbit, which doubles as a tanker. I could have just parked a tanker there, or even expendable boosters. But in the end I felt uncomfortable with the length of time people spend in suits.

Now, I'm happy with an ascent vehicle that has no pressurised compartment for crew. A faring with seats. But a store of air. Comfortable for a few hours, with a contingency for a couple of days in an emergency. The thing I didn't like when I came at this problem myself was that launching people direct from the surface to a high Mars orbit not only requires a lot more fuel on the surface and a larger vehicle, it also places constraints on navigation. Far easier to aim for a nearly circular low orbit, refuel, climb into a pressurized space and enjoy a beer, and wait for right launch window to take you to high orbit. Now, I'm pretty sure that I haven't considered all possible options here.

As for aerocapture, I'm a little easier about it. And its doable. It does require testing though. What I'd prefer to see happen is that the vehicle has the capability to propulsively capture at both ends. Which might lead to a program where initially the vehicle is propulsively captured, but as more confidence is gained then aerocapture becomes the norm.

Aerocapture less so aerobraking means at the very least a stowable solar array. I can't get around that. Keeps the crew busy I guess.

My architecture isn't entirely dissimilar in that I have a reusable transit vehicle. It goes from a high Mars orbit to a high Earth orbit. Now, what I still have to think about is what happens then. From a high Earth orbit to ISS orbit its a relatively simple task to aerobrake. Its just a case of whether the days to weeks involved in a gentle aerobrake is too long to keep the crew confined. This is one of the reasons why I swapped from having a Mars lander that remained near Mars and one that came back with the return vehicle. Its to allow the flexibility to get the crew from high Earth orbit back to the ISS in a hurry. Its by no means essential though. But having the couple of days of emergency air on the lander is important.

I agree that having a reusable transit vehicle is important politically in the sense that having it up there keeps the game going - rather like the ISS really.

I am divided on how to approach the word "reusable". In other words, how many missions do you want to fly the same vehicle. Initially I figured on a lifespan (with refurbishment) of about 13 years. That's 5 missions. Enough to amortize the cost of the vehicle. Problem is that the vehicle itself is not actually the big ticket item - its the development program surrounding it. When you add that to the fact that inevitably "upgrades" will happen and inevitably you end up with hybrids if not total replacements. The other thing I kept coming across is that the bulk of the mass of a transit vehicle is not vehicle - its fuel and consumables. All of which has to be launched from Earth. Hence I'm comfortable with a reusable transit vehicle, but not necessarily one that lasts a decade or more.

Now, I stand to be convinced about the merits of making return fuel on Mars. Fuel that gets you into Mars orbit yes. High orbit maybe. But given the decision to send fuel to Mars and leave it there in low orbit, I then had to decide on what vehicle. And ultimately that came down to a decision to use a transit vehicle as the tanker. And I'm surprised no ones yet criticised that because I'm not absolutely happy with it either. However, a transit vehicle is a good thing to have in low Mars orbit. For the reasons discussed above, and because it has the tankage needed. Now of course using it as a fuel store for the lander means the transit vehicle parked in low Mars orbit stays there. And if you send one on every mission you end up with a small fleet of them. A little wasteful but also handy because the more there are, the more places in low Mars orbit you can find a refuge with life support.

I'm also not entirely keen on landing ascent fuel using one of my landers. That's the single biggest use of fuel taken from low Mars orbit and without that I could probably make the transit vehicle parked in low Mars orbit last two missions. Nevertheless, taking what I just said as a starting point I'm left with a bit of a juggling act.

Instead of just having one transit vehicle that shuttles back and forth, I've actually got three. Although two of them effectively act as the one vehicle. The manned vehicle. The other lone transit vehicle is sent unmanned, propulsively captured into high Mars orbit, then aerobraked into low Mars orbit. Now, what I'm thinking of is rotating vehicles (like car tyres.). So the life of a transit vehicle goes like this. First, its launched from Earth containing all the spares and consumables and fuel needed for a manned mission. Now its partner might join it in LEO (not decided on this). Various masses are distributed and the vehicle checked out. The whole kit is boosted to a high Earth orbit and waits there for the crew to arrive. A separate flight brings the crew. Both vehicles transit to Mars (high orbit) and remain there for the duration of a mission. They return to high Earth orbit. The crew transfer to low orbit and then back to Earth. Now what happens here is that newest transit vehicle now waits in Earth orbit ready for a bran new vehicle to join it. It then does the Mars and Earth mission. After the second mission and on return back to Earth, it then becomes the vehicle that is sent ahead on an un manned flight to become the vehicle parked in low Mars orbit. So its already served two complete missions at that point.

So for every manned mission, one complete new transit vehicle is launched from earth. Which then becomes half of the actual vehicle sent to Mars.

Part of the reason I've done all of this is to limit the range and size of boosters needed. I figured on being able to use a booster capable of launching a bit over 50 tonnes to LEO - basically a Falcon heavy. For each manned mission there are 6 launches. Goes something like this.

First two launches are for the unmanned portion. What is delivered into LEO is a fueled transit vehicle (maybe a lander or if not some other vehicle designed to land fuel on Mars). And a smaller hydrolox booster. These are paired and sent on a lower energy trajectory that ultimately arrives in low Mars orbit ahead of the crew. One launch is in the order of 45 tonnes. The other is in the order of 30 tonnes.

Next four launches are..

A transit vehicle, fully fueled and carrying supplies. Topping out at around 50 tonnes.
A hydrolox booster, again around the 50 tonne mark.
A fuel tank providing fuel for the other transit vehicle. Plus the lander. Another 40 tonnes.
Finally, the crew. To the ISS in presumably something like a Dragon.

So I'm thinking in terms of a shade under 250 tonnes in low Earth orbit terms. That's per manned mission, ignoring the other one-way stuff that's actually landed on Mars.
Single biggest improvement on that would be using aerocapture. With aerocapture you could take 80 tonnes off that.

As far as what gets landed on Mars (and stays there) I'd personally be pleased to have at least one fixed, rigid structure. But I've no problems with rovers etc. If only Mars drive could get their power problems worked out.. One thing I'd like to see is a fixed pressurized structure that can nevertheless be disassembled into tow-able units. Certainly I think that the kinds of materials you're talking about are inevitably going to be part of this.

I'm starting to feel like a camping/caravan holiday here

Ok, probably left out a fair bit of detail here. Have to post and run. Let me know if I've missed anything you feel should have been talked about.

Josh,

I started this thread to discuss an overall architecture. And more than that to encourage wider discussion about comparative architectures. There are many more possible architectures than those currently being championed. And too often what you see are people responding to a particular problem, often at the expense of causing other problems. What I'm trying to do is to get a more holistic approach. I'm happy to discuss partly versus mostly propulsive landings, but that's a small part of this.

As far as the overall architecture goes, I think there is a lack of discussion about the detail. Simple things like how you get from low Mars orbit to high Mars orbit. What you do exactly on return to Earth space. What you contingencies are. And so on. Those considerations flow back into the design of your vehicles and the overall development effort.

Early in this thread I pointed out that my architecture as much as anything is about simplifying the development effort - by having only two vehicles involved in the "getting there and back" problem. I've not spoken much about the development of large, aerobraked vehicles that land stuff on Mars on a one-way trip because in essence, that's something common to every architecture. And I'm perfectly happy to talk about that. And I'm even willing to say that I agree that such big landers can be optimised to 50% landed mass, and possibly a bit beyond that.

