Commercial HLV and R&D

[mmeijeri]

That is exactly what he said, see Jon's quote above in the thread.

Not trying to be difficult here, trying to make what I believe is a crucial distinction. I'll reply in more detail to Greason's point to show why it is important to make it more precise.

A “big enough” lifter for exploration can be built using the industrial base already in the U.S. for boosters shared with other customers (at ULA, SpaceX, and OSC).

Agreement, though he doesn't state which booster would be the minimum for good (not optimal) solutions. He seems to be saying that would be EELV Phase 1, and I would agree that that is big enough and highly desirable, but not that it is the smallest booster that is big enough for good solutions.

The one piece which is missing for that is political more than technical; to do it you need a large hydrocarbon booster engine in the 750,000-1,000,000 lbf class and ideally with at least 1970’s class chamber pressure 1500+ psi).

This on the other hand contradicts what ULA says about EELV Phase 1, since the Delta would have a payload of 50mT. Unless Greason believes something even bigger than EELV Phase 1 would be required.

Fairing size does matter; you can get fairings in the 7.5m class with that approach rather than 10m. I found pieces difficult to stuff in 5m fairings but none I didn’t believe could be stuffed in a 7.5m fairing. Of course, reasonable people can disagree on this in part because they have different mission models in mind.

Well, he's being reasonable and says people can disagree on this. I disagree more than 5m are necessary for good solutions. Mentioning the 6.5m would have been better.

“Big enough”, in my opinion, is a cost-optimized 25-30mT booster of which you can afford to launch a lot, and occasional capacity to go to 50-75mT for, at most, a few missions a year in a “heavy” variant. A reasonable mission model has few enough of the very big launches that maintaining all-dedicated production lines for that purpose is not as cost-effective as making a “heavy” version of a somewhat smaller booster with same engines, tank tooling, etc.

OK, he did say that, it's the bits above I disagree with. He doesn't quite state existing EELVs would not be big enough for good solutions and doesn't quite state EELV Phase 1 is the smallest launcher that would be big enough. But in light of his previous remarks it seems he believes existing EELVs would lead to marginal solutions. I disagree with that.

This may seem like an academic distinction since I'm strongly in favour of EELV Phase 1, even for many of the same reasons you mention. The crucial point is that the (paraphrased) statement ">6.5m, 40-50mT and a new hydrocarbon engine are necessary to stay away from marginal solutions" can be misconstrued as supporting even bigger HLVs or SDLVs whereas a statement like "EELV Phase 1 would be highly desirable" couldn't.

On reflection, he seems to be aiming for EELV Phase 2. If so, I'd be opposed to that. What do you think, is he arguing for Phase 1 or Phase 2?

[MikeAtkinson]

Say you take a fairly standard Mars mission, 20T landed on Mars (largest chunk), 40T transhab (including supplies). The lander + descent system does not fit into 20T and probably not into 5m, so it needs to be assembled. The transhab will need to be supplied in space with food and other dry goods. Both the lander and transhab would may have to be launched dry and so would then need water and oxygen transfer and refuelling. The EDS and Mars departure stages would need refuelling and all the components docked together and checked out in a complex dance.

Good example. I agree with 20mT as a reasonable target for the largest piece you want to be able to land in one piece. The wet mass of the lander will obviously be much more than that. Are you saying the dry mass would have to be higher too? You seemed to be implying that it could fit if dry. If so, I agree with that too.

I would disagree with two things: I believe the lander could easily fit in 5m fairings and even more easily in 6.5m fairings. And with automated rendez-vous and docking using more flights wouldn't be a problem at all.

If it turns out the maximum mass you need to land in one piece is 25T, then a 20T launch limit causes severe problems.

But that wasn't what I meant, the lander + descent systems are more than 20T. The heat shield could be the areoshell, but that is marginal to impossible at 5m. The parachutes and other landing systems (including an alternative inflatable heat shield) would need high levels of integration with the lander, but as they go over the 20T limit would need to be installed on orbit. This isn't docking which has been done hundreds of times with little fuss, but genuine external construction, which judging from the ISS experience is very expensive and difficult.

You've suggested powered descent, that almost certainly pushes up the IMLEO, but also involves a large descent stage, volume limitations with a 5m fairing (depending on fuel) make this marginal. Unless you do a cleaver design (e.g. like the ACES lunar lander) you end up with the payload a long way above the ground. I happen to think that a powered descent is the way to do it, but adding extra constraints just makes a difficult design even harder.

