Are hydrogen upper stages worth it or not?

[Pipcard]

I've seen some posts like these, this one in response to someone proposing that SpaceX could put a scaled-up version of an Advanced Cryogenic Evolved Stage (a "super-ACES") on the Big Friggin' Rocket:

Why use an ACES? A hydogen upper stage would make integration hugely complex and expensive. An expendable BFR upper stage cannot be that expensive in comparison. But with refuellling in LEO I guess it would be possible to send a spacecraft into a high energy orbit and still return the upper stage. Maybe with a simple storable fuel booster that provides extra kick and/or orbit insertion at the destination.

There I go, optimizing for performance instead of cost. Habits die hard, it seems. A Refueled BFR upper stage would probably be a better option (even if expended), especially since refueling seems to be the logical architecture for MCT anyway.

There's an idea going around that hydrogen is overrated as a rocket fuel: "it's a pain to handle because of its cryogenic nature," "it makes ground handling and manufacturing more complex and expensive," "it's an optimization for performance instead of cost." So why are Blue Origin & ULA going to use BE-3s? Why are Arianespace and Mitsubishi still going to use hydrolox on their next generation rockets? Are they continuing to fall for the "siren song" of specific impulse or are they somehow making a wise decision?

[hkultala]

I've seen some posts like these, this one in response to someone proposing that SpaceX could put a scaled-up version of an Advanced Cryogenic Evolved Stage (a "super-ACES") on the Big Friggin' Rocket:

Why use an ACES? A hydogen upper stage would make integration hugely complex and expensive. An expendable BFR upper stage cannot be that expensive in comparison. But with refuellling in LEO I guess it would be possible to send a spacecraft into a high energy orbit and still return the upper stage. Maybe with a simple storable fuel booster that provides extra kick and/or orbit insertion at the destination.

There I go, optimizing for performance instead of cost. Habits die hard, it seems. A Refueled BFR upper stage would probably be a better option (even if expended), especially since refueling seems to be the logical architecture for MCT anyway.

There's an idea going around that hydrogen is overrated as a rocket fuel: "it's a pain to handle because of its cryogenic nature," "it makes ground handling and manufacturing more complex and expensive," "it's an optimization for performance instead of cost." So why are Blue Origin & ULA going to use BE-3s?

Because BE-4 is too big for a space tug engine. For LEO, they are not going to use BE-3, only BE-4s.

Why are Arianespace and Mitsubishi still going to use hydrolox on their next generation rockets? Are they continuing to fall for the "siren song" of specific impulse or are they somehow making a wise decision?

Mostly because they have only hydrolox and solid engines, they have no methane or kerosine engines.

[Pipcard]

Because BE-4 is too big for a space tug engine. For LEO, they are not going to use BE-3, only BE-4s.

Just should've asked, why are they using hydrogen if it makes integration significantly more expensive? Owlon was proposing a hydrogen third stage for BFR.

Mostly because they have only hydrolox and solid engines, they have no methane or kerosine engines.

Sure, they'd have to develop those engines. But ultimately, shouldn't they should transition into all-kerolox or all-methalox architectures to avoid the cost and complexity of hydrogen?

[R7]

Sure, they'd have to develop those engines. But ultimately they should transition into all-kerolox or all-methalox architectures. You know, to avoid the cost and complexity of hydrogen, isn't that right?

It's not that simple. France and Japan have special interest to have big solids in their LVs. It has very little to do with minimum cost and a whole lot more with maintaining strategic capability to manufacture such solids for less civilian purposes. Hydrolox is there to offset solid subpar Isp.

[jongoff]

I've seen some posts like these, this one in response to someone proposing that SpaceX could put a scaled-up version of an Advanced Cryogenic Evolved Stage (a "super-ACES") on the Big Friggin' Rocket:

Why use an ACES? A hydogen upper stage would make integration hugely complex and expensive. An expendable BFR upper stage cannot be that expensive in comparison. But with refuellling in LEO I guess it would be possible to send a spacecraft into a high energy orbit and still return the upper stage. Maybe with a simple storable fuel booster that provides extra kick and/or orbit insertion at the destination.

There I go, optimizing for performance instead of cost. Habits die hard, it seems. A Refueled BFR upper stage would probably be a better option (even if expended), especially since refueling seems to be the logical architecture for MCT anyway.

There's an idea going around that hydrogen is overrated as a rocket fuel: "it's a pain to handle because of its cryogenic nature," "it makes ground handling and manufacturing more complex and expensive," "it's an optimization for performance instead of cost." So why are Blue Origin & ULA going to use BE-3s? Why are Arianespace and Mitsubishi still going to use hydrolox on their next generation rockets? Are they continuing to fall for the "siren song" of specific impulse or are they somehow making a wise decision?

Are there real costs associated with using LH2 in an upper stage? Sure. But there are also many benefits, and it's not at all clear that the costs outweigh the benefits. Especially when you start talking about high-energy upper stages and eventually in-space refueling. When I got a tour of SLC-3E last month, I asked some of the pad guys how hard LH2 was to handle, and while they agreed it was more annoying than LOX, they didn't seem to think it was that big of a deal.

I wouldn't use it for a booster propellant (unless I was doing an SSTO or air-launched vehicle), but for an upper stage I think it isn't obvious that it is a bad choice, *even if you are designing for cost*. For ULA at least, I doubt they'd save anything by replacing Centaur with a LOX/Kero upper stage (and growing their first stage big enough to handle the much heavier upper stage). In fact, I'm almost positive it would make their system more expensive.

~Jon

[spacenut]

I think eventually metholox will win out.  Higher ISP than kerolox, but lower than hydrolox.  Temperature of liquid methane is only about 20 degrees different from liquid oxygen.  So, it would be easier to design for and handle vs hydrogen. 

Kerolox for first stage is like a solid rocket.  Lots of thrust or lift capacity, but burns out fast.  At least both will get you to the edge of space for a second stage with say metholox to take over. 