As far as the nature of the manned lander goes. Is it partly or mostly propulsive. We've reached a point where there is little to be argued about regarding mass. A single use lander, accompanied by an ascent vehicle that itself sits on top of its own lander, will always be more mass than a mostly propulsive lander/ascent vehicle. So lets exclude that option. That leaves us with a lander and ascent vehicle that is partly propulsive - it uses a heat shield. Where do we get to with that? Well, a vehicle that is a few tonnes less mass than fully fueled mostly propulsive lander. Does those few tonnes count? In isolation yes. In context no.

One of the things drilled into me in engineering is the phrase "the devil is in the detail". You might start out with something that is conceptually simple but end up with a monster. A jet engine is conceptually very simple. But look at the decades of effort and thousands of parts that make a modern engine. And all that detail has its price. The engine that failed in a Qantas A380, failed because of a slightly misaligned drill hole - a tiny mistake in a huge file of numbers.

All else being equal - and this is pretty much the case here when comparing the partly and mostly propulsive landers - I'd settle anytime for the simpler, more serviceable and to my mind, safer machine.

I'd like someone to tell me the deal-making advantage of having to rely upon a heat shield - without simply deferring to "conceptual simplicity". A heat shield is a necessary evil for larger landers. But with a smaller lander you've got a choice. And I really don't want to go into here all the operational and safety advantages that come with a mostly propulsive concept. Been there, done that.

But I will repeat one issue, since I'm talking about the devil in the detail. Supersonic retro propulsion is a problem that will be solved. But, it may be solved with certain caveats. Certain no-go zones in terms of the placement of the retros, the density and velocity. I've done two things that I believe make this problem easier. One is to move the engines away from the body of the vehicle. So the turbulence has less to react against. That's simple physics. The other thing is I'm able to maintain constant ignition through the whole procedure. That may or may not matter. But what people may discover is that you probably can operate a retro into a higher velocity stream than you can start one up at. If you've ever watched videos of people blowing out oil well fires with a jet engine then you'll understand what I mean.

The inevitable result is that SRP parked behind a shield will almost certainly operate within a tighter envelope than those don't and remain lit for the whole procedure. Which means a vehicle that uses a heat shield and then goes directly to SRP is going to have a shorter, harder terminal brake, and with less margin.

NO way I would like to ride that myself.

I'll stand by the conclusion that with comparable margins, the mostly propulsive lander is only somewhat more massive than one that uses a heat shield. And the resulting ascent vehicle is lighter. Yeah ok, I'm repeating myself again.

Now you can call me uncharitable but I'm just not enraptured by the tendency to ignore the details. For instance, I've noted several times here that trash is an issue. But gotten no response. Ok, you start with a Mars hab. That's got to be landed with a heat shield and once that comes to rest you probably don't want to land a similar vehicle anywhere within a few tens of Km. Now what happens when you land a smaller, crewed lander? Well you have to target a landing ellipse for the heat shield that's well away from base. Then what happens with the next lander? Things get complicated, and people start needing longer distance rovers. Anyhow I'll stop at that for now. I'd just urge people to consider that the detailed design.. the boring engineering stuff.. really does matter.

And in the end, simplicity of actual detailed design (as opposed to simplicity of concept) does matter.

Now, please, can we talk about overall architectures? Are we going to park a return vehicle in low Mars orbit, or high Mars orbit? Are we going to return that vehicle to Earth space in a reusable condition, or trash it? Where are the designs for small "taxi" vehicles designed to transfer people between orbits? Should we integrate visiting Phobos in the process? Should aerocapture be a requirement before we go to Mars, or simply a later refinement? What level of redundancy do we need? Are we going to trust a singular life support system, or (as I have done) split it over two self contained units? What happens when we develop leaks? How about leaking fuel? should we build in the capability to transfer fuel?

That's what I'd like to see a discussion about. Thanks

Btw, although I'm proposing to use a lander (on its last mission) to ferry fuel to the surface, I'm still not convinced myself that's worth the cost in fuel. It may make more sense to make that a more conventional landing. But when I did try this approach before I kept running into issues that had to do with where you store surface fuel.. in a rover like vehicle.. in a stationary tank? Etc. Unresolved.

If I don't land a lander as a fuel ferry then obviously there's got to be spares supplied somehow else.

But as far as an ascent vehicle goes, the least mass option is to land it with the crew. A conventional lander, plus an ascent vehicle carried down by its own lander will always end up costing more in mass than a lander that doubles as an ascent vehicle. And I'll repeat, you can't beat the very large mass savings that come from putting your transit vehicle and return fuel into a high Mars orbit. Having a decent ascent vehicle that acts as an inter-orbit ferry is part and parcel of that.

As far as shock impingement goes, I do take that message to heart. That's why I'm thinking in terms of Mach 4 at most. At least in densities of any consequence. Now the configuration puts the engines above the center of mass and the engine bell shadows the guts of the engine. Its the boom that sits in the path of shock waves. And the answer to that, at least at these sorts of speeds is a selective thermal protection layer on a small portion of the boom. If you really wanted to design it to the limit you could use fuel as coolant for the boom - since you're expecting a small amount of fuel flow for the whole trip.

And I've already addressed this four or five times. The primary reason is its safer. And let me explain why, again.

An 8000Kg lander with a 6m diameter heat shield will slow to Mach 3 at around 6Km altitude with about 30 seconds before impact. Give or take.

By comparison MSL with an entry mass around 3000Kgs and a heat shield 4.5m in diameter needed to slow to Mach 2.2 before chute deployment which was nominally about 9 to 10Km altitude.

http://ntrs.nasa.gov/archive/nasa/casi. … 006430.pdf

To give a crewed lander the same sort of margin. That is to get the craft down to Mach 2.2 at 10Km means inflating the heat shield diameter to more like 9m.

Yes, quite doable, and yes still possibly less mass than a fully propulsive lander.

But, now the single riskiest item is the parachute. Its massive, or you've got more than one. All of which leads to more failure modes. I can live with that if there's no one on board. You want to know the statistics of how many people are killed by parachutes here on Earth?

GW is quite right in that beyond a certain point (in terms of increasing lander mass) you're better off ditching parachutes and going directly to retro propulsion as soon as possible. I don't agree with him on the exact numbers for two reasons. One is that when you take gravity into account your "projectile" doesn't keep the same angle to the horizontal. As it slows down below about 30Km, gravity kicks in. The other is that real margins reflect the variability of the Mars atmosphere.

This is why in my mind, landing large masses on Mars does mean having a reliable aerobraking system. But parachutes are not part of the solution. Higher velocity/temperature decelerators deployed further up might be useful. At the very least no one can avoid SRP.

With a mostly propulsive landing you've got complete authority over what speed your'e doing and when. As I said before, you can land on Olympus if you want to. Certainly a few Km above MOLA and it makes no difference - you've simply set your propulsion to the settings needed. But with aerobraking those last few Km count. Or put another way, if you land propulsively, higher landing sites simply mean less gravity loss.

You see, when all sources of error are accounted for - the main one being atmospheric variability - a landing that depends on aerobraking gives you only a couple of Km and a matter of seconds to play with. A mostly propulsive landing ensures you are where you want to be with more precision. Less variability. More margin.

Mars has big diurnal variations in its atmosphere and combining that with direct entry means over sized heat shields (just as on MSL). Think about that one. Diurnal variations trouble a mostly propulsive lander less, but even so, this is part of the reason why I prefer using low Mars orbit as a starting point. It gives you more choice as to when you land.

As well as parachutes there are other things that can go horribly wrong in a conventional landing. Separation of the back shell, and more importantly separation of the fore shell (main heat shield). Now you can throw time at this, but that's less time you have to throttle up your terminal descent.