[mmeijeri]

If it turns out the maximum mass you need to land in one piece is 25T, then a 20T launch limit causes severe problems.

Agreed. Note that existing EELVs are capable of >=25mT, closer to 30mT even. How about 25mT as the "smallest biggest piece" for good (not marginal) solutions? Your argument would still apply if 30mT turned out to be necessary.

But that wasn't what I meant, the lander + descent systems are more than 20T. The heat shield could be the areoshell, but that is marginal to impossible at 5m.

That I can believe. You no doubt know far more about this than I do, as does Jeff Greason. I don't contend you can do this with an aeroshell. I'd be happy to drop the aeroshell if that solved the problem, even if it turned out not to be optimal, as long as it was still good, not just marginal. If it truly did turn out to be marginal, that would be reason for a rethink. And since I'm so strongly opposed to HLV for other reasons, I'm really eager to find out if it truly would be marginal.

The parachutes and other landing systems (including an alternative inflatable heat shield) would need high levels of integration with the lander, but as they go over the 20T limit would need to be installed on orbit.

This sounds very plausible.

This isn't docking which has been done hundreds of times with little fuss, but genuine external construction, which judging from the ISS experience is very expensive and difficult.

More agreement.

You've suggested powered descent, that almost certainly pushes up the IMLEO,

Perhaps, but IMLEO is not the same as cost. Another consideration is that aerobraking and aerodynamic EDL require a lot of R&D whereas propulsive EDL doesn't. But I'd attack the IMLEO by using small SEP tugs for the propellant. That would give you much higher Isp than even the NTR in the DRA 5.0.

but also involves a large descent stage, volume limitations with a 5m fairing (depending on fuel) make this marginal. Unless you do a cleaver design (e.g. like the ACES lunar lander) you end up with the payload a long way above the ground.

Shouldn't a kerolox or LOX/methane lander fit comfortably? It would likely be horizontal like the ACES lander you mention. I wouldn't call that a marginal solution. And you can use the same half a dozen or more solutions that can make a moon lander fit to make a Mars lander fit. Using the NTR transfer stage to do orbit insertion, refueling it and then using it to bring the lander to a dead stop for a vertical powered descent sounds straightforward enough.

I happen to think that a powered descent is the way to do it, but adding extra constraints just makes a difficult design even harder.

Why would it be difficult, assuming propellant transfer in Mars orbit? It looks as if doing it propulsively instead of aerodynamically saves you a lot of R&D. Not that you wouldn't want to take advantage of that where you could, but that sounds like a later embellishment.

Would you say fully propulsive (as opposed to merely strongly propulsive) descent is a marginal solution? I'd happily pay that price, especially since I'm looking for commercial propellant flights as soon as possible.

[MikeAtkinson]

Any problems need to be found and fixed in space with the clock ticking towards the launch window.

You have years to fix this, so that isn't really a problem. Also, the transhab could and probably should be reused, so you'd have to do this infrequently. Every mission would still require new propellant and consumables of course. For this to work, you'd station the hab and transfer stage at a Lagrange point.

I meant problems during the construction, fitting out and refuelling. There are lots of potential problems, launcher failures, equipment failures, delays due to various reasons, etc. They cannot be sorted out ahead of time, at best contingency plans can be made. But contingency plans and provisions themselves cost money and add to complexity.

Reusing the hab is a good idea, but has lots of problems of its own, it will require thousands of man-hours of repair, refurbishment and replenishment. A man-hour at a Lagrange point will cost several times a man-hour in LEO (a man-hour costs $400,000 or so at ISS), so it looks like we will need at least 2 orders of magnitude reduction in man-hour costs for it even to be considered. Reusing habs, makes it difficult to use a block manufacturing approach, each one will likely be different [yes there are ways round that, but each one has costs and difficulties associated with it].

[mmeijeri]

I was thinking of using Bigelow habs, assuming they'd be in use for other purposes too. Why would they require extensive maintenance?

[notsorandom]

Setting aside the debate over the necessity of HLV for a second it looks like a commercial SDHLV beats the other HLV options. It can reach IOC quicker then EELV. Both the cost for each flight and kg to LEO is significantly lower for SDHLV then an EELV HLV. Its difficult to evaluate the development cost of the Atlas Phase 2 vs J-130 because I don't know the cost to only develop the J-130. The JUS takes a significant portion of the development budget though. EELVs in the over 100 mt class cost $5.3 Billion more then the J-246. Additionally the SDHLV can be scaled up past the J-246 if needed. It looks like if an HLV 70mt or greater is needed the SDHLV is the best path.

