I would have thought so too. But we hotfired an N2O thruster and sampled the exhaust gases- no detectable NOx. I am no chemical engineer but I suspect it has to do with the temperature and pressure conditions driving the thermochemisty. There is a company that holds a patent on a N2O breathing air generator- clearly if you control things you can avoid getting a lot of undesirable species.
As some one who works with hydrogen on a daily basis let me assure you that it is a true pain to deal with as compared to many other gases. It diffuses through many polymers and leaks are extremely dangerous due to its wide combustion mixtrue ratio range. WIth an invisible flame you can walk right into a large hydrogen fire. To get decent densities for storage you are working with either very high pressures or liquified H2. Both of these are problematic. One imposes hydrogen embrittlement issues, large heat of compression losses and many materials are useless and the other demands exquisite thermal control and imposes many other materials limitations. Hydrogen is a great fuel but only for certain uses and I would not say that everyday transport is one. It MAY be acceptable for fixed-base use in industry and less possibly in homes.
Transport batteries ( I think we all agree that is what we are discussing here) require a few things to be practical: low cost of materials and ease of fabrication, high energy density, ease of movement of the material from one vessel to another and finally ease of synthesis and also conversion efficiency. Non-toxicity is important as is the effect on the atmosphere. There are very few materials that can match or better liquid hydrocarbons.
There is one candidate that should at least be considered. Nitrous Oxide. N2O is a saturated fluid under about 750psia at room temperature and it has a density the same as hydrocarbons. This means that vessels to store it are efficient. It is non-toxic although it is an anesthetic gas. It is very safe to handle and compatible with nearly all materials. This means that the devices to handle it are cheap to make. It is a liquid so heat of compression losses for movement are minimized. If it leaks it has a distinct odor and will generally not pose an explosion hazard- at least compared to H2.
N2O is a monopropellant- in other words it will decompose to N2 and O2 when passed over a heated catalyst. It reacts very completely and almost no NOx species are produced- good for pollution. Better still it has a high flame temperature which makes for high thermodynamic efficiency. So a turbogenerator running N2O does not have to have a compressor- it can work at least part of the time off of the storage tank source pressure. Heat from the environment or directed waste heat from the exhaust can help keep the remaining N2O warm and vapor pressure high. N2O has a decent energy density but more importantly you can add any fuel and increase the power release enormously. So you power with N2O when you can and add fuel when you need to accelerate. The power increase is rapid and significant.
It does have problems though- synthesis is complex and not presently at large scale. What would be great is to develop a catalytic system that could take atmospheric N2 and O2 and under proper conditions directly synthesize N2O which could then be stored. Sounds hard to me but you never know. In any case there is no shortage of the precursors. It is however a nasty greenhouse gas. This could be its worst issue- lareg releases of unreacted N2O could be worse than CO2. But at least these are accidental and incidental- not part of everyday operation.
Anyway it is something to ponder. I always thought that a N2O vehicle with ethanol fuel assist sounded pretty good- and what a party car!
Well you didn't include the cost to develop metal and to liquify cryogens either so where does it end? Here is the situation: there is hardly a single piece of hardware that was developed for the original Atlas or Centaur that is on the present Atlas V. Every system has seen multiple upgrades and simplifications. Elon Musk also hired a bunch of folks with some previous experience so he too benefited from prior work. Lets be honest and compare near term costs. The Atlas V had a pricetag and it was not $12 billion.
The RD-180 engine as shown is an all-up ready to fly engine with a complete operational set of instrumentation as well as the basic plumbing for moving propellants around. That adds a few tubes and wires here and there. The photo of the Merlin is a prototype on a test stand. The test stand itself replaces much of that flight hardware. In terms of basic function they are very similar. A pump, injector, control valves and chamber. Except that one has never failed in flight and the Merlin failed its first time and had a raft of problems before then. Ironically the image you chose is in an article describing how they cooked their control wires on the test stand leading to hardware damage. This behavior has been known for DECADES but they chose either by ignorance or hubris to take no preventive measures. What does this say about the level of sophistication of the team?