Thrusting through a heat shield has twin problems. First you're carrying that much more mass to land. Meaning more landing fuel. The other is that if you're going to run into show stopping turbulence its going to happen because you're riding on the plume from the engine. This is why I've opted for the concept of putting the landing engines outboard and above the center of mass of the vehicle. That on the face of it gives a lot more stability.

Now, with a very large vehicle you may find you have no choice but to fire through the heat shield, or else let the heat shield disassemble under you. But it still makes sense to keep the landing engines on the periphery. Now in that scenario you're still experiencing some of the backwash from the landing thrust. Net result is that you're sitting inside a tighter envelope of speed and thrust setting. In essence, putting the engines outboard and above the center of mass means you can use higher throttle, sooner. But you can only do this if you're not exposing the engines and their supports to the extremes of hypersonic heating.

Any way you look at that it means safer. Any way you look at that it means a larger vehicle has less margin. And a vehicle that keeps the engine behind the heat shield has less margin.

There are other reasons why a mostly propulsive lander is safer. They have to do with simplicity and fewer systems that have to be kept in order, and monitored by a crew that may only have seconds to correct errors in automated systems. With a mostly propulsive lander its a lot easier to "ride the stick" if you have to.

And another issue. And I mentioned this before too. With a conventional lander your main source of error as far as exactly where you land is the atmosphere. And because you're likely dropping heavy items onto the surface you're forced to land some distance from base. With a mostly propulsive lander you don't have the raining trash issue so all things consider your landing is likely to be a lot closer to base. Things can go wrong between the lander and the safety of base and if your trip is 1Km rather than 5Km that's also a safety issue.

And if I really wanted to get into boring detail I'd get into how all the accumulated junk of heat shields and parachutes on the surface of Mars will eventually require a clean up. Anything you have to do out there comes with risk.

Moving away from the actual landing itself, lets take a look at what else I'm doing here. Now, my lander is somewhat more massive in terms of mass at low Mars orbit. Its 15 tonnes and the alternative might be 8-10 tonnes. That's a few tonnes saving relative to the many tens of tonnes you're otherwise transporting to Mars. I'll live with that.

But, where my architecture gets more decadent is in using a lander (one that has been used in a previous mission) itself as a transport for fuel to the surface. Now, admittedly I could have just dispensed with that and landed the fuel in a conventional pod. But the reason I'm doing this is that if you want spares, you've got spares. You send one ascent vehicle to Mars and you're ready to return home and the ascent vehicle loses an engine soon after take off and you'd better get used to survival rations. If you take off on my ascent vehicle and it does the same thing you've got yourself a week or two in which to swap engines. Safer? Yes.

Now, look, you could provide some spares separately and at less cost. And I remain open about that particular feature. It is nice however to have a vehicle that's already landed on Mars that can in a pinch be converted to a fully functional ascent vehicle. Take away this element of the architecture and a reusable lander and ascent vehicle actually requires less mass.

How about multiple redundancies in space brought on by the fact that you've got propulsion and fuel storage in the lander as well as in the transit vehicle? Safer? Yes.

How about approach to Earth. You've taken damage and lost some fuel. Transfer the fuel to the lander. As a last resort abandon the transit vehicles and use the lander to achieve Earth orbit. There's a host of scenarios like this where if you've got your lander with you, you're safe. If you've relied upon an expendable ascent vehicle, or worse, launched directly from Mars, then there's always the satisfaction of knowing your place in history, right? Yep, the first person in history to go all the way to Mars, and, um.. almost returned home.

I'm struggling to follow you here. The first question you ask is "what do you do if your lander crashes?". Which lander in which scenario is my first question. So lets go through some possibilities here.

I think most Mars architectures call for some things to be landed robotically. Correct me if I'm wrong there. Now, for instance if you were to land an ISPP plant and then wait two years before landing humans, what do you do if the robotic lander crashes, or the equipment fails? Give up? Or do you fix the error, launch again, and tell the crew they're going to have to wait? As I see it, this particular outcome is a definite possibility in any architecture that seeks to establish mission critical infrastructure ahead of a manned landing. Can't be avoided.

Now, are you considering the type of mission where mission critical infrastructure is launched during the same launch window as the crew? If so you've got two basic options. One is you land all your eggs in one basket. That's the land everything on one very big lander approach. As we discussed above, this has some serious safety issues. Now an alternative which you've raised above is to send the mission critical stuff slightly ahead of the crew. Meaning separate landings.

Thing about this approach is now you've got two separate landings. And you're headed down the track where the manned lander (the one delayed by 2 weeks) is fairly minimal. And to the extent you're not weighing that lander down with stuff that could have been landed 2 weeks sooner, and thus you've got a lighter lander and one that you can afford to build more margin into, then I'm happy with it.

Problem though is that you've reached a point where if something goes wrong with the first lander then the crew are either committed to hanging around in Mars orbit for a long time, or a free return trajectory. And what you've done here is trade the possibility that the crew will have to wait another 26 months on Earth with the possibility that the crew will have to go through the return journey and then again wait until someone figures out what went wrong and there is an opportunity to fly again.

Now, I'm committed to safety wherever technically possible. Which means a principle of anything mission critical functioning, or people waiting. You've pulled me up on the consequences of that, but the reality is that mission scrubs and long delays are a possibility in absolutely every architecture.

I could have chosen to trade off safety in various ways. For instance, a free return trajectory is well within the capability of my system. So if you were really in a blind hurry you could indeed send everything on the one launch opportunity. Ok, so the non manned lander crashes. What then? I've sent two fully fueled transit vehicles towards Mars. They don't capture. Instead they return home. Nice thing here is that instead of having to be very fussy about the trajectory in order to guarantee a free return, I'm in a position to plan non free return trajectories that can be converted to a free return. Because I've got fuel reserves both in the lander - that was tasked to ferry the crew to low orbit, and in the two vehicles. At last resort I can limp home in on transit vehicle.

The other nice thing about not having to be too fussy about having a free return trajectory in the first case is to minimise the length of time the crew spend in space. An issue you've noted.

Now, honestly, faced with a choice between that scenario, and simply opting to send the non manned mission out 26 months earlier, I'd choose the later even before scratching my head about it.

You simply don't get a choice. There will always be situations where you have to scrub missions. And its better to do that with your crew waiting on Earth than all the other alternatives.

Quite apart from the ethical considerations of sending people into space in its own right, do you stop to think about the risks to reputation and public enthusiasm that are not just oh dear we'll have to wait but oh no we just cratered a 60 tonne lander and there were people inside? People will forgive mistakes if its just stuff. They won't forgive mistakes if needless risks were taken with people and you can stop and think about the risks taken with the shuttle program.

By far and away the safest Mars architecture is one where do everything to protect the crew, including avoiding hurling people through the thin Martian atmosphere strapped to a very large brick. And if your concern is that Mars should be an ongoing venture, the biggest risk to that is not delays - its the commission of inquiry that will follow a loss of crew and the revelation that you could have done it more safely.

If you don't get why an aerobraking landing is necessarily more physically stressful you don't get the fact that the Mars atmospheric density profile is essentially an exponential curve. The only way you avoid higher g forces is to slow down earlier. That's what a mostly propulsive landing does. To get the similar effect with drag you need bigger.. MUCH bigger. So that you're already getting significant drag higher up - around 80Km. Worse part is that anything you use to create drag either has to have a variable size, or you again suffer the fact that if you've created enough drag high up, you've created way too much drag further down.

Inflatable heat shields designed to create significant drag in the higher atmosphere will be huge - 50m for a small manned lander and well over 100m for the sorts of landers you're contemplating. And what do you do when the stress becomes too high? You have to shed it and rely upon a smaller drag device. This staging effect adds mass, complexity and risk.