[MATTBLAK]

Yes. Methalox could probably always have a common bulkhead in the propellant tankage to reduce stage mass and volume.

[TrevorMonty]

Methlox for long term storage especially for Mars trips. Hydrolox still superior for cislunar when storage is measured in days.

Here a couple paper lunar cargo landers, roughly based on XEUS.
Hydrolox 460ISP, 3t dry + 5t cargo + 23t fuel =31t. DV 6.1km/s
Methlox 360ISP, 2t dry +5t cargo + 33t fuel =40t. DV 6.1km/s.

I was being generous with Metholox dry mass, probably more like 2.5t, tanks smaller but engines, landing gear etc would be same as Hydrolox.

[jongoff]

Yes. Methalox could probably always have a common bulkhead in the propellant tankage to reduce stage mass and volume.

But Centaur has shown that common bulkheads for LOX/LH2 aren't that hard to do once you have the bugs worked out. And Centaur-style fabrication methods result in a pretty lightweight stage that isn't that expensive to make.

~Jon

[Steven Pietrobon]

I worked out that in expendable mode, using a hydrolox upper on the Falcon IX increases GTO payload mass from 8.3 t to 10.4 t, an increase of about 25%. Whether you think that is worth the hassle is up to you! Calculations below.

http://spaceflight101.com/spacerockets/falcon-9-ft/
Falcon 9 v1.2
mp1 = 409.50 t
ms1 =  22.20 t
mp2 = 107.50 t
ms2 =   4.00 t
mpf =   1.75 t
ms = ms1+ms2+mpf = 27.95 t
mp = mp1+mp2 = 517.0 t
mc = 8.30 t
mi = mp+ms+mc = 553.25 t
mc/mi = 1.50%
mc/ms = 29.7%
(ms1+mpf)/mp1 = 5.85%

http://spaceflight101.com/spacerockets/atlas-v-401/
mp1 = 284.089 t
ms1 =  21.054 t
mis =   0.947 t
msa =   0.182 t
mp2 =  20.830 t
ms2 =   2.243 t
mpf =   2.487 t
ms = ms1+mis+msa+ms2+mpf = 26.913 t
mp = mp1+mp2 = 304.919 t
mc = 4.750 t
mi = mp+ms+mc = 336.582
mc/mi = 1.41%
mc/ms = 17.6%
(ms1+mis+msa+mpf)/mp1 = 8.68%

We can see the major reason why Falcon IX does better, its superior mass fraction for the first stage. The mass efficiency for Falcon IX first stage is 5.85% compared to 8.68% for Atlas V. This comes about from several factors with Falcon IX. The 44% greater propellant load (non-linear scaling means more propellant mass gives better efficiency), use of lighter weight Al-Li, a common bulkhead, sub-cooled propellants and lighter weight payload fairing. If we were to replace the Falcon IX upper stage with a hydrolox stage we have

Atlas V   mt2a = mc+ms2+mp2 = 4.75+2.243+20.83 = 27.823 t
First stage delta-V = g*338*ln(mi/(ms+mc+mp2)) = 6159 m/s
Second stage delta-V = g*449.7*ln(1+mp2/(mc+ms2)) = 6090 m/s
Total delta-V = 12249 m/s

Falcon IX mt2f = mc+ms2+mp2 = 8.3+4.0+107.5 = 119.8 t
First stage delta-V = g*311*ln(mi/(ms+mc+mp2)) = 4110 m/s
Second stage delta-V = g*345*ln(1+mp2/(mc+ms2)) = 7701 m/s
Total delta-V = 11811 m/s

Required delta-V for Falcon IX hydrolox second stage is dv = 12249-4110 = 8139 m/s. A higher value is used since we include additional gravity losses in the second stage. With more thrust, we can reduce the gravity losses.

dv = g*449.7*ln(mt2f/mf2)

which gives mf2 = 18.921 t. We have mp2 = mt2f-mf2 = 100.879 t. Using non-linear tank scaling model where ms = fm*mp^0.848, we have fm = 2.243/20.83^0.848 = 0.1708. Then ms2 = fm*mp2^0.848 = 8.547 t. This gives mc = mf2-ms2 = 10.374 t and a payload efficiency of mc/mi = 1.88%. This shows that using a hydrolox second stage does give a better payload fraction.

However, a better comparison in terms of performance would be mc/ms, the amount of cargo for the amount of rocket (not including propellant). We see that Falcon IX does much better, at 29.7% compared to 17.6% for Atlas V. Using hydrolox with Falcon IX, the ratio increases to 10.374/(22.2+8.547+1.75) = 31.9%. This makes using hydrolox for the second stage slightly more mass efficient, but at the expense of having a different engine and the complexity of using hydrolox.

The other case is to have Falcon IX mass efficiency in the first stage of Atlas V, where the the first stage performs a higher delta-V. We have fm = ms1/mp1^0.848 = 22.2/409.5^0.848 = 0.1353. This gives ms1 = fm1*284.089^0.848 = 16.281 t. This increases mt2 by dm = 21.054-16.281 = 4.773 t. We also add the interstage and adapter mass of mis = 0.947 t, msa = 0.182 t and the payload fairing difference mass dpf = 2.487-1.75 = 0.737 t to give the new mt2 = mt2a+dm+mis+msa+dpf = 27.823+4.773+0.947+0.182+0.737 = 34.462 t.

dv = g*449.7*ln(mt2/mf2)

where dv = 6090 m/s. This gives mf2 = 8.662 t. We have mp2 = mt2-mf2 = 25.800 t. Using non-linear tank scaling model where ms = fm*mp^0.848, we have fm = 2.243/20.83^0.848 = 0.1708. Then ms2 = fm*mp2^0.848 = 2.689 t. This gives mc = mf2-ms2 = 5.973 t and a payload efficiency of mc/mi = 1.77%, again showing the performance increase of hydrolox.