Your next comment on complexity is pretty funny- they are proposing to use NINE of these engines on a Falcon 9. They must work in synchrony and not interact too much with each other. This is non-trivial and ended up burning the Russians in the past. Surely you will agree that a single RD-180 is not nine times more complex than a Merlin. This complex rocket can get you all of 8.7mt to LEO and a measly 3.1mt to GTO. A base level Atlas V 401 with ONE booster engine delivers 12.5mt to LEO and 5t to GTO- fully 50% greater capability. This means you will have to launch a lot more of these non-existent F5's to match delivered payload-if in fact their performance calcs are correct- they have never actually made it to orbit. Think that a mistake there is unlikely?- the Delta folks got severely burned on the Delta III performance because they chose to play it cheap and omit a propellant utilization system- a few $100K. And F5 has no real capability to GSO.
So Elon consumes 15 Merlins per 1 RD-180 on a payload basis. All of those Merlins must work or there is a mission failure. Of course I could fly an Atlas 551 and get performance up to 20mt to LEO by adding some flight proven solids- the very same ones that got us going to Pluto with a nuc powered payload. There- I use 6 booster engines and Elon uses 9 and I get over twice the delivered performance. Yep I'm twice as expensive- is that a ripoff?
Lets say YOU had to launch a $70M payload which is time critical. Lives may be at stake. You are in the news. Astronauts are depending on the replacement widget you MUST deliver. Which would you bet your money on- a rocket made by a team with a dicey track record or one that has not failed in well over a decade? Now change the scenario to you being responsible for spending someone else's money and having a duty to do the right thing. Your job and reputation are on the line. Do you play it cheap and take a chance or spend a few millions more for a virtual guarantee? Remember this no hobby or game this is FOR REAL. With Atlas you as the customer get complete access to technical assessments of the health of the rocket by a time-proven team. You can see every change made and demand complex analyses be undertaken to address your concerns and doubts. You know they will be done by some of the best in the industry. You have hundreds of flights of data to compare to. You can see that minor issues are worked as if they are show-stoppers and brought to resolution in a open forum with you, your engineers and your mom there if you like. The best efforts of the best team are focused on YOUR success.
The "intended" launch rate for the Atlas V and Delta IV was 20 vehicles per year. They are right now flying 5/year. The cost of the metal in the machines is not the issue. The Atlas V especially is a very simple machine with far fewer elements and components than an equivalent Falcon. Costs are in the people to support the missions and keep the machine alive through years.
How much do you think LM and the USG spent on the Atlas V? It was a bargain- for less than $1.5B you got an entire family of launch vehicles- 8 variants in all including an upper stage that can send you direct to GSO or provide the energy to go to Pluto. You got two brand new state of the art, high launch rate launch pads. Advanced avionics, the most advanced guidance, fluids and propulsion systems. This means they freaking work when they need to (not just when you get lucky). Like when the payload is people. Or something important for ISS. You got 5 different payload fairings including one that would hold most of a Falcon 1.
For equivalent performance and capability Elon will have spent just about the same. Of course he is sitting at 0% reliability right now. Atlas is at 100% for the past 79 launches-including the last two decades worth of NEW vehicle first flights. I bet it cost Elon many millions to just figure out what happened on his first flight and fix the problems. I've heard there were mulitple issues uncovered by the failure analysis. You get to add that to flight costs- unless this is a hobby of his.
Here is the bottom line: NASA selected the two "finalists" not with an eye to replacing the shuttle with commercial vehicles for ISS resupply but to provide a justification for the continued development of the horrid CLV "stick" launch vehicle. They know that these two teams will fail to meet the technical and cost targets. Especially the Kistler design- that is a total loser. Their failure will show that NASA just HAS to use the CLV even though it is probably five times the cost. They can say they tried but that industry was simply not up to the task. Total crap. The Atlas V was teamed with one of the COTS proposers and lost since they were decremented because of their intent to use existing, proven, low cost launchers. No one who does not build their rocket in a third world country can compete with an Atlas V 401 on cost per pound to orbit. But this obvious winner was not selected in favor of high-risk and low capability systems.
The whole thing is a sick joke- that the taxpayers get to foot the bill for.
By the time Elon has met all of NASA's demands for manned ops and ISS rendezvous and actually learned to fly successfully his total costs will be nearly identical to those of Atlas or Delta. These are not big teams contrary to popular conceptions. When you are in a prototyping phase you can afford to screw around and reduce operational discipline. His first launch attempt was a comedy of errors. Collapsed tanks, lost propellants- it was almost funny to watch. If he was not footing most of the bill the whole thing would have been shut down on the basis of inherent incompetence.