How about a rigid heat shield? Well for a smaller lander (under 10 tonnes) you might consider 6m diameter providing you want to hit 4g lower down. But if you want to make the same heat shield big enough to create enough drag higher up to lower the g forces, then you're going for an in Earth orbit assembled 30m affair. Obviously the mass goes over the top. Try and do the same thing with an inflatable hybrid and you need to lose the inflated section before it shreds itself. More risk.

Essentially the laws of physics are against you. There is no known way to build something that provides enough drag higher up in the Mars atmosphere that would give you the benefit of lower g forces. And do so without defeating the purpose which is trying to avoid weight.

Now, there is lift. You'll notice that MSL used lift. But it did so at the expense of throw away weights.

You could really light weight wings. But you'd either need those to fold or more likely to jettison them. Again, more mass.

Even though a minimal lander might cost you 5 tonnes in low Mars orbit, there is the extra structure and the reason for that is that it's reacting against the shell of the heat shield. That shell has to be stiff in order to avoid buckling and cause aerodynamic instability. Then you have to have a back shell under aerobraking conditions. And every time you decide to make the heat shield bigger, you also scale the back shell. You're lucky to get out of it for 7 tonnes. Now, add your parachute or other drag devices if you deem them necessary (they may just make things worse) and then the terminal descent fuel which at 500m/s using lower Isp fuel gives you a further mass multiplier of 1.22. So a 7 tonne craft is now 8.5 tonnes. And that's not the end of the story.

See, I'm creeping up on 2 to 1 here? Now, we may do better than that. But not significantly better. Not when the scheme of comparison include the system wide consequences. A mostly propulsive landing will always require more mass in LMO. How much more doesn't particularly matter. What matters, and I've said this before, is that when you don't think about landing in isolation, and think about the whole architecture, it makes a whole lot of sense.

Now, some failure modes, please?

I think the point you've not fully understood is that I've always operated on the basis that anything critical to safety will be landed on Mars and verified as operational before the crew leave Earth. So the safety issue you imagine isn't there. It a very simple principle. An unmanned mission lands what you need on Mars. Crew follows. There's no such thing as crew landing with life critical equipment.

I'd put it to you that the shoe is on the other foot here. You land crew with cargo and if the cargo doesn't function, you've got a safety issue. Even if you have landed successfully. Safety alone dictates you take the alternate route and land anything critical to safety before the crew leave Earth.

Once you've understood this, its possible to decouple the quite separate issues of landing large masses, and landing crew. In doing so you can provide more safety for the crew than you can possibly do were you to land the crew on a large vehicle.

I stand by a mass multiplier of 2 to 1. That's a 50% landed mass to entry mass. No landing on Mars so far has come close to this. NASA thinks that to get 30+ tonnes to Mars you need 100 tonnes at entry. Zubrin is a lot more optimistic. I'm adopting the middle ground. I think that a mass multiplier of 2 to 1 is possible with large payloads.

However with smaller vehicles you'd be quite lucky to get to that 2 to 1 (landed versus entry mass) figure. I was being fair. Zubrin btw is primarily interested in large vehicles.

I tend to agree that an evolved heat shield may amount to 15-20% of the overall entry mass. However one factor I keep repeating that doesn't seem to be fully understood is that apart from the heat shield the entire vehicle needs to be heavier because its being exposed to much higher forces during aerobraking than would be the case for a mostly propulsive landing.

So for instance, if you wanted to land an expendable ascent vehicle using aerobraking (inside an expendable lander), the payload itself - that is the ascent vehicle - must be heavier than if it were landed mostly propulsively. Hence one of the advantages of the lander and ascent vehicle being one and the same vehicle is that you can make the ascent vehicle lighter than could possibly be the case if the ascent vehicle were landed conventionally. Lets be more concrete about this. A mostly propulsive landing can be engineered with a peak deceleration of 18m/s/s. (About 2 Earth g). A landing involving aerobraking involves upwards of 40m/s/s. The reason for this in simple terms is that you've burned off most of the velocity in the initial burn and the final burn is of comparable scale to a terminal descent from aerobraking - so around the 500m/s mark, give or take. From there to zero vertical velocity in 40 seconds without needing to exceed the 18m/s/s mark. Now with aerobraking you're a prisoner to the density profile of the atmosphere and you're having to burn through all that velocity much closer to the ground.

Now, adding more drag. A larger heat shield. Parachutes. Ballutes. Anything you can imagine. They can only slow you down faster in the denser layers of atmosphere. Which means the more margin you want to build in, the higher the stress on the vehicle structure. And that affects everything. Framing, tankage, even down to the little bolts that secure the plumbing.

There are of course ways out of this. Generating more drag higher up will give you lower stresses. But this means something huge, and light, and thus fragile. And probably something you will have to jettison in favor of a smaller drag device (going from ballute to parachute) at a relatively high altitude and as well as creating more things that can go wrong, will be pushing the limits of the second tier drag device. There is also lift that buys you more time but also buys you into more complexity and more pushing of the limits of materials. Even a trailing ballute device engineered to provide lift still has the problem that as the air becomes more dense it will generate too much drag and either self destruct or put too much load onto the vehicle.

Worse, all of this gets harder and harder to do, the larger your vehicle is. Buying margin for your crew is a lot easier and less expensive on a smaller vehicle. Taking a vehicle that masses 60 tonnes at entry and trying to give it the same margin as is possible with a smaller vehicle is, well, frankly bordering on show stopping. And there will always be the temptation to cut corners. You either end up with outrageous situations where you've reached final maneuvering (most of your vertical velocity is gone) with a Km to go, or you end up discovering that all your margin is gone and the only way to make up for it is to land people on lower terrain.

The initial burn of a mostly propulsive landing takes away most of the velocity and most of the difficult and thus safety critical materials issues. I'll say that again. Aerobraking is a difficult and dangerous process because you're stressing materials to their limits. And no I'm not particularly worried about ablatives there. I'm talking about the materials you're going to use for ballutes and the like. I don't mind those things if they're part of landing a non manned cargo. (After all, your crew isn't going to leave Earth until critical cargo has been landed safely and is functional, right?). That way the risk of very large drag devices isn't an issue of crew safety. And lets not forget also that landing large masses on Mars absolutely begs for these kinds of devices.

Another point about safety. When you land propulsively you're given back the ability to set the parameters. You've restored authority over the situation. You can land anywhere on Mars. You can choose the margin. It does cost fuel sure. But you're in a position where any other drag device you deploy merely adds to the overall level of comfort but you can still land safely even if it fails. Crew survival doesn't then absolutely depend on the correct functioning of such devices. With a small vehicle and an aerobraking landing that depends on drag devices you're taking a risk on those devices. With a much larger craft that has to land cargo and at the same time maintain margins for crew you're further pushing the materials. You're further increasing risk.

Final point on this issue. Much of uncertainty in an aerobraking landing is due to uncertainties in the atmosphere itself. You can to a certain extent compensate for that. That's what MSL was programmed to do by using a small amount of thrust to essentially fly its way out of these uncertainties. Nevertheless this problem exists because you're relying on air to slow you down. A mostly propulsive landing is affected less by the uncertainty in the atmospheric profile and also has more ability to correct for it. If you care about crew safety its little details like that that matter. And landing crew with a large vehicle makes this problem also much more difficult. And more difficult still if the large vehicle is relying even more heavily on cutting edge drag devices, and has more inertia and is thus harder to "fly".