Payload to structure mass efficiency (using a SpaceX payload fairing) is 5.973/(16.281+2.689+1.75) = 28.8%. That's a lot better than the original Atlas V at 17.6%.

[Oli]

Sure, they'd have to develop those engines. But ultimately they should transition into all-kerolox or all-methalox architectures. You know, to avoid the cost and complexity of hydrogen, isn't that right?

It's not that simple. France and Japan have special interest to have big solids in their LVs. It has very little to do with minimum cost and a whole lot more with maintaining strategic capability to manufacture such solids for less civilian purposes. Hydrolox is there to offset solid subpar Isp.

The US and India use big solids as well while Russia and China do not. All of them are nuclear powers.

In terms of $ per thrust or $ per total impulse solids easily beat liquids.

[Pipcard]

Are there real costs associated with using LH2 in an upper stage? Sure. But there are also many benefits, and it's not at all clear that the costs outweigh the benefits. Especially when you start talking about high-energy upper stages and eventually in-space refueling. When I got a tour of SLC-3E last month, I asked some of the pad guys how hard LH2 was to handle, and while they agreed it was more annoying than LOX, they didn't seem to think it was that big of a deal.

I wouldn't use it for a booster propellant (unless I was doing an SSTO or air-launched vehicle), but for an upper stage I think it isn't obvious that it is a bad choice, *even if you are designing for cost*. For ULA at least, I doubt they'd save anything by replacing Centaur with a LOX/Kero upper stage (and growing their first stage big enough to handle the much heavier upper stage). In fact, I'm almost positive it would make their system more expensive.

~Jon

I thought it wasn't about mass optimization, it was about cost optimization, and that additional engine manufacturing lines and ground support equipment to handle different types of fuel, especially hydrogen (regardless of whether or not the workers were used to it), made things more expensive.

Look, it's perfectly clear hydrogen has more energy per kg than kerolox, and hence allows a lighter first stage for the same performance.   That's simple physics and not in dispute.  But hydrogen has drawbacks as well, and hence may not be the most economical choice.  It's not a good first stage fuel (not dense enough).  So now you need a two-fuel system.  This implies different engines for the different stages, more specialists on your launch team, and now your second stage engine is produced in low volume.  All of these can be solved, but it costs money.  On the whole, is the hydrolox upper stage cheaper?  Like all engineering, it's a question of tradeoffs.

Take Ed's example of a hydrolox upper state for Falcon, then reducing the first stage to 7 Merlins.  That's three less Merlins, which are rumored to cost about $1 million each.  How much does a BE-3 cost?  If it's more than 3 million you are already behind.  Even if it's less than 3 million, hydrogen might still be a losing proposition once you add in the ground infrastructure and support, amortized over missions.  And if they get re-use working, then the cost of that additional first stage mass may be smaller yet, reducing hydrogen's advantage still more. 

[pippin]

Why is that? Do you have data on this assumption?

[spacenut]

I think the reason SpaceX and Blue Origin are going to methane engines is simplicity.  Both booster and upper stage metholox.  Simplifies fueling.  Then, for reusability, methane is cleaner than kerolox.  Also, again, the liquid temperature is close to LOX.  This allows for less weight on tanks with a common bulkhead.  Then when they ran the numbers, an all methane rocket got slightly better payload capacity over a mix of fuels.  Rocket stages can be slightly smaller than hydrolox, slightly larger than kerolox.  Overall a good compromise. 

[Robotbeat]

And if you really crank flight rate and rapid reuse up to 11, then price of propellant could start to matter (it does for ITS). Liquid methane is almost as cheap as liquid oxygen.

[Pipcard]

Why is that? Do you have data on this assumption?

I wish there were, but the closest thing to data that we have is the Falcon 9 costing less per kg than an Atlas V. Some people could argue labor costs are the real main factor here, but it could also mean the extra complexity of systems to handle LH2 result in extra costs.


http://www.airspacemag.com/space/is-spacex-changing-the-rocket-equation-132285884/?no-ist

Musk says that overhead starts with how the launch vehicle is designed. The workhorse Atlas V, for example, used for everything from planetary probes to spy satellites, employs up to three kinds of rockets, each tailored to a specific phase of flight. The Russian-built RD-180 first- stage engines burn a highly refined form of kerosene called RP1. Optional solid-fuel strap-on boosters can provide additional thrust at liftoff, and a liquid hydrogen upper stage takes over in the final phase of flight. Using three kinds of rockets in the same vehicle may optimize its performance, but at a price: “To a first-order approximation, you’ve just tripled your factory costs and all your operational costs,” says Musk.

[pippin]

We don't even know whether it costs less, we only know it's being priced lower.

[Lar]

We don't even know whether it costs less, we only know it's being priced lower.

Where have I heard that argument before?

[pippin]

It's not really an argument, we just don't know an awful lot about what the cost difference would be. It depends on a thousand different factors and most of them have to do with your system design and your manufacturing processes which will vary widely between vendors.
I think it boils down to: the only difference that is really only related to the fule is the cost of the stuff itself, the ground handling, the tankage.
Engine design _might_ be more expensive or it might not because LH2 has quite a number of thermodynamic advantages over other fuels.
And then we'd have to compare this with the advantage of being able to build a smaller first stage - of which we don't really know the cost structure either...

[Oli]

Are there real costs associated with using LH2 in an upper stage? Sure. But there are also many benefits, and it's not at all clear that the costs outweigh the benefits. Especially when you start talking about high-energy upper stages and eventually in-space refueling. When I got a tour of SLC-3E last month, I asked some of the pad guys how hard LH2 was to handle, and while they agreed it was more annoying than LOX, they didn't seem to think it was that big of a deal.