There are a few points I want to make about the size of LV debate. Trade offs exist between development cost of launchers versus development costs of smaller payloads that have to be linked in space. The best solution from that perspective is the one where the total of the two is lowest. Another factor we have to consider is how long the system will be operational. Eventually operational cost will exceed the  development cost. This has happened in the Shuttle program but didn't happen with the Apollo program. The longer a program goes on the more important operating costs become relative to the development cost. A larger LV will have a lower per mission cost but higher development cost. Finally there is a limit on the possible flight rate. A launch pad can only be recycled so quickly after a launch. Smaller rockets will have to fly more frequently. For 3 lunar mission a year it would take three Ares Vs, six SDHLVs or at least 21 EELVs. For the EELVs that means almost one a month for both Atlas and Delta. Falcon 9 coming on-line would help this but it is still a very crowded manifest and schedule conflicts become more of a problem at The Cape.

I realize that what I am saying on the LV size debate is very qualitative, unfortunately I don't have the numbers to make it a more quantitative discussion. It would be great if we could figure out where some of those sweet spots are in development payload vs LV costs and operating vs development costs are. I suspect that they are hiding somewhere between Ares V and EELV class launchers.

[mmeijeri]

Setting aside the debate over the necessity of HLV for a second it looks like a commercial SDHLV beats the other HLV options. It can reach IOC quicker then EELV.

Not quicker than EELV Phase 1 I think.

Both the cost for each flight and kg to LEO is significantly lower for SDHLV then an EELV HLV.

Depends on the flight rate. It would have to be very high for SDLV to win.

Its difficult to evaluate the development cost of the Atlas Phase 2 vs J-130 because I don't know the cost to only develop the J-130. The JUS takes a significant portion of the development budget though.

You still need an EDS. EELV Phase 1 could do that. Phase 1 looks like the best possible commercial HLV. If you truly need more than that, SDLV starts to look attractive.

EELVs in the over 100 mt class cost $5.3 Billion more then the J-246. Additionally the SDHLV can be scaled up past the J-246 if needed. It looks like if an HLV 70mt or greater is needed the SDHLV is the best path.

That may be true. But I don't think that much would be needed.

The best solution from that perspective is the one where the total of the two is lowest.

Cost is not the only consideration, benefits are important too. One such benefit could be contributing to commercial development of space. That should be part of deliberations too.

[MikeAtkinson]

Mars missions are usually in the 600-1200T IMLEO range. Packing constraints and technology limitations due to the 20T and 5m limits probably will push that up to 700-1400T IMLEO. fuel for docking, station keeping (perhaps even orbit reboost) will push that up to 800-1600T IMLEO.

Total launch volume for near term moon and Mars missions is a better argument than fairing sizes. I do think your numbers are very high.

It doesn't matter what the total volume is, if the biggest piece that is not easily divisible won't fit on the launcher, mass, volume or diameter, then the mission won't work. Not only that but there should be a large margin for growth. We've seen what trying to fit Orion onto Ares I did for cost, schedule and capability. At this early stage and without knowing the missions (let alone the mission architectures) the only wise course is to estimate the likely parameters and then add a healthy margin. I think that most chunks will be in the 20-30T (5m) range with a few up to 50T (7m). Therefore planning for a 30T launcher with a 75T heavy option seems the sensible thing to do.

DRM 5.0 is about 750T with NTR, Augustine said 1200T for chemical (LOX/H2).

[mmeijeri]

I think that most chunks will be in the 20-30T (5m) range with a few up to 50T (7m).

What do you think could require 50mT and 7m? Anything else than aeroshells?

Therefore planning for a 30T launcher with a 75T heavy option seems the sensible thing to do.

One of my goals in this discussion is to find out if EELV Phase 2 might be needed after all. From the point of view of commercial development of space I'd be very, very reluctant to see that. Even if it turned out to be very useful for exploration, as long as the alternatives would be good, not just marginal. I would suggest that the correct course of action would be to find a robust architecture that would fit on EELV Phase 1 or even current EELVs and not to sacrifice the potential for commercial development of space to the hunt for technically optimal solutions.

DRM 5.0 is about 750T with NTR, Augustine said 1200T for chemical (LOX/H2).