Lets go to the issue of engine reliability. Reliable engines are not black magic. It simply requires competent engineering. The problem with most engines are they are deliberately designed down in mass knowing they only have to work once. There are wel l known examples of engines that can survive multiple restarts and long burns. The price to be paid of course is extra mass. But that's extra mass on top of a component that overall is not a large fraction of your overall mass budget. Put it this way, given the choice between an engine with a 100:1 thrust to weight ratio (relative to Earth gravity) that is designed for single use and an alternative engine with a 70:1 thrust to weight ratio (its heavier) that is ultra reliable. Which one are they going to specify, even if the mission only (nominally) requires one firing?

I'll also repeat something I said before. There are many Mars architectures that rely upon methane/LOX engines that are capable of multiple firings and long burns. We cannot avoid either designing such an engine or testing it.

Putting an Earth return vehicle on low Mars orbit, or even resorting to landing it on Mars does not obviate the need for reliable engines. Instead a sensible mission involves the capability of multiple firings if not for course corrections then for unforeseen maneuvers or contingencies. Engines deliberately designed down (and I do mean deliberately.. its a choice) to the point of being single use, should never be sent to Mars.

Having a minimal mass ascent vehicle means less propellant manufactured on Mars. Less propellant (even if its just oxygen) means less mas both for the ISPP plant and for the energy source. Some architectures seem to take ISPP too much for granted. I'm merely making the point that if your sole concern is how much mass you need to get into low Earth orbit, sooner or later you need to deal with the mass of the ISPP system and thus the mass of the ascent vehicle itself. And the lightest possible ascent vehicle is the one you've landed via mostly propulsive means, since its a gentler landing.

Even in a single stage to orbit Mars ascent vehicle you still have to incorporate features that are specific to landing. Otherwise you've no abort to land. And designing a lander and ascent vehicle together guarantees that the ascent vehicle is guaranteed to have abort to land capability. How many Mars architectures have been thought through this far?

I'm not asking my engines to be reused over and over again for multiple missions. They're being used once for a crewed landing. Once for a crewed ascent. During those two critical times, there is redundancy. At all other times there is even more redundancy. Although I'm using the lander as the primary propulsion source for trans-Earth that's in the context where both the transit vehicles have their own secondary propulsion using 2 of the same engines. Beyond return to Earth the lander is recycled as a vehicle that will and fuel on Mars. That's a non manned mission. And a non safety critical process for exactly the same reason discussed above. And beyond this the lander becomes spare parts.

What I'd ask you to do is to specify exactly the failure mode and scenario and then I'll address a response to that.

And for the sake of completeness I'll point out again that one thing I like about a mostly propulsive lander and ascent vehicle is that its inherently simplicity. Not just less things to go wrong. Its also about servicability. Its all out in the open. Its easy to visually inspect. Its easy to repair. And the lander that ferried fuel to the surface is a ready made source of spares, including whole engines, which at this scale can be man handled. The crewed landing is effectively being done on pristine engines. You land, you inspect, you swap components if necessary. You've got an ascent vehicle that has a firing sequence that if there are problems in the first seconds you do a soft landing and go back and fix the problem. And having transit vehicles in both low and high Mars orbits gives you a very long effective launch window so these sorts of problems can be fixed.

Again, specific failure modes please?

As for your last question. Its aerobraking that has obvious and hard to fix safety issues. Its time to think laterally.

As for L2. Using L2 does not depend on solar tugs. But having a base at L2 might benefit from solar tugs - provided that we're doing this on a very large scale and over many missions. Otherwise the development cost of a solar tug doesn't save enough to be worth it.

My architecture does not depend on the means by which you get to high Earth orbit, or for that matter which high Earth orbit it is. It could be L2, or it could be simply a high, elliptical orbit. The latter being slightly less energy.

I have noted that the only real advantage (to me) in using L2 is that what you have at L2 isn't traversing the radiation belts on each orbit. Arguably a problem for systems. Definitely a problem for people. But again, the latter has workarounds. And the former probably means running systems in safe mode at times.

The big advantage for others in having a base at L2 is proximity to the moon. I'm not fussed. Its also having a steady platform at which you can do manned assembly. Again I'm not fussed because what I've proposed is bump together assembly.

As for large unmanned cargo deliveries I have no problem with simply very large heat shields. As much fun as inflatable heat shield extensions are, there is doubt in my mind that they cannot be matched by light weight but still rigid heat shields of the order of 30m diameter. Which of course requires manual assembly. And that to me means using low Earth orbit. Doing so (with a rigid heat shield) means taking the most advantage of emerging light weight ceramics (GW would be happy to hear me say this). Anyhow I'm not wedded to any particular solution and your mileage may vary.

Ok well in response to that one what I'm thinking here is that for my purposes most of the surfaces that need some form of heat protection are actually cryogenic tanks. The surface is actually kept cold by the contents so the purpose of the thermal protection is really to reduce the rate of boil off. Hence small shingles even with gaps is quite sufficient.

I've not forgotten that there is more to be landed than just crew. Indeed I've acknowledged that. But I've also treated it as out of scope. Why have I done that?

Because everything else that must be landed on Mars is on a one way trip. And conventional, heat shielded entry vehicles are almost certainly the way to land those masses. And I've deliberately left that part of the enterprise out of scope because essentially it costs what it costs. In other words, landing the crew in a mostly propulsive lander does not change the cost of landing other materials. It simply changes the cost of landing the crew. Hence I can logically separate the two.

It would be unfair to describe the mass multipliers for a mostly propulsive landing as horrific. In simple terms, the mass multiplier is 3, versus 2 for a conventional landing. The question is not the mass multiplier, its the mass being multiplied. As I've made clear, in rough terms, to get a crew down to the surface of Mars and back again, it costs around 35 tonnes per crew in Mars orbit. To do that conventionally the cost is around two thirds that. That 10 to 15 tonnes extra, delivered to low Mars orbit is a real cost, and I acknowledge that. But mass isn't the only thing that matters. Other metrics such as flexibility and safety matter more. Most missions that rely upon conventional landers have other points of failure. Its those points of failure I'm trying to address. So this isn't just about the landing method per se, its about having the right vehicles. And its about those vehicles performing multiple roles, and thus allowing for simpler, but more focused development.

As far as loss of hab goes. Yes, it doesn't do any good to your mission timeline when things get cratered. But its not loss of crew if all that mass is landed, and verified before a crew leave Earth. Most people agree on this point anyway so I didn't bother to mention it. I'll repeat the point though that in landing crew along with a large object, such as a hab, you actually increase the risk to the crew or the cost, because in the one case you're dealing with higher velocities and lower margins (and quite probably lower landing sites as a consequence) and in the other case you're giving a huge entry vehicle more margin because there is crew on board, but obviously that costs.

Like it or not, the sanest method of getting people to Mars is on purpose built vehicles. Ones with low mass and where you can then afford to suffer other mass multipliers because ultimately the last few Km of descent does matter, a lot.

In factoring a 50% loss of hydrogen, I'm not saying you're losing half your hydrogen, I'm saying you're losing a third for what you started with. Just to make that clear. Secondly its not just boil off losses. Hydrogen has a habit of permeating most materials. Even in a liquid state. And because you've got complex shapes inside a ISPP including manifolds, heat exchangers and so on its nearly impossible to avoid some loss. The only way to really reduce loss with hydrogen is to quickly form it into water, and that requires previously generating a commensurate volume of oxygen. Now, that's feasible. But, then there is always a cost.