I wouldn't use it for a booster propellant (unless I was doing an SSTO or air-launched vehicle), but for an upper stage I think it isn't obvious that it is a bad choice, *even if you are designing for cost*. For ULA at least, I doubt they'd save anything by replacing Centaur with a LOX/Kero upper stage (and growing their first stage big enough to handle the much heavier upper stage). In fact, I'm almost positive it would make their system more expensive.

~Jon

I thought it wasn't about mass optimization, it was about cost optimization, and that additional engine manufacturing lines and ground support equipment to handle different types of fuel, especially hydrogen (regardless of whether or not the workers were used to it), made things more expensive.

I suppose that's an important argument at current low flight rates. It gets less relevant the higher the flight rate though.

And if you really crank flight rate and rapid reuse up to 11, then price of propellant could start to matter (it does for ITS). Liquid methane is almost as cheap as liquid oxygen.

Once fuel is produced with renewable energy hydrogen should be cheaper. I admit that will take a while.

[Robotbeat]

For first stage, methane may still be preferable unless you go LOx-rich or use a tripropellant design. For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

[TrevorMonty]

For first stage, methane may still be preferable unless you go LOx-rich or use a tripropellant design. For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

Thread title is about US only, best keep booster fuel types out of the discusses.

[Oli]

For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

If we assume the methanation process is for free, and if we assume mixture ratios of 5 and 2.77 for hydrolox respectively methalox, methalox would be 80% of the cost of hydrolox per kg.

So it might just barely make sense to use hydrolox for a first stage if we look at fuel costs alone.

But this is a purely hypothetical scenario of course. Methane will be available as a fossil fuel for a long time to come and fuel costs are unlikely going to dominate launch costs for an equally long time to come.

[Pipcard]

For first stage, methane may still be preferable unless you go LOx-rich or use a tripropellant design. For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

Thread title is about US only, best keep booster fuel types out of the discusses.

The question is whether or not having commonality is more important than upper stage performance. However, hydrogen is not a good common fuel because of low density and low thrust in the first stage. So which is more economical, having a hydrocarbon first stage + LH2 second stage, or the same hydrocarbon (kerosene or methane) on both stages? And if the latter is more economical, why are Blue Origin and ULA wasting their money on hydrogen upper stages? All that we know so far is that a Falcon 9 is cheaper per kg than an Atlas V, along with anecdotal statements on this forum about hydrogen infrastructure and systems being more expensive, like this one:

LH drives up the cost. All the plumbing and valves have to be LH compatible, the valves and plumbing for Kero are cheaper.

[Robotbeat]

You can't just ignore first stages when deciding whether hydrogen is preferable for the upper stage. If you optimize the two stages completely separately, you will end up with basically two entirely different rockets, which doubles development costs and some operational ones as well. At high enough flight rates, this may not matter as much, but generally it does.

[Katana]

For first stage, methane may still be preferable unless you go LOx-rich or use a tripropellant design. For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

Thread title is about US only, best keep booster fuel types out of the discusses.

The question is whether or not having commonality is more important than upper stage performance. However, hydrogen is not a good common fuel because of low density and low thrust in the first stage. So which is more economical, having a hydrocarbon first stage + LH2 second stage, or the same hydrocarbon (kerosene or methane) on both stages? And if the latter is more economical, why are Blue Origin and ULA wasting their money on hydrogen upper stages? All that we know so far is that a Falcon 9 is cheaper per kg than an Atlas V, along with anecdotal statements on this forum about hydrogen infrastructure and systems being more expensive, like this one:

LH drives up the cost. All the plumbing and valves have to be LH compatible, the valves and plumbing for Kero are cheaper.

Centaur upperstage could go directly to GEO, while Falcon9 could only reach GTO.

[Robotbeat]

For first stage, methane may still be preferable unless you go LOx-rich or use a tripropellant design. For the first stage, it's not so much energy but mass that's needed to throw out the nozzle, and methane would be cheaper per ton than hydrogen even if both made renewably.

Thread title is about US only, best keep booster fuel types out of the discusses.

The question is whether or not having commonality is more important than upper stage performance. However, hydrogen is not a good common fuel because of low density and low thrust in the first stage. So which is more economical, having a hydrocarbon first stage + LH2 second stage, or the same hydrocarbon (kerosene or methane) on both stages? And if the latter is more economical, why are Blue Origin and ULA wasting their money on hydrogen upper stages? All that we know so far is that a Falcon 9 is cheaper per kg than an Atlas V, along with anecdotal statements on this forum about hydrogen infrastructure and systems being more expensive, like this one:

LH drives up the cost. All the plumbing and valves have to be LH compatible, the valves and plumbing for Kero are cheaper.

Centaur upperstage could go directly to GEO, while Falcon9 could only reach GTO.

1) I don't think it's true that Falcon 9 doesn't have the delta-v performance to do direct GEO. However, it doesn't have the lifetime (right now, that we know of). Falcon 9 FT is pretty dang high performance and can hit very high delta-v IF you have a very small payload...

2) Yeah, the advantage of hydrolox is best seen at very high delta-v. For LEO, it almost seems like a waste of time to try hydrogen, but for escape velocity or similar high delta-v (like your example of direct to GSO), hydrogen pulls ahead.

3) But you can get hydrogen-like performance by using more boosters or stages. This is what the Russians do. Just use a ton of stages. But if you're constrained to 2 stages, need to focus on high energy orbits, and have an unlimited budget, hydrogen looks like a very good option. So it fits DOD/three-letter-agency requirements very well.

[Steven Pietrobon]

If we assume the methanation process is for free, and if we assume mixture ratios of 5 and 2.77 for hydrolox respectively methalox, methalox would be 80% of the cost of hydrolox per kg.

Methalox oxidiser to fuel mixture ratio is more like 3.5 to 1.

[Oli]

If we assume the methanation process is for free, and if we assume mixture ratios of 5 and 2.77 for hydrolox respectively methalox, methalox would be 80% of the cost of hydrolox per kg.