Assuming NTR before you go to Mars itself might not be an unreasonable thing to do. Not strictly necessary, but perhaps wise. Chemical + SEP is good enough for Phobos and Deimos.

[MikeAtkinson]

Construction crew launch and their logistics will push that up to 1000-2000T (at least) [note only about half of that is propellant because construction and fitting out on orbit requires launches of crew and equipment. The effect of this is that there is little increase in the cheap propellant launch and a large increase in expensive crew launch]

I dispute the need for construction. Docking should suffice. And with reusable hardware costs can be amortised over multiple missions.

I would think that there should be no need for external construction for the ISS either, but inspite of its difficulty and a large number of very cleaver people trying to reduce it, there has been some. I was more talking about internal construction. Inflatable habs have all their services down the core, it is likely that they will require some internal fitting out or construction for extended BEO use.

Several things mitigate against reuse:
   1. the technology becomes obsolete
   2. production runs are low, leading to higher unit costs
   3. reuse always costs more in money, design time and mass.
   4. the design has to be tested longer to ensure that it can stand up to repeated reuse.
   5. you still need to keep the development and production teams and equipment together and available to fix problems.

[Ben the Space Brit]

Setting aside the debate over the necessity of HLV for a second it looks like a commercial SDHLV beats the other HLV options. It can reach IOC quicker then EELV.

Not quicker than EELV Phase 1 I think.

FWIW, ULA have said that it would take them about 48 months to get the ACES upper stages to flight readiness.  The DIRECT team have consistently said that their contacts at NASA have said that a core+SRB only SDLV would take about 36 months.  Take those figures as you will.

[MikeAtkinson]

If anything goes wrong with this complex operation and a earth departure window is missed, effectively 2 years of fixed cost (many billions) are added onto the project for the same return.

How so? What fixed costs?

Because you still need to keep all your capabilities in place even if they are not being used. Fixed costs are the development, production and ops teams and facilities (the portion that isn't shared with other missions). Also things like astronaut training, simulators, extensions to production contracts, management.

[notsorandom]

Setting aside the debate over the necessity of HLV for a second it looks like a commercial SDHLV beats the other HLV options. It can reach IOC quicker then EELV.

Not quicker than EELV Phase 1 I think.

ULA has stated that it will take 30 months to make the Heavy version. Some parts of Phase I could be done in parallel but 30 months is the absolute minimum. It would also take a while for propellant depots to be operational. ACES development is going to take longer then the WB Centaur. EDIT: Thanks to Ben the Space Brit, Aces takes 48 months.

Both the cost for each flight and kg to LEO is significantly lower for SDHLV then an EELV HLV.

Depends on the flight rate. It would have to be very high for SDLV to win.

Flight rate helps but a SDHLV still beats the EELVS. Take a look at this chart. http://www.launchcomplexmodels.com/Direct/charts/LV_Cost_per_kg_to_LEO.gif At a very realistic flight rate of 3 or more an SDHLV beats EELVs regardless of how many times they fly.

The best solution from that perspective is the one where the total of the two is lowest.

Cost is not the only consideration, benefits are important too. One such benefit could be contributing to commercial development of space. That should be part of deliberations too.

I would argue that an HLV launched depot would do the same. Direct's plan is to turn the JUS in to a Depot. This allows for single launch SDHVL moon missions and makes quite a bit of sense. Depots will be developed since they are so useful under either architecture. The HLV depot is also more capable since the tanks are larger. Development cost can be lowered somewhat too since the thermodynamics are better. The larger tanks have less surface area per kg of propellant. Since the architecture would cost less to operate it would allow NASA to peruse other programs with benefits to commercial space.

The problem with modifying the EELVs is that any improvement you make are not needed by the normal customers today and will increase the price to them. Furthermore, at least some of the tanker flights would have to be flown on EELVs since the Falcon 9 can't launch fast enough to meet the demand. The problem is that the EELVs just became more expensive since they are going to be upgraded to carry humans, an upgraded to phase I is going to compound that.

[mmeijeri]

Another consideration to throw into the mix: even if it turned out that you do need more than the 6.5m fairings that EELV or EELV Phase 1 could support in order to achieve good solutions, we certainly won't need that any time soon. One good thing about the EELV upgrade plan is that it can be done incrementally. Start with Phase 1 and if there's a need for Phase 2, we can always upgrade later. Phase 1 would require a new upper stage, which has major synergy with an EDS, which we'll need anyway. On the other hand, spending money now on a new first stage for the EELVs, developing new engines and making major infrastructure modifications for things we will not need for another 20 to 30 years is wasteful. The same argument can be made for sticking with current EELVs for now, we can always upgrade to a new upper stage later. Above all we need a spacecraft.