There are NO short cuts to successful space flight.
Be aware that there already has been a revolution in vehicle design in which basics like the rocket equation are respected and they result in cost/pound to LEO levels that are on the order of ten times better than either Saturn or most especially Shuttle. Stop thinking so narrowly about LEO- think about how you get to the moon or Mars. This thought process will show you that winged vehicles and the like are horrible solutions. They are optimized for the final 30 minutes of flight- not the months that you must spend in deep space in order to perform serious exploration. This is sort of like choosing what shoes to run a race in based on what would look best at the awards party- yes the Italian heels would be better but will not help you in the least during the race- where it counts.
Generally speaking placing mass forward on a rocket is a GOOD thing- ignoring aero induced loads for the time being. Moving Cg forward allows greater control authority - engine out is also easier.
Anyone who thinks that all innovation is captured in patents is clearly not involved in the design process. I've got a bunch of patents but the coolest ideas we NEVER commit to a patent unless there is an overriding strategic/competitive reason. Innovation is best kept secret and imbedded. Most innovation is invisible to the user and it simply does not pay in most cases to reveal an underlying technique to the whole world who can then modify the process in some small way and gain 90% of the benefit of the patent. And there is no patent police- you have to find a clear violation. All this and you get a whole 20 years from time of filing- and you have to pay to maintain the thing too. Most innovation has a lifetime that is only a few years at most- so really you get maybe 10 years max of real utility before you are overcome by events.
This report is pure crap and is based on an incredibly blinkered perspective. Innovation- and by that I mean real workable ideas that can be used in the marketplace and have been debugged- is exploding. The changes in materials alone will swamp you. It is barely possible to stay current in say polymers and metals and their associated processing techniques- throw in optical materials and its all she wrote.
Clearly, as you say, you cannot throw design rules to the wind and expect rockets of arbitrary design to fly with arbitrary loads. However to make my point about scaling - the key thing is not the total dry mass of a particular stage but instead the mass fraction of the propellant compared to total mass. This flows directly into the rocket equation which relates the change in mass of a stage to delta V. In general the greater the capacity of a tank the greater its mass fraction if the design rules it followed are consistent. You can, of course, create poor designs with excessive L/D ratios and choose crummy manufacturing and materials to undermine this. My point is that scaling up volumes and masses is not in any way exhausted as of today. Most designers would agree that even further significant reductions will be available in the next ten years. There are a multitude of manufacturing and materials technologies that will aid this.
Booster Engine throttling is commonly done on modern rockets and in general you can get around a 50% turn down on most engines before bad things happen. This is of course dependent on whether the engine designer anticipated this. This allows the suppression of peak acceleration which would otherwise occurr at the end of booster engine burn. This capability is often overlooked and is a powerful tool.
The airbreathing concepts you mention are interesting but seem to require elaborate and costly infrastructure that has to be maintained etc. I've not done the math but they seem to have only small advantages with large attendant costs. remember you have to keep that infrastructure alive and crews in practice even when you don't launch. Those overheads can be trouble. This is the lesson of shuttle. Each system architecture has a range of launch rates and utilization over which it is cost effective. Fall out of that range and you are dead. Most reuseable systems have launch rate optima that are WAY outside the present demand. but in 20 years who knows...
As someone who designs next gen systems for a living perhaps some additional thoughts:
1) As rockets grow their structural weight does not balloon like a skyscraper. All rockets are mostly propellant ( if they are well made) and that propellant ( talking liquids here) has a vapor pressure associated with its temperature. Since it has to fill a tank at sea level this generally means that the O2 or H2 must have a vapor pressure around 15 psia- unless the propellants are subcooled. Otherwise the empty space above the liquid (the ullage) is pressurized with a gas of some sort. Maximum loads are not generally at liftoff- they are at max Q- roughly 40,000 to 60,000 ft where the ambient pressure is 5 psia or less. This means generally that you get 10 psid of internal pressure for free which acts to alleviate axial loads. Check out how much upforce is available in a cylinder that is 27 feet in diameter as on ET @ 10 psi. It is....large. Want more? Increase pressure. Also remember liquid head effects. SO! all modern vehicle designs make extensive use of internal pressure to stabilise primary structures. Also the tank diameter directly bears on the bending stiffness of the vehicle which is essentially a beam in bending - the main force being lateral aeroloads. We have not reached any sort of limit on rocket size - but transporting these massive structures does get costly and tedious. Anything over 18 ft is a pain due to transport issues.