Methalox oxidiser to fuel mixture ratio is more like 3.5 to 1.

I got my numbers from http://www.braeunig.us/space/propel.htm.

I know it depends on the pressure, but I'm merely interested in the mixture ratios relative to each other.

[Steven Pietrobon]

This page is more reliable.

http://web.archive.org/web/20090203154304/http://dunnspace.com/alternate_ssto_propellants.htm

[Oli]

This page is more reliable.

http://web.archive.org/web/20090203154304/http://dunnspace.com/alternate_ssto_propellants.htm

It does not say at what combustion chamber pressure, which I find a bit suspect.

Anyway, with those numbers methalox would be ~75% of the cost of hydrolox per kg, under my simplifying assumptions of free methanation and no cost difference in storage.

[Proponent]

Dunn states that he assumes "optimal" mixture ratios, except in the case of lox-hydrogen, where he assumes 6, to increase bulk density.  By "optimal," I presume he means vacuum maximum specific impulse.  But maximum specific impulse does not in general correspond to maximum performance, especially in the case of a ground-launched stage.

Attached is a plot of the specific impulse of lox-methane as a function of mixture ratio for chamber and exit pressures respectively of 1000 psi and 1 atm (the conditions assumed above in the source suggesting 2.77 as a mixture ratio).  This was generated with RPA Lite 1.2.8 with shifting equilibrium and default efficiencies.

Specific impulse peaks at ratio of about 3.2 on the ground and about 3.4 in vacuo.  You might well want to operate at higher ratios, to improve bulk density at the expense of a bit of specific impulse, but I can't see why you'd want to operate at a lower ratio.  And for an upper stage, where mass is particularly important, you're likely to want to stay pretty close to maximum specific impulse, i.e., about 3.5.

[edkyle99]

1) I don't think it's true that Falcon 9 doesn't have the delta-v performance to do direct GEO. However, it doesn't have the lifetime (right now, that we know of). Falcon 9 FT is pretty dang high performance and can hit very high delta-v IF you have a very small payload...

Correct that Falcon 9 appears on paper able to lift significant mass directly to GEO, but first the Falcon 9 second stage will have to demonstrate, or be given, a reliable second restart (third burn) capability after a 5.5 or so hour coast.  I think that this capability will be required for some EELV missions that SpaceX likely wants to win.

 - Ed Kyle

[Pipcard]

https://forum.nasaspaceflight.com/index.php?topic=35754.msg1294057#msg1294057

ULA has a lot of experience with handling liquid hydrogen and the infrastructure is already in place so for them it's difficulty may not be as big an issue as it would be for a company trying it for the first time.

Legacy infrastructure means legacy costs.

Why don't they just abandon it, then?

(I know, I've asked this before on the Vulcan thread. But "legacy costs" were not emphasized enough in that discussion.)

[Robotbeat]

Because:
1) It still works for their bread and butter high energy orbits.
2) Their primary customer (govt) doesn't like change.
3) ULA doesn't have such freedom.
4) And hydrogen really isn't a bad solution for an upper stage at all. Once you have everything else, it IS pretty high performance.

[Pipcard]

Because:
1) It still works for their bread and butter high energy orbits.
2) Their primary customer (govt) doesn't like change.
3) ULA doesn't have such freedom.
4) And hydrogen really isn't a bad solution for an upper stage at all. Once you have everything else, it IS pretty high performance.

I can understand reasons 2 and 3, but aren't they still falling for the same old "optimize for performance" trap?

[Robotbeat]

Optimizing for performance isn't always a bad idea. I mean, think about it this way:

You've built a launch pad, a nice first stage and maybe some boosters. Optimizing the upper stage will help get a big payback by making all those bits more productive.

Rocketry is exponential, so compensating for a non-optimum design by brute force can end up costing a lot more than careful attention to performance.

The Russians take a different approach and just use a bunch of stages. But that's not necessarily cheaper and can have negative reliability consequences.

I think that picking a single propellant combo makes more sense if you're vertically integrated than if you're not.

[Pipcard]

Optimizing for performance isn't always a bad idea. I mean, think about it this way:

You've built a launch pad, a nice first stage and maybe some boosters. Optimizing the upper stage will help get a big payback by making all those bits more productive.

Rocketry is exponential, so compensating for a non-optimum design by brute force can end up costing a lot more than careful attention to performance.

I thought brute force was what amortized the fixed costs of launch infrastructure and made them more productive, and what might make the MCT (now ITS) architecture more economical - by "brute forcing" the launching of propellant and disregarding IMLEO optimization.

You're not going to save money by developing 5 different rocket stages and a couple different EDL schemes at large scale over one BFS which can refuel multiple times.

Things look very different when IMLEO is no longer the primary constraint!

The Russians take a different approach and just use a bunch of stages. But that's not necessarily cheaper and can have negative reliability consequences.

And yet Blue Origin is wanting to do both, a hydrolox third stage. And despite what you've said, Russia is also developing a hydrolox third stage (KVTK), for Angara A5.

I think that picking a single propellant combo makes more sense if you're vertically integrated than if you're not.

Okay, then.

[Robotbeat]

ITS is optimized to all heck.

[TrevorMonty]

Blue covered both bases, common fuel and engine for LEO 2nd stage and light high performance hydrogen 3rd stage for BLEO missions.

The 3rd stage benefits directly from NS development. For NS a LNG engine would have been better choice as fuel costs are lot lower along with associated components. Blue made a strategic decision to develop a LH engine for BLEO applications, with NS being its first use.

[RanulfC]

Cryo-propane still isn't a consideration I suppose?

Randy

[spacenut]

When they went to the moon, it was determined that kerolox was the best first stage and hydrolox was the best for upper stages.  They also had no monitary restraints, just the need to get to the moon first. 