Developing those 6.5m fairings ASAP, if for no other reason than to make sure it's possible, sure. Developing EELV Phase 1, perhaps. But above all we need a spacecraft. For some reason people want to start at the bottom: new engines, then new first stages, then new upper stages and only then new spacecraft. This makes no sense to me.

I've said it before and I'll say it again: to go to the moon we need a lander. We already have launch vehicles.

[MikeAtkinson]

I'll just say one more thing about heat shields and powered landing then I'm done for the day.

Ideally you would want a heat shield to do aero-capture into Mars orbit and then reuse it for the descent. Just using it for aero-capture is marginal, especially as a fully powered descent would need more mass in Mars orbit which would require a larger larger heat shield.

Again ideally a fully powered descent lander would also be reused for ascent after ISRU refuelling. This limits fuel choices a bit but not in my opinion too badly. The descent lander would be used for both crew and cargo so as to only do a single design and get confidence that it works with cargo before trusting crew to it. Therefore it must work with 40T of cargo, doing horizontal landings when you have a 10T descent stage with 40T of cargo at one end might be difficult. Mars conditions (gravity, thin atmosphere) are also more difficult than the Moon. I don't have a good feel about whether this makes an ACES style horizontal lander a bit harder, much harder or effectively impossible.

For a vertical lander, having the cargo on top places it a long way above the ground, there may also be toppling issues unless the ground is flat with no appreciable slope. If there are no diameter limits,  on the moon you can get away with slinging the cargo underneath the rocket. On Mars this seems impractical if you want to use the lander for ascent as well.

This discussion shows how difficult (and fascinating) Mars mission design is. In my opinion the Mars mission designs I've seen have been far from ideal. Designing a new 30T launcher with 75T heavy version is probably < $4B, especially if using existing engines. In contrast an excellent overall mission architecture (for Moon, NEO and Mars) will probably save over $10B in the next twenty years over a good one and > $50B over a poor one (which is what we have now). IMHO the POR altair was poor, ACES is pretty good, DRM 5.0 is poor and there are good reference missions which integrate them with NEO access.

That is where we should be putting our effort. Lets design a 30T launcher with a 75T heavy version, in fact lets do more than one. With reasonably assured HLV access to LEO, then the other portions of the missions can be slotted into place.

I did some playing with a hypothetical TAN rocket the other day, taking a Merlin-1C F9 and morphing it into a Merlin-2 TAN 5 engine first stage. At the same lift off mass only 5 engines are required and payload is increased so that cost/kg goes down by a factor of almost 2. Increasing the engine thrust so that payload is ~30T gives another small cost/kg saving (at the same flight rate), a heavy version (at ~90T) would give another modest cost/kg saving. If the F9 heavy is getting 32,000kg to LEO for $100M, then this hypothetical F5 heavy would give a price/kg of under $2000 an may even approach $1000. The cost/kg would probably be less, but we have no idea of F9 costs only the price they charge to customers. Given these prices, a 1200T Mars mission would cost < $2.4B in launch costs (every 2 years), which seems sustainable. Obviously, similar calculations can be made for other launcher families and would I believe give similar results.

[Serafeim]

A eelv or falcon heavy lift(50-75t) how soon can be ready?
2015 possible perhaps? Not manrated maybe its better..
with that in hand everything else for beyond Leo is possible..

[mmeijeri]

At the end of the day, while it might be possible to do your all-hypergol approach, it might actually be less bad to go with slightly upgraded EELV-class vehicles and cryo depots to relax the constraints and allow a more open design space.

OK, comm check: I do not now nor have I ever advocated using an all-hypergol approach. My preferred approach would be EELV Phase 1 + fully reusable single stage hypergolic lander + cryogenic depots. If that doesn't fit the budget some deep thinking is required to find an incremental path that does fit the budget.

Can I get a roger on that?

[William Barton]

One interesting route to a larger kerolox engine would be a Merlin-1C-derived 4-chamber engine. That would be a 500klbf kerolox engine that would have few technical challenges in its implentation path. Sort o a half-step on the way to a latter-day F-1 BFE for SpaceX, and less than enormous cost, possibly something they could absorb from the presumptive profits of a working F9/Dragon. I wish RS-84 development could be resumed, but no one has suggested so far PWR would be willing (or able?) to fund this on its own, so there's a political issue. I assume same goes for TR-107 and Northrop Grumman.