2) Although ramjet/scramjet systems are certainly interesting their total operational time in an efficient trajectory will be measured in small numbers of seconds and the total impulse delivered by them will be small. And they are not zero weight. They want to fly at high Q for extended periods- just what the rest of the vehicle does not want- you are intensifying loads on everything else.
3) It is important to recognize that there are really two parts to a launch trajectory. The first is all about THRUST- getting the thing off the deck with at least 1.2 to 1.8 Gs. Otherwise you are accelerating too slow and pay a big penalty. You want to get up quickly to 5-6 G acceleration and HOLD it for some time. The optimal solution for this is fairly obvious by now- it is Lo2/kerosene and/or solids. With this big thrust and modern Isp's you can heave a high Isp upper stage up to around 1/3 to 1/2 of orbital velocity. THAT stage must strike a balance between thrust, dry weight and Isp. That balance is different depending on whether you are going to LEO or to the moon. Suffice it to say that BURNOUT WEIGHT is KING. Notice that this is not just dry weight- residuals are super important. Bungle those are your performance will suck.
4) Systems for getting tourists to LEO using PRACTICAL methods at reasonable cost ( might be a bit more than a million though) have been evaluated and can be executed if someone wanted them for real. Suffice it to say that they are unlike anything yet proposed- very cool. The real key to prices normal people can afford is RATE production. And there has to be a "THERE" there. If the machine is hand built like Rutan or anyone else it is a non-starter. It needs to be bulletproof and not require engineering hand-holding. Our handholding is some of the most expensive you can buy. That is NOT so easy.
The bottom line is that with hardware that is already in production it is possible to not only achieve a launch architecture that can not only match Saturn but do it at probably 30-50% of the recurring cost. There is NOTHING magic about Saturn- it was a complex, single purpose and extraordinarily expensive way to get to orbit. That is just about all Saturn was able to do- that and a little bitty trans-lunar injection burn. Short missions and you need a Navy to recover the puny little capsule. The real solution lies in using rate-production hardware that can meet the entire architecture- from earth to LEO to lunar orbit to lunar surface and back with extensibility to Mars. This needs to be done with economy otherwise we will NEVER to Mars because it will be too expensive. Going to the moon is the litmus test of overall economic viability. If we cannot afford at least 4-8 missions to the moon in one year we are deluding ourselves about Mars.
Slashdot Mirror
I would have thought so too. But we hotfired an N2O thruster and sampled the exhaust gases- no detectable NOx. I am no chemical engineer but I suspect it has to do with the temperature and pressure conditions driving the thermochemisty. There is a company that holds a patent on a N2O breathing air generator- clearly if you control things you can avoid getting a lot of undesirable species.
Transport batteries ( I think we all agree that is what we are discussing here) require a few things to be practical: low cost of materials and ease of fabrication, high energy density, ease of movement of the material from one vessel to another and finally ease of synthesis and also conversion efficiency. Non-toxicity is important as is the effect on the atmosphere. There are very few materials that can match or better liquid hydrocarbons.
There is one candidate that should at least be considered. Nitrous Oxide. N2O is a saturated fluid under about 750psia at room temperature and it has a density the same as hydrocarbons. This means that vessels to store it are efficient. It is non-toxic although it is an anesthetic gas. It is very safe to handle and compatible with nearly all materials. This means that the devices to handle it are cheap to make. It is a liquid so heat of compression losses for movement are minimized. If it leaks it has a distinct odor and will generally not pose an explosion hazard- at least compared to H2.
N2O is a monopropellant- in other words it will decompose to N2 and O2 when passed over a heated catalyst. It reacts very completely and almost no NOx species are produced- good for pollution. Better still it has a high flame temperature which makes for high thermodynamic efficiency. So a turbogenerator running N2O does not have to have a compressor- it can work at least part of the time off of the storage tank source pressure. Heat from the environment or directed waste heat from the exhaust can help keep the remaining N2O warm and vapor pressure high. N2O has a decent energy density but more importantly you can add any fuel and increase the power release enormously. So you power with N2O when you can and add fuel when you need to accelerate. The power increase is rapid and significant.