Today it seems that BO and SpaceX are going metholox for both stages.  Cost is the biggest factor today.  Metholox seems to be the best to solve the capability, cost, and re-usability.  Solids are not good for upper stages because they can't be shut down and restarted and weight is a factor.  Hydrogen is expensive and has a supercold storage and boiloff problem especially if loitering.  Liquid methane is about the same temperature as liquid oxygen which is needed for any upper stage being considered. 

[Pipcard]

When they went to the moon, it was determined that kerolox was the best first stage and hydrolox was the best for upper stages.  They also had no monitary restraints, just the need to get to the moon first. 

Today it seems that BO and SpaceX are going metholox for both stages.  Cost is the biggest factor today.  Metholox seems to be the best to solve the capability, cost, and re-usability.  Solids are not good for upper stages because they can't be shut down and restarted and weight is a factor.  Hydrogen is expensive and has a supercold storage and boiloff problem especially if loitering.  Liquid methane is about the same temperature as liquid oxygen which is needed for any upper stage being considered. 

You seem to have forgotten that BO is going for a hydrolox third stage, which is what has been boggling my mind.

Blue covered both bases, common fuel and engine for LEO 2nd stage and light high performance hydrogen 3rd stage for BLEO missions.

The 3rd stage benefits directly from NS development. For NS a LNG engine would have been better choice as fuel costs are lot lower along with associated components. Blue made a strategic decision to develop a LH engine for BLEO applications, with NS being its first use.

But what about guckyfan's claim that a hydrogen upper stage would be prohibitively "complex and expensive" for integration, and that even an expendable methalox upper stage would be more cost efficient for BLEO applications? (e.g. Owlon's proposed "direct-injection/fast trajectories for planetary science missions")

[RanulfC]

When they went to the moon, it was determined that kerolox was the best first stage and hydrolox was the best for upper stages.  They also had no monitary restraints, just the need to get to the moon first. 

Today it seems that BO and SpaceX are going metholox for both stages.  Cost is the biggest factor today.  Metholox seems to be the best to solve the capability, cost, and re-usability.  Solids are not good for upper stages because they can't be shut down and restarted and weight is a factor.  Hydrogen is expensive and has a supercold storage and boiloff problem especially if loitering.  Liquid methane is about the same temperature as liquid oxygen which is needed for any upper stage being considered.

Just to be historically correct, (and a bit nit-picky-ish ) kerolox as a booster was a compromise as it was already 'developed' and in use and hydrolox was in development as the "best" possible chemical propellant as long as you didn't seriously consider 'additive' chemicals or elements. (Which is to say they in fact DID consider just about everything from fluorine to boron additives to aluminum and everything short of the kitchen sink* but only pursued such in specific cases. *=Yes there were studies of suspended ceramic materials a couple of times )

In terms of both cost and performance 'at-the-time' most work agreed that the "optimum" solution was large solid booster stages with all hydrolox upper stages as per the Air Force SLS and later Saturn-II studies. However it can be noted that there were a LOT of 'assumptions' and built-in bias' that are questionable today that were not in fact questioned then.

Big solid boosters were supposed to be quite cheap to build and operate by the metrics of the day. So much so that significant money and effort was put into developing a way to throttle, shut-down and re-start them which was successfully tested but never employed because it, (as you might imagine) significantly increased the cost and complexity of the basic solid motor. In addition formulations were found that greatly decreased the burn instability and "shudder" that most SRBs had but again it costs a lot less to NOT do that if you can get away with it.

Similarly the one thing everyone pretty much agreed on was once we had hydrolox propulsion we could pretty much 'do-anything' from easy SSTO to hydrogen powered airplanes and cars with filling stations on every corner. In truth, eh, not so much. But, again in the 'thinking-at-the-time' anything LESS than hydrolox was not really worth considering seriously as hydrolox was, (obviously) THE best combination despite any engineering or operational concerns. After all fixing those concerns/issues/problems was 'only' engineering and operations work, and of course money, but we had that readily available didn't we?

It's taken decades to break away from most of those "basic" assumptions and 'common-knowledge' to the point where alternatives are becoming more common place. Methalox is currently leading the way due to a combination of users, higher awareness that there ARE alternatives and the need to find more efficient, economic, and effective ways of accessing space. As an 'on-going' process it will continue, er, going on and growing and changing as time goes on but I would not expect it to be a smooth or easy process and fully expect that todays assumptions and 'common-knowledge' will change over time, again.

Randy

[guckyfan]

But what about guckyfan's claim that a hydrogen upper stage would be prohibitively "complex and expensive" for integration, and that even an expendable methalox upper stage would be more cost efficient for BLEO applications?

I was thinking in the context of SpaceX launch operations. Horizontal integration and erection of the stack with the TEL. Can you even do that with a Centaur or ACES?

Also integrating LH into the TEL. They might have to vertically integrate that stage together with the payload, if that is possible. A huge headache and something SpaceX would not do and would cost a lot in their structure. If you design a rocket and a pad all dedicated to that LH upper stage the situation may be different. But then again only for use in cislunar space including earth departure burns to outside cislunar. A different propulsion system would be needed on arrival at the destination.

[Pipcard]

But what about guckyfan's claim that a hydrogen upper stage would be prohibitively "complex and expensive" for integration, and that even an expendable methalox upper stage would be more cost efficient for BLEO applications?

I was thinking in the context of SpaceX launch operations. Horizontal integration and erection of the stack with the TEL. Can you even do that with a Centaur or ACES?

Also integrating LH into the TEL. They might have to vertically integrate that stage together with the payload, if that is possible. A huge headache and something SpaceX would not do and would cost a lot in their structure. If you design a rocket and a pad all dedicated to that LH upper stage the situation may be different. But then again only for use in cislunar space including earth departure burns to outside cislunar. A different propulsion system would be needed on arrival at the destination.