Now I won't quibble with you about the precise percentage of hydrogen loss. In the scenario above you need theoretically 722Kg of hydrogen. I accounted for 50% over and above that for both boil off and other losses. That made it 1083Kg. If the loss accounted for is only 20% then you need to import 866.4Kg. The net saving is 216.6Kg. Its called diminishing returns. As I said, given the risks you'd normally adopt a more conservative margin for hydrogen than you would for methane. Nevertheless I can adopt a margin of 20% and it changes nothing to the argument. The cost of converting hydrogen to methane ultimately boils down to more mass having to be landed. At the risk of repeating myself, the ISPP plant needed to be built will be more massive (over one that purely produces oxygen) and the power source mass will also scale. Now, I'd like to think that such an ISPP plant would service 2 or 3 missions. But its a risk to assume it would last longer. And the main reason I think hydrogen isn't as good a mass-save as you'd assume is that the vehicle that lands the hydrogen has to be larger and heavier. And that's a mass multiplier that gets worse for a conventional lander because a conventional lander has higher forces and thus that extra tankage and support structure for the hydrogen scales.

And again, the wider picture is that even with some optimistic assumptions, landing 866Kg of Hydrogen versus 3000Kg of methane translates to a mass saving in the order of 10 tonnes back in low Earth orbit. Now, it must be said that any mass saving is, on the face of it, a good one. But compared to the 200+ tonnes of mass launched to low Earth orbit that's a difficult judgement call as to whether you want to spend the extra $20M or so on launch costs, or you want to face the development costs, and risks of going directly to importing hydrogen on your first mission.

Again, I'll state that hydrogen does make good sense if its obtained from the planet itself. That's a goal I'm fairly confident will happen. But as I said also, I think that the process of scraping Martian dirt and processing it, at scale, is probably going to require some human supervision. Making 100% propellant production a goal for the 3rd mission. Now, as and when that happens, my architecture evolves accordingly. We then don't need to land any fuel, and as a consequence we only need one lander landing. And as confidence in the technology builds it might be possible to reuse the lander for more missions. Again, that's another reason why I feel its a good idea to take the lander back into Earth space with you because it gives you an opportunity to check out and refurbish.

Let me make it clear. I don't think the benefit of importing hydrogen (over the 4 times heavier methane) vanishes entirely because of the problems I've noted (particularly as a consequence of scaling the lander that lands the hydrogen). But practicalities do reduce the benefit obtained. And in the end, you have to maintain a sense of perspective and make judgement calls where a few tens of millions in launch costs are not worth certain added complexities and risks.

By the way, I'm actually in favor of testing hydrogen to methane synthesis even at an earlier robotic stage. I hope that clears things up.

As far as safe Martian aerocapture goes. That would be wonderful. The hardest decision for me is choosing not to use aerocapture on either end. Were I to be convinced of the long term stability of light weight ceramic materials as GW is talking about, I'd actually be tempted to do aerocapture into a high Earth orbit first. And the reason is that on return to Earth you can put the crew into the lander and it has the capability to both capture and put the crew into low Earth orbit, and presumably dock with the space station. If the aerocapture for the main vehicle doesn't work, well, you're going to have to build a new one. But if aerocapture does work then you can extend the same concept to Mars. This time you give the lander more fuel and its capable of doing both propulsive capture and docking in low Mars orbit (see the beauty of having some form of habitat there?). Those two procedures knock about 40% off the original mass. So its not to be sneezed at. But when you look at the numbers, the whole mission is perfectly doable even without aerocapture. Its just going to be cheaper with.

However you ran aerocapture and aeroentry together and that's a bit unfortunate. Aerocapture has the potential to save nearly a hundred tonnes IMLEO. Aeroentry - at least for the crew - is a second order saving.

I've done a lot of thinking about every single technique that can be used to trim mass. Before coming back to realise that mass simply isn't everything.

Now the reason why there has never been a truly reusable rocket on Earth is that Earth has a much larger gravity well. Its simply a lot more difficult to do. 9Km/s versus 4Km/s. That's the long and short of it. The reality is that a mostly propulsive lander for Mars requires a bit less fuel to land than it does to ascend. You can't do that on Earth of course. The mass ratios would kill you. But on Mars the mass ratio is about 3.6 going up and about 3 going down. Well within the realms of well understood engineering.

I'm not sure why we're talking about Ad-Astra. I think their engine is way cool. I also think it would require a legendary power source to make practical as trans-Mars and trans-Earth vehicle. More to the point though when it comes to dealing with Earth's gravity well you don't need a fancy new engine. A NEXT ion drive (or a bank of them) would do the job satisfactorily. I'm not advocating doing that. Indeed I'm happy with the 200+ tonnes IMLEO that would be required to do this with conventional boosters and a fair chunk of that obviously would be ditched before the whole project left the Earth domain. I'd be even happier with that of course, if aerocapture could be relied upon. I've just not specified aerocapture as essential because when its all said and done, if we spend a few hundred million launching rocket fuel (per mission) then I can live with that until its clear how to safely improve on that.

There is a case for L2 and its with solar tugs between there and LEO. But the development cost for a MW class solar tug I believe is such that it might be ruled out as a possibility for the first mission. I could be wrong. These guys who want to go to the moon again, and establish L2 might win their argument. And in which case you'd piggy back off that infrastructure.

Let me also bring in something that is more speculative but reinforces my feeling that the design has some truth to it. When you build a space hab or transit vehicle if you wish, there's a minimum mass, partly because of consumables and water. And also because mass is a desirable thing when you're avoiding radiation. Now, in my architecture you're travelling home with a reserve of fuel that's intended for Earth orbital capture. So up until relatively late in the trip, the fuel you've kept in reserve creates extra shielding, at least for a portion of your combined vehicle. If you do without that you've got to deal with the trade off somewhere. That's one thing. The other is that because you're more than likely to carry several tonnes of water you've got yourself a good heat sink. It turns out that in aerocapture you can simply point the blunt end of your vehicle into the air stream and youu've got yourself a big kettle. Now you probably need a modest amount of thermal insulation (light weigh ceramic) and the rest is handled by the water. You might end up with several tonnes of water heated to 60C (anyone for a hot tub?) but essentially its a fairly non stressful way to do things. Again, provided you get the thermal protection right. Can such a thermal protection layer also compensate for and act as part of the radiation shield. I really don't know.

Yes, you're correct. These were expected objections

The mass penalty for a mostly propulsive landing is what it is. Two thirds of your mass on entry is going to be fuel. So question is the difference in mass between a mostly propulsive landing and a conventional landing. And the more important question is whether that can be justified in terms of:

a) overall dollar budget
b) development cost and complexity
c) simplicity of design
d) robustness, resilience and safety

Considering only the mass involved in transporting the crew from low Mars orbit to Mars surface and back again, my proposal looks like this.

The crewed lander is posited as being 5 tonnes of mass, without fuel, but including crew and life support consumables.
The lander which delivers methane to the Martian surface is posited as having a dry mass of 3 tonnes.
The landed methane. That's 3 tonnes.
The fuel needed to land the methane. 12 tonnes.
The fuel needed to land the crew. 12 tonnes.

So accounted for in this way, the actual process of landing a crew on the surface of Mars, and bringing them safely back into a low orbit, involves 35 tonnes delivered to low Mars orbit. I hasten to add that with the architecture I propose, all the mass (fuel, hab and lander) delivered to low Mars orbit has first been delivered to high Mars orbit and then aerobraked to low Mars orbit. (And strictly accounted for, the mass of the crew does not belong in the above accounting, but I'll leave it this way to be conservative.)

Now before going to look at alternative approaches to the problem, I'm going to make the observation that the 35 tonnes delivered to low Mars orbit in order to perform this function, compares to the 70 tonnes of mass delivered to high Mars orbit, which is the transit vehicles and their fuel supply. In other words, if a third of your mass budget is spent on safely landing the crew and returning them as far as low Mars orbit, then I'm comfortable with the result.