It does have problems though- synthesis is complex and not presently at large scale. What would be great is to develop a catalytic system that could take atmospheric N2 and O2 and under proper conditions directly synthesize N2O which could then be stored. Sounds hard to me but you never know. In any case there is no shortage of the precursors. It is however a nasty greenhouse gas. This could be its worst issue- lareg releases of unreacted N2O could be worse than CO2. But at least these are accidental and incidental- not part of everyday operation.
Anyway it is something to ponder. I always thought that a N2O vehicle with ethanol fuel assist sounded pretty good- and what a party car!
The RD-180 engine as shown is an all-up ready to fly engine with a complete operational set of instrumentation as well as the basic plumbing for moving propellants around. That adds a few tubes and wires here and there. The photo of the Merlin is a prototype on a test stand. The test stand itself replaces much of that flight hardware. In terms of basic function they are very similar. A pump, injector, control valves and chamber. Except that one has never failed in flight and the Merlin failed its first time and had a raft of problems before then. Ironically the image you chose is in an article describing how they cooked their control wires on the test stand leading to hardware damage. This behavior has been known for DECADES but they chose either by ignorance or hubris to take no preventive measures. What does this say about the level of sophistication of the team?
Your next comment on complexity is pretty funny- they are proposing to use NINE of these engines on a Falcon 9. They must work in synchrony and not interact too much with each other. This is non-trivial and ended up burning the Russians in the past. Surely you will agree that a single RD-180 is not nine times more complex than a Merlin. This complex rocket can get you all of 8.7mt to LEO and a measly 3.1mt to GTO. A base level Atlas V 401 with ONE booster engine delivers 12.5mt to LEO and 5t to GTO- fully 50% greater capability. This means you will have to launch a lot more of these non-existent F5's to match delivered payload-if in fact their performance calcs are correct- they have never actually made it to orbit. Think that a mistake there is unlikely?- the Delta folks got severely burned on the Delta III performance because they chose to play it cheap and omit a propellant utilization system- a few $100K. And F5 has no real capability to GSO.
So Elon consumes 15 Merlins per 1 RD-180 on a payload basis. All of those Merlins must work or there is a mission failure. Of course I could fly an Atlas 551 and get performance up to 20mt to LEO by adding some flight proven solids- the very same ones that got us going to Pluto with a nuc powered payload. There- I use 6 booster engines and Elon uses 9 and I get over twice the delivered performance. Yep I'm twice as expensive- is that a ripoff?
Lets say YOU had to launch a $70M payload which is time critical. Lives may be at stake. You are in the news. Astronauts are depending on the replacement widget you MUST deliver. Which would you bet your money on- a rocket made by a team with a dicey track record or one that has not failed in well over a decade? Now change the scenario to you being responsible for spending someone else's money and having a duty to do the right thing. Your job and reputation are on the line. Do you play it cheap and take a chance or spend a few millions more for a virtual guarantee? Remember this no hobby or game this is FOR REAL. With Atlas you as the customer get complete access to technical assessments of the health of the rocket by a time-proven team. You can see every change made and demand complex analyses be undertaken to address your concerns and doubts. You know they will be done by some of the best in the industry. You have hundreds of flights of data to compare to. You can see that minor issues are worked as if they are show-stoppers and brought to resolution in a open forum with you, your engineers and your mom there if you like. The best efforts of the best team are focused on YOUR success.
How much do you think LM and the USG spent on the Atlas V? It was a bargain- for less than $1.5B you got an entire family of launch vehicles- 8 variants in all including an upper stage that can send you direct to GSO or provide the energy to go to Pluto. You got two brand new state of the art, high launch rate launch pads. Advanced avionics, the most advanced guidance, fluids and propulsion systems. This means they freaking work when they need to (not just when you get lucky). Like when the payload is people. Or something important for ISS. You got 5 different payload fairings including one that would hold most of a Falcon 1.