So what you're saying is that it would be okay for the ITS booster (BFR) to have a hydrolox upper stage (super-ACES in the place of the ITS spaceship) that was supplied by a third party that wanted to develop industry on the Moon*, now that the plan has been revealed to involve vertical integration (via giant crane) and Pad 39A (which has already handled launch vehicles with hydrolox)?

*I know LCROSS apparently found carbon on the Moon, but then one poster (jongoff) claims that it's disputed based on statements by Paul Spudis. (who is an ITS skeptic)

[jongoff]

But what about guckyfan's claim that a hydrogen upper stage would be prohibitively "complex and expensive" for integration, and that even an expendable methalox upper stage would be more cost efficient for BLEO applications?

I was thinking in the context of SpaceX launch operations. Horizontal integration and erection of the stack with the TEL. Can you even do that with a Centaur or ACES?

I don't see why not. Didn't at least some of the old Atlas missile silos have the rockets stored horizontally and then tipped up for firing? I really don't see anything about Centaur or ACES that would preclude horizontal processing if you really felt it was necessary.

Also integrating LH into the TEL. They might have to vertically integrate that stage together with the payload, if that is possible. A huge headache and something SpaceX would not do and would cost a lot in their structure. If you design a rocket and a pad all dedicated to that LH upper stage the situation may be different. But then again only for use in cislunar space including earth departure burns to outside cislunar. A different propulsion system would be needed on arrival at the destination.

Why would integrating LH2 for an upper stage into a TEL be that hard? Sure, it's less dense and colder than LOX, but these are solvable problems. The Centaur LH2 umbilical hardware I got to examine at SLC-3 back in Vandenburg didn't really look that hard to deal with compared to their LOX umbilicals.

I also don't get why an LH2 upper stage is only useful in cislunar space for earth departure burns.

~Jon

[Nathan2go]

I think the hydrogen fueled rocket stage question depends on your goals; these issues are probably more important than the launch pad issues:

- If you want to refuel on Mars, then methane is the easiest fuel to make.  This works well with the SpaceX vision for Mars.  Methane is also easier to store in LEO than LH2 (LEO is much warmer than in higher orbits due to proximity to warm mother Earth).

- If you want to refuel on the Lunar poles or at Lagrangia with Lunar-made fuel, then hydrogen is the easiest fuel to make.  This works well with the Blue Origin vision for lots of people living and working in near-Earth space.   One could also argue that for trips to Lagrangia or the Moon, that stopping to refuel in LEO would add too much mission complexity and travel time, thus adding a third stage is preferred (LEO refueling of Mars ships is great as a LEO fuel depot allow utilization of the launch facility and ground crew outside of the Mars departure window).

- If you want the US government to pay you to develop an engine using the highest performing fuel, or if you want to tell your customers that your rocket builds on decades of rocket heritage, then an RL-10 derived hydrogen engine is the right answer (thus the ULA choice).  Also, hydrogen is more suitable to expendable upper stages which are paired with expendable first stages, since in that case, the upper stage is a smaller part of the whole, thus cost increases for the small upper stage have less impact to the total cost (paring an expendable upper stage with a reusable first stage makes the upper stage a bigger percentage of the total flight cost).

I don't think there is much reason to try propane as a launch fuel.  As I understand it, the performance would be about the same as methane, except that the heavier-than-air fumes could collect on the ground around the rocket, creating a fire/explosion hazard; plus it's a more expensive fuel.

On the other hand, propane could be interesting as a separate fuel for long storage periods and landing, since it's storage temp (Tboil =231 K, Tmelt=86K) is more compatible with people, without freezing in the presence of LOX (Tboil=90 K).  I suspect that a Martian plant for making methane could easily co-produce propane as 5-10% of the produced fuel, by simply tweaking the catalyst.

[Robotbeat]

For the record, the hydrogen third stage for New Glenn strikes me as a good idea. Most launches won't need it, but it should /double/the performance to high energy orbit. At least double.

[Pipcard]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

[Oli]

Methane is also easier to store in LEO than LH2 (LEO is much warmer than in higher orbits due to proximity to warm mother Earth).

If you use subcooled methalox you almost certainly need cryocoolers on both depot and departure stages though.

[TrevorMonty]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

I'd surprised if 3rd stage doesn't sharing a lot of NS components, maybe even tanks.

[Pipcard]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

I'd surprised if 3rd stage doesn't sharing a lot of NS components, maybe even tanks.

The New Glenn third stage is apparently depicted as having the same diameter as the rest of the rocket (7 m). But yes, the BE-3 is going to be used for both.

[hkultala]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

I'd surprised if 3rd stage doesn't sharing a lot of NS components, maybe even tanks.

The New Glenn third stage is apparently depicted as having the same diameter as the rest of the rocket (7 m). But yes, the BE-3 is going to be used for both.

both what?


Same diameter for methalox 2nd stage and hydrolox third stage makes a lot of sense as methalox is much more dense than hydrolox, but upper stages should always be much lighter, so the tank sizes do not differ very much.

Just like Saturn V had same diameter in it's first(kerosine) and second(hydralox) stages.

[Pipcard]

I meant both the New Glenn upper stage and New Shepard, as mentioned by TrevorMonty; sorry for the confusing language.

[Robotbeat]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

Not necessarily. Think of the hydrogen upper stage as their "heavy." And the hydrogen tech will be very well proven from suborbital flights.

Most Atlas V flights don't use solids, but the tech is still well amortized. (Though pad requirements are easier with solids, of course.)

[Jim]

Just like Saturn V had same diameter in it's first(kerosine) and second(hydralox) stages.

Not a relevant comparison.  The relative size of the LOX tank to the fuel tank was the opposite in both stages.

[notsorandom]

For the record, the hydrogen third stage for New Glenn strikes me as a good idea. Most launches won't need it, but it should /double/the performance to high energy orbit. At least double.

I know you are aware that most launches these days are to high energy orbits. So I'm curious to hear thoughts on why you think most launches will not need the 3rd stage. Do you think that Blue will be doing a different mix of mission types or do you think that the 2 stage NG will still have the performance to do those missions economically?