I should also add that there is additional mass in the habitat vehicle that is transferred to low Mars orbit and is used in part as a fuel depot. Its difficult to know where to account for that because in this context, much of the mass of that vehicle is actually life support and serves the purpose of providing a home in orbit for the crew. But I do acknowledge that these numbers are neither perfect nor final.

A conventional single use lander suffers its own mass multipliers. It needs to withstand the forces of a conventional landing. Meaning several tonnes of basic structure. Add to that the heat shield. Add to that unavoidable propulsion system which still needs to provide more or less the same thrust even if for less duration. And now its easy to see why Dragon heads north of 5 tonnes. Factoring in the uncertainties and the need for larger drag devices for Mars, I'd suggest that 7 tonnes is a conservative figure for a fully fueled lander of this type. Some have suggested much higher than this.

In addition you need an ascent vehicle landed separately. Now, with a minimalist vehicle (such as my own) then you have a take off mass of 5 tonnes (plus fuel). Some of that mass is crew. And some of it is life support consumables that can be produced on the surface. So lets be fair and call it 4 tonnes of ascent vehicle.

Then there is the imported methane fuel (I'll consider hydrogen shortly). A 5 tonne ascent vehicle needs 13 tonnes of fuel. That's 2.88 tonnes of methane. If its imported you probably need 3 tonnes of methane to start with. Now to import the methane you need a vehicle to land it in. The most economical way of course is to land the methane with the ascent vehicle, bringing the total ascent vehicle landed mass to 7 tonnes.

A conventional landing craft has a mass penalty of about 2 to 1, so 7 tonnes of payload means 14 tonnes at entry.

Which takes us down to a comparison between 35 tonnes delivered to high Mars orbit, then aerobraked. Versus 21 tonnes at Mars atmospheric entry (we assume direct entry). The former amounts to 111 tonnes in LEO. The latter amounts to 57 tonnes in LEO.

Total saving: 54 tonnes in LEO.

Now, that's nothing to be sneezed at. At $2000/tonne that's a saving of a bit over $100M in launch costs per mission.
But that should be seen against the backdrop of billions in development costs. And development costs are driven by complexity and time.
I do agree that in the long term, if we decide we want to return to Mars many times over, then an indigenous source of hydrogen is the key.

----

Now, to do a fair comparison regarding the importation of hydrogen I'm going to use as a baseline the conventional mission described above. A conventional 7 tonne single use lander and a previous landing of an ascent vehicle with methane fuel on board which added up to 14 tonnes.

From this baseline we derive a version where hydrogen is imported instead of methane.

I'm first of all going to introduce hydrogen losses into the equation. Its not just boil off, its also the losses incurred by passing hydrogen through the plumbing and heat exchangers of a ISPP plant. In short, hydrogen is innately leaky. I'm going to assign a factor of 50% for overall loss. That's even with very good insulation and some measure of active refrigeration, but allowing for leakage and for a safety margin (we're assuming the hydrogen is our first and only source of ascent fuel). With that we need 1125Kg of hydrogen.

Landed in a conventional lander, 1125Kg of hydrogen versus 3000Kg of methane translates to a difference of 3.75 tonnes at Mars entry. With the direct entry delta-V of 4.6 assumed above, that translates to a saving of 12.8 tonnes initial mass in low Earth orbit.

Now that's the scale of the maximum possible theoretical saving. About 13 tonnes of mass, and about $26M of launch costs but in return all the issues and risks inherent in hydrogen.

Of course if you decided you wanted to use mostly propulsive landers for other reasons, and still wanted to deliver hydrogen rather than methane. Then the saving in IMLEO terms in delivering hydrogen would be roughly 6 tonnes. Of course that's the theoretical maximum saving.

Beyond that theoretical best case the real world detail doesn't look pretty.

For a start, the ISPP plant needed to produce methane from hydrogen is likely to be more massive than the alternative which simply produces oxygen. Possibly by a factor of 2. The reason is that it involves more parts and more plumbing and heat exchangers (or else you sacrifice efficiency). Now its fair to say that such a plant is likely to see service for years so it can't be accounted for on a per mission basis, but the issue is there. The same applies to the power source. And again the task of producing methane from hydrogen is much more energy intensive. For two reasons. One is the inefficiencies in the process, especially if done on a small scale. The other is the refrigeration costs. Often we're dealing with very hot gases but the ultimate product is a hard cryogen. I won't go further into this but its a real issue.

More importantly in terms of its effect on mass per mission, is that 3 tonnes of methane occupies a bit over 7m3 of tankage. 1.125 tonnes of hydrogen occupies closer to 16m3 of tankage. That presents problems both on the ground and more importantly in terms of how you land it. Any lander, mostly propulsive, or conventional suffers the issue that larger tankage translates to larger structural elements. Worse, the conventional lander is hit hardest because with the higher stresses, a large tank of hydrogen just multiplies the mass of structure and heat shield. This factor alone probably removes half of the real advantage of hydrogen.

There are also other system wide mass costs from importing hydrogen that go all the way back to supply of further power and refrigeration equipment right back to low Earth orbit. In the end the advantage of importing hydrogen over methane (even assuming conventional landers) may evaporate.

This is not to say that hydrogen isn't the future. It is. Its indigenous hydrogen we need. And for that we need machines that dig and process dirt. And that's quite probably an argument for having people there, to manage the prototypes. But I'm getting off track here.

------

Back to the big picture. A singular lander and ascent vehicle design is an exercise in reducing complexity and saving development costs. It also affords you a platform that serves many needs. For one thing, a lander becomes the taxi that transports crew between high and low Mars orbits, thus enabling the practice of parking the transit vehicles in high orbit. That factor alone saves the 2.6Km/s of delta V otherwise involved in bringing the transit vehicles down to low orbit and back. That would double the overall mass in and of itself.

Next I'm using the lander as the primary propulsion source for the return journey. Now the transit vehicles have their own engines but these are for backup purposes and its because we're relying upon the lander as the main engine we can then mass reduce the transit vehicle.

Next, the lander is capable of running surface to surface hopping missions. Meaning that its capable of long distance hops from base to base.

Next, the lander is capable of simple missions to the Phobos and Deimos. Indeed for a reasonable amount of fuel one could stop to claim samples from those moons on the way back to high Mars orbit.

Next, a mostly propulsive landing isn't as given to the vagaries of the atmosphere and has more capability to land on higher landing sites.

These synergies actually reduce the total mass of doing all of these things in other ways.

....

Inflatable heat shields are a great idea - if you plan on an expendable vehicle. I've no doubt that they, or something similar will come in very handy when it comes to landing large objects on Mars. But the problem with an inflatable heat shield is that it is still added, mass complexity and risk. And the bigger it is the more stability issues it has. Again, a necessary evil for large payloads, but lets make these large payloads unmanned.

The more you rely upon stuff that you eject during landing, the more you end up with the cost (and risk) of garbage.

Now, I would be willing to consider that in the case of a mostly propulsive, reusable crewed lander, a supersonic decelerator (probably a towed one) could be useful. But only to provide further margin. In other words you can land without it but it means you have more time to consider your landing site. Here you have two advantages. Firstly you're not dealing with extremely high dynamic loads and temperatures and you're not contemplating operating above Mach 3. Which means a lighter decelerator. Think in terms of being man handled and replaced easily. The other advantage is that because of the lower velocities involved you can afford to deploy the decelerator virtually from interface, or as soon as there is enough density to make it worthwhile.