For equivalent performance and capability Elon will have spent just about the same. Of course he is sitting at 0% reliability right now. Atlas is at 100% for the past 79 launches-including the last two decades worth of NEW vehicle first flights. I bet it cost Elon many millions to just figure out what happened on his first flight and fix the problems. I've heard there were mulitple issues uncovered by the failure analysis. You get to add that to flight costs- unless this is a hobby of his.
Here is the bottom line: NASA selected the two "finalists" not with an eye to replacing the shuttle with commercial vehicles for ISS resupply but to provide a justification for the continued development of the horrid CLV "stick" launch vehicle. They know that these two teams will fail to meet the technical and cost targets. Especially the Kistler design- that is a total loser. Their failure will show that NASA just HAS to use the CLV even though it is probably five times the cost. They can say they tried but that industry was simply not up to the task. Total crap. The Atlas V was teamed with one of the COTS proposers and lost since they were decremented because of their intent to use existing, proven, low cost launchers. No one who does not build their rocket in a third world country can compete with an Atlas V 401 on cost per pound to orbit. But this obvious winner was not selected in favor of high-risk and low capability systems.
The whole thing is a sick joke- that the taxpayers get to foot the bill for.
By the time Elon has met all of NASA's demands for manned ops and ISS rendezvous and actually learned to fly successfully his total costs will be nearly identical to those of Atlas or Delta. These are not big teams contrary to popular conceptions. When you are in a prototyping phase you can afford to screw around and reduce operational discipline. His first launch attempt was a comedy of errors. Collapsed tanks, lost propellants- it was almost funny to watch. If he was not footing most of the bill the whole thing would have been shut down on the basis of inherent incompetence. There are NO short cuts to successful space flight.
Be aware that there already has been a revolution in vehicle design in which basics like the rocket equation are respected and they result in cost/pound to LEO levels that are on the order of ten times better than either Saturn or most especially Shuttle. Stop thinking so narrowly about LEO- think about how you get to the moon or Mars. This thought process will show you that winged vehicles and the like are horrible solutions. They are optimized for the final 30 minutes of flight- not the months that you must spend in deep space in order to perform serious exploration. This is sort of like choosing what shoes to run a race in based on what would look best at the awards party- yes the Italian heels would be better but will not help you in the least during the race- where it counts.
Generally speaking placing mass forward on a rocket is a GOOD thing- ignoring aero induced loads for the time being. Moving Cg forward allows greater control authority - engine out is also easier.
not only can internal insulation be done on cryogenic tanks but it WAS done on SAturn and is done on Ariane today
Anyone who thinks that all innovation is captured in patents is clearly not involved in the design process. I've got a bunch of patents but the coolest ideas we NEVER commit to a patent unless there is an overriding strategic/competitive reason. Innovation is best kept secret and imbedded. Most innovation is invisible to the user and it simply does not pay in most cases to reveal an underlying technique to the whole world who can then modify the process in some small way and gain 90% of the benefit of the patent. And there is no patent police- you have to find a clear violation. All this and you get a whole 20 years from time of filing- and you have to pay to maintain the thing too. Most innovation has a lifetime that is only a few years at most- so really you get maybe 10 years max of real utility before you are overcome by events. This report is pure crap and is based on an incredibly blinkered perspective. Innovation- and by that I mean real workable ideas that can be used in the marketplace and have been debugged- is exploding. The changes in materials alone will swamp you. It is barely possible to stay current in say polymers and metals and their associated processing techniques- throw in optical materials and its all she wrote.
Clearly, as you say, you cannot throw design rules to the wind and expect rockets of arbitrary design to fly with arbitrary loads. However to make my point about scaling - the key thing is not the total dry mass of a particular stage but instead the mass fraction of the propellant compared to total mass. This flows directly into the rocket equation which relates the change in mass of a stage to delta V. In general the greater the capacity of a tank the greater its mass fraction if the design rules it followed are consistent. You can, of course, create poor designs with excessive L/D ratios and choose crummy manufacturing and materials to undermine this. My point is that scaling up volumes and masses is not in any way exhausted as of today. Most designers would agree that even further significant reductions will be available in the next ten years. There are a multitude of manufacturing and materials technologies that will aid this. Booster Engine throttling is commonly done on modern rockets and in general you can get around a 50% turn down on most engines before bad things happen. This is of course dependent on whether the engine designer anticipated this. This allows the suppression of peak acceleration which would otherwise occurr at the end of booster engine burn. This capability is often overlooked and is a powerful tool. The airbreathing concepts you mention are interesting but seem to require elaborate and costly infrastructure that has to be maintained etc. I've not done the math but they seem to have only small advantages with large attendant costs. remember you have to keep that infrastructure alive and crews in practice even when you don't launch. Those overheads can be trouble. This is the lesson of shuttle. Each system architecture has a range of launch rates and utilization over which it is cost effective. Fall out of that range and you are dead. Most reuseable systems have launch rate optima that are WAY outside the present demand. but in 20 years who knows...