[Ben the Space Brit]

A eelv or falcon heavy lift(50-75t) how soon can be ready?
2015 possible perhaps? Not manrated maybe its better..
with that in hand everything else for beyond Leo is possible..

ULA are endlessly saying 'four years' for any new development and SpaceX's announced schedules can be... optimistic.

That said, the EELV Phase 1 has only one new component: The ACES-41 upper stage.  ULA say that can be done in 4 years.  The Atlas-VH cores are all-but-done and the Delta-IV+ core with the new RS-68A engine will fly for the first time in 2011.  Human rating is an iffy issue - what does that really mean? That said, I am convinced that, with funding, the two 50t capacity EELVs could be operational by 2015 as cargo launchers and certainly by 2018 as crew launchers.  Indeed the Delta-IVM+ (5,0) and the Atlas-V-504 medium lifters could operate the lighter LEO commercial crew taxis.

Beyond that, you need the proposed 5.4m-diameter, 2 x RD-180 Atlas-V-2 core and/or the modified five-core Delta-IV.  Certainly, I'm sure ULA could have that done by 2020.  Along with the ACES-41, this can launch up to 100t - in other words, it is a direct ULA competitor to the two directly-derived SDLV heavy lifters.

However, there is a fly in this ointment.  How much funding? Remember that ULA will be sharing with people like Bigelow, SpaceX and Orbital for commercial funding from NASA.  That may lead to some delays as no one will likely get as much as they want.

That said, if NASA gets off of its launch cradle now and tells the contractors to get moving, I have no doubt that an Orion-class orbiter could be operational by 2015 and a J-24x-class lunar-capable heavy lifter by 2020.  The question really is: Would they have anywhere to go?

[clongton]

On reflection, he seems to be aiming for EELV Phase 2. If so, I'd be opposed to that. What do you think, is he arguing for Phase 1 or Phase 2?

I think he would like to see Phase 2 capability available, but designed in such a way that the launch vehicle could be de-configured to enable it to fly with Phase 1 performance capability, perhaps by removing parallel staging boosters or a different upper stage engine count for example. That would give you Phase 1 for most uses and Phase 2 when required.

What do you think could require 50mT and 7m? Anything else than aeroshells?

Surface habs need to be larger than what would fit in a 5m fairing. Even though 5m may be spacious in zero-g, where almost every cubic meter is available, the crew will now be living and working on the surface in a gravity well where 5m will be restrictive. Remember they will be living in this thing for a long time and the need for space is a well documented physiological need in humans. Even a 7m fairing may prove constrictive in surface gravity as opposed to 5m in zero-g.

One good thing about the EELV upgrade plan is that it can be done incrementally. Start with Phase 1 and if there's a need for Phase 2, we can always upgrade later.

Moving from EELV Phase 1 to Phase 2 is not just an upgrade - it's an entirely new launch vehicle development program. The Atlas V Phase 1 has a 3.8m common core while the Phase 2 is a 5m common core. That's not an upgrade; that's a new rocket. That is a design change that should be avoided because then it simply recreates the Ares debacle of having 2 different launch vehicles, 2 different manufacturing infrastructures and 2 different launch infrastructures. That just drives costs way up. It's better to find a way to do everything with a single common core design, and configure that for whatever performance is needed for any specific launch need.

That's what the DIRECT team did with the 8.4m common core that was mandated by Congress. And it was configured to fly with or without the upper stage, to provide medium or heavy lift as required, like I think Greason was alluding to in the first quote above. Remember in DIRECT v1.0, the medium lift was 45mT and the heavy lift was 90mT; exactly the range I think would have made him happy. The same philosophy could be applied to the Atlas V Phase 1 and 2 difference, but it would mean that the Phase 1 core would need to go to 5m in order to maintain the commonality.

Just as an aside, the only reason we moved from DIRECT v1.0 to v2.0 was all the criticism about the redesigned RS-68 regen engine, all of which NASA ultimately had to back away from. We *could* go back to that configuration and satisfy both Greason and the Congressional mandate to use the SSP hardware, personnel  and infrastructure. We had budgeted $1 billion for the RS-68 program. That would lower the IMLEO to be more in line with what you're looking for. Not as low as you would like, but closer.

 The bottom line is that we *must* avoid have 2 different launch vehicles, which is what your "upgrade" thought would lead to.