I agree that by the time we get to land on Mars we will probably have real time, detailed observations of the atmosphere. Even a mostly propulsive landing would (especially if the landing site is higher) require some consideration as to the atmospheric conditions.

Which brings me to another point. One bug bear I have with direct entry is that you don't get much choice about the atmospherics. But if you based your architecture around having a safe (life support etc) haven in low Mars orbit, you're then free to choose when you land. Which means you can time it for the best atmospherics. And again, having an outpost in low Mars orbit does come at a cost, but it does have its advantages in terms of safety issues like this, plus as I said, you can keep most of your transit mass in a high orbit.

----

In some ways I'm not particularly fussed about L2 versus LEO. Certainly if you're in the business of large scale fabrication (over sized heat shields, trusses etc) then I'd definitely go for LEO. Remember that what I propose does not involve anything more than bump together assembly. No EVA required.

However the bigger issue is that you get the choice between either assembling (or rather docking) everything in LEO and then using conventional propulsion to Mars, or if you're going to adopt a solar tug then L2 is possibly the better option. Yes, you could just use a solar tug to lift the whole thing into a high elliptical Earth orbit but the whole thing would be traversing the radiation belts on every orbit. You'd then have to time the crew transfer well and get them on board quickly. Possible, yes.

L2 does give you more time to do check out systems (again that's in the context of using solar tugs).

I guess if you pushed me, I'd go for a conventional launch from LEO. At least at first. But remember that if you're seriously concerned about mass delivered to low Earth orbit, you wouldn't do that. You'd develop the solar tug first. Again, its useful to have a sense of perspective.

....

Now regarding the final point about reliability. For me there really are two enabling technologies that we need to go to Mars under any architecture that we don't have.

One is supersonic retro propulsion. A fact of life even for more conventional landers. And I need to emphasise this. Large payloads are never going to go subsonic in the Mars atmosphere. We need to understand SRP. Simple as that.

The other is enabling technology is developing a methane/lox engine that's very reliable. And that's the case even without a propulsive lander. As soon as you rely upon methane/lox engines to get you home, then you need an engine that can survive long burns, multiple restarts and long idle periods in space conditions.

It also has to be added that there's no such thing as a non propulsive lander. Sooner or later you're going to have to use retro propulsion. And you'd hope your engines are reliable.

What I've done though in having an architecture that relies upon a mostly propulsive landing for the crew is that I can achieve the goal of having one universal engine design for all purposes. And only having two basic vehicles. Which means a lot of the supporting hardware is of common design. Which means the capability to swap components.

(The above sentence ignores the big dumb H2/LOX booster which I think everyone takes for granted).

Then I've built in multiple layers of redundancy. Typically the lander itself would be flying with 8 engines. Four directly down. Four pointed about 15 degrees to the side. There's a certain amount of redundancy there that I won't go into right here. Where redundancy also comes in is that instead of landing the ascent fuel in a purpose built, expendable vehicle, I am instead landing the fuel in another lander. Which means that there's a source of spare parts and spare engines waiting on the surface. Indeed it might be possible in an emergency to retrofit the fuel lander as a crewed ascent vehicle.

The redundancy continues in space. Instead of discarding the lander, I'm bringing it back to orbit and then on to high orbit. There it provides the primary source of propulsion for the return journey. But if it fails there are still 2 engines on each of the transit vehicles. Again, its designed so that you can swap over engines if needed.

Your final fall back is to abandon one of the space habs, transfer fuel across and limp home (Ok, its cramped but you're safe). I don't see any other architecture providing that final layer of safety.

And technically there is one final refuge, and that's the lander itself.

----

I hope that answers things?

Btw, I was hoping to be asked about turbulence

Edit: Ok, now I've managed to get this post into some readable shape.

Edit 2: Since I brought up turbulence I might was well answer it.

This is somewhat more speculative. I'm simply making an argument that a good solution may be possible, at least for a vehicle without a major heat shield.

What little evidence I have about SRP is what can be found on the web. Some of it discusses the reduction in effective drag coefficient caused by an engine pushing the air away from the base of the vehicle. Some of it is the video showing clearly the turbulence that is caused by firing an engine into a supersonic air stream. Its fair to say that a lot of discussion about SRP is framed by the assumption that the engine or engines will be project through a conventional heat shield. Obviously I'm stepping outside that convention. And what I can't find is information that relates the nature of the turbulence to the density and relative velocity of the stream. Nor is there information about the nature of the turbulence. What the physical and time scales are and again, how those are governed by density and so on.

There are basically 3 regimes in which a mostly propulsive lander would have to work. The first is essentially outside the atmosphere. Here you have no constraint since there is no atmosphere to provide turbulence. In doing my rough math I've figured that this regime extends down at least to about 100Km. Now the reason for getting this point as low as possible is that the lower you do your initial large braking burn, the more efficient you are because of lower gravity losses.

The key question is, when does the atmosphere become dense enough for significant turbulence to apply. It might exist, but the forces might be too small to be of consequence. If I recall correctly, at around 100Km the air is still measured in micro grams per cubic metre. Its also quite possible that thermal considerations might kick in before turbulence from propulsion becomes an issue. In other words, it becomes necessary to apply (some of) the initial braking burn to keep heating below an acceptable level. The acceptable level I'm thinking of is around 0.3W/cm2. In other words something akin to being exposed to 3KW per m2. That's about 2 suns in Earth orbit. 

The third regime is below the point at which full retro propulsion is required in order to effect a safe landing. Which is probably somewhere below 15Km. The key here is that there is a trade off involved. Wait longer and normal atmospheric drag will continue to slow your descent and thus lower the effect of turbulence when you apply retro propulsion. Wait longer and you need to apply more thrust.

The second regime is obviously between those two extremes.

Now, one thing that can help is to move your engines away from the body of the vehicle. This has two potential benefits. One is that the region of lower pressure created by the engine can extend away from the body of the vehicle (especially with engines directed away from the vertical) and this should have the effect of reducing the effect of propulsion on normal drag. A lot of numbers I have seen give effective drag coefficients based on engine thrust, but do so on the assumption that the engine is directly below the vehicle and thus will create lower pressure under it. The other reason for moving the engines away from the vehicle is to remove as much of the turbulent wake of the engine as possible away from the body of the vehicle. Now the theory here is that the effects of turbulence are only important to the extent that changing air pressure can react against something.

So in essence you end up with a design that has engines mounted on a four star truss that projects beyond the body of the vehicle and also keeps the engine as high as possible. Quite probably integrated with the design of the landing legs. On each point you have two engines. One directed downwards. The other projected about 15 degrees off vertical and outwards.

Now, in regime 1, the key is to brake as late and as hard as possible. So you use all 8 engines at first, and then switch to the 4 engines that are pointed outwards and continue to complete the initial brake. My hope is that you can do this well below 100Km.

In regime 2, its yet to be determined what the optimum is. One suspects the optimal strategy is either a continued low thrust, or none at all (for the sake of engine efficiency at lower powers). An advantage of continued low thrust perhaps is to avoid re-ignition problems further down.

In regime 3 I suspect that you'd use the outer 4 engines at low thrust initially and then gradually build thrust as velocity falls. As velocity and thus turbulence becomes less of an issue you then bring in the 4 vertically projected engines and quickly go to full power. And of course as you approach final landing you'd throttle back and use the outer 4 engines.

Now all of that is based on the educated guess that turbulence can and will develop but so long as it has minimal surfaces to react against, its manageable.

And the nice thing about a mostly propulsive landing is that its easier to get away with doing that. Because your engines don't have to be hidden behind a heat shield they can be positioned where you like them and away from the body of the vehicle