As someone who designs next gen systems for a living perhaps some additional thoughts: 1) As rockets grow their structural weight does not balloon like a skyscraper. All rockets are mostly propellant ( if they are well made) and that propellant ( talking liquids here) has a vapor pressure associated with its temperature. Since it has to fill a tank at sea level this generally means that the O2 or H2 must have a vapor pressure around 15 psia- unless the propellants are subcooled. Otherwise the empty space above the liquid (the ullage) is pressurized with a gas of some sort. Maximum loads are not generally at liftoff- they are at max Q- roughly 40,000 to 60,000 ft where the ambient pressure is 5 psia or less. This means generally that you get 10 psid of internal pressure for free which acts to alleviate axial loads. Check out how much upforce is available in a cylinder that is 27 feet in diameter as on ET @ 10 psi. It is....large. Want more? Increase pressure. Also remember liquid head effects. SO! all modern vehicle designs make extensive use of internal pressure to stabilise primary structures. Also the tank diameter directly bears on the bending stiffness of the vehicle which is essentially a beam in bending - the main force being lateral aeroloads. We have not reached any sort of limit on rocket size - but transporting these massive structures does get costly and tedious. Anything over 18 ft is a pain due to transport issues. 2) Although ramjet/scramjet systems are certainly interesting their total operational time in an efficient trajectory will be measured in small numbers of seconds and the total impulse delivered by them will be small. And they are not zero weight. They want to fly at high Q for extended periods- just what the rest of the vehicle does not want- you are intensifying loads on everything else. 3) It is important to recognize that there are really two parts to a launch trajectory. The first is all about THRUST- getting the thing off the deck with at least 1.2 to 1.8 Gs. Otherwise you are accelerating too slow and pay a big penalty. You want to get up quickly to 5-6 G acceleration and HOLD it for some time. The optimal solution for this is fairly obvious by now- it is Lo2/kerosene and/or solids. With this big thrust and modern Isp's you can heave a high Isp upper stage up to around 1/3 to 1/2 of orbital velocity. THAT stage must strike a balance between thrust, dry weight and Isp. That balance is different depending on whether you are going to LEO or to the moon. Suffice it to say that BURNOUT WEIGHT is KING. Notice that this is not just dry weight- residuals are super important. Bungle those are your performance will suck. 4) Systems for getting tourists to LEO using PRACTICAL methods at reasonable cost ( might be a bit more than a million though) have been evaluated and can be executed if someone wanted them for real. Suffice it to say that they are unlike anything yet proposed- very cool. The real key to prices normal people can afford is RATE production. And there has to be a "THERE" there. If the machine is hand built like Rutan or anyone else it is a non-starter. It needs to be bulletproof and not require engineering hand-holding. Our handholding is some of the most expensive you can buy. That is NOT so easy.
The bottom line is that with hardware that is already in production it is possible to not only achieve a launch architecture that can not only match Saturn but do it at probably 30-50% of the recurring cost. There is NOTHING magic about Saturn- it was a complex, single purpose and extraordinarily expensive way to get to orbit. That is just about all Saturn was able to do- that and a little bitty trans-lunar injection burn. Short missions and you need a Navy to recover the puny little capsule. The real solution lies in using rate-production hardware that can meet the entire architecture- from earth to LEO to lunar orbit to lunar surface and back with extensibility to Mars. This needs to be done with economy otherwise we will NEVER to Mars because it will be too expensive. Going to the moon is the litmus test of overall economic viability. If we cannot afford at least 4-8 missions to the moon in one year we are deluding ourselves about Mars.