[Nathan2go]

If you use subcooled methalox you almost certainly need cryocoolers on both depot and departure stages though.

For a departure stage which lingers in LEO for only a few hours, certainly no coolers would be needed for reasonably low boil-off.  And boil-off can be made to happen at any desired temperature between Tfreeze and Tcritical, by adjusting the tank pressure: lower pressure means lower temperature boiling (on Earth, tank pressure can't go below 1 bar to avoid tank collapse, which is why subcooling creates temperature gradients in the rocket, but this is not an issue in vacuum conditions).  Just remember to bring the tanks to flight pressure a few seconds before starting the engines.

For storage of many weeks at a LEO depot, yes maybe cro-coolers are needed.  But holding 95k for methane and 60k for LOX is 3x easier than 20k for LH2.  Also, the Earth departure burn requires only 60% as much delta-V as the trans-Earth injection from Mars surface, and 78% of the delta-V for the second stage to reach orbit; so deep subcooling is not as critical.  So maybe the depot holds the propellants at more like 95/80K (4x easier than LH2).

[Nathan2go]

For the record, the hydrogen third stage for New Glenn strikes me as a good idea. Most launches won't need it, but it should /double/the performance to high energy orbit. At least double.

What?  Were you thinking of hydrolox vs solids (like the cancelled Shuttle-Centaur vs IUS)?  Methalox and hydrolox are much closer in performance.

The rocket equation for mass ratio is exponential in Delta-V/Isp, so it grows faster as the Delta-V gets bigger.  The Isp of hydolox should be 460 s, versus 380 s for methalox, or 21% better.  For low mass ratio/low delta-V burns (eg. LEO to GTO, or LEO to Vescape), all you get is about a 21% benefit. 

For two stages, the delta-V (and mass ratio) of stage 2 must be higher, thus, you'd get a bigger benefit for hydrolox: maybe 40% to Vescape.  But with three stages, each stage contributes less performance, thus there's less benefit to upgrading.

[Robotbeat]

For the record, the hydrogen third stage for New Glenn strikes me as a good idea. Most launches won't need it, but it should /double/the performance to high energy orbit. At least double.

I know you are aware that most launches these days are to high energy orbits. So I'm curious to hear thoughts on why you think most launches will not need the 3rd stage. Do you think that Blue will be doing a different mix of mission types or do you think that the 2 stage NG will still have the performance to do those missions economically?

New Glenn won't be launching just what we launch today. New Glenn will be launching tourists to LEO and is already being marketed for that purpose. And perhaps LEO constellations will take market share from GSO. But anyway, the main reason is that not all payloads will need all 3 stages to get to a high energy orbit. Two stages is enough. It's a big launch vehicle.

[TrevorMonty]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

I'd surprised if 3rd stage doesn't sharing a lot of NS components, maybe even tanks.

The New Glenn third stage is apparently depicted as having the same diameter as the rest of the rocket (7 m). But yes, the BE-3 is going to be used for both.

Can't locate picture with NG in it but it showed two versions. The 2 stage is same diameter as 1st (7m) but 3rd stage is considerably smaller.

[Pipcard]

Most launches won't need it

Doesn't that mean that it doesn't get amortized enough (low flight rates)?

I'd surprised if 3rd stage doesn't sharing a lot of NS components, maybe even tanks.

The New Glenn third stage is apparently depicted as having the same diameter as the rest of the rocket (7 m). But yes, the BE-3 is going to be used for both.

Can't locate picture with NG in it but it showed two versions. The 2 stage is same diameter as 1st (7m) but 3rd stage is considerably smaller.

The fairing has a smaller diameter for the two-stage version, the third stage in its respective version has a 7 m diameter.

[guckyfan]

I also don't get why an LH2 upper stage is only useful in cislunar space for earth departure burns.

~Jon

What else could it do? It can't brake near Neptun or Uranus. For that you design a long living stage with hypergols or methalox.

[jongoff]

What else could it do? It can't brake near Neptun or Uranus. For that you design a long living stage with hypergols or methalox.

There's more to the solar system than TLI injection and Neptune/Uranus. Once you get away from planetary bodies, passive thermal control alone can suppress LOX boiloff entirely, and get LH2 boiloff down to levels that work for everything this side of the main belt. Active cooling also becomes a lot easier when the amount of heat you have to reject is modest.

But seriously, for the foreseeable future, we're talking Moon and maybe Mars and/or the asteroids. And LOX/LH2 works fine for enough of that system to be worth serious consideration. I probably would go LOX/CH4 for mars landers, but I also wouldn't go with the SpaceX architecture where your launch vehicle upper stage, interplanetary stage, and mars lander are all the same stage. But that's me.

~Jon

[Pipcard]

but I also wouldn't go with the SpaceX architecture where your launch vehicle upper stage, interplanetary stage, and mars lander are all the same stage. But that's me.

But I was told that it would save on development costs, and that the requirements "overlap" anyway?

[Robotbeat]

They do overlap if your upper stage is VTVL reusable anyway.

[oldAtlas_Eguy]

For the near and even farther into the future on-way outer solar system the use of hydrolox for the US to get fast transits for low cost is a definite positive. But once you start talking re-usability and inner solar system then the systems are more specific to their task and the planetary bodies they service. For airless bodies that have water/ice then a hydrolox system does make some sense. But for Mars it does not.

[Patchouli]

While avoiding that hydrogen did allow Spacex to get flying sooner it also forced some design decisions such as densified propellants that later came to haunt them when they decided to upgrade their vehicle to EELV class payloads.

Though the need for stuff like precooled lox also probably could have been eliminated by simply adding a third stage for missions to high energy orbits.

[jongoff]

but I also wouldn't go with the SpaceX architecture where your launch vehicle upper stage, interplanetary stage, and mars lander are all the same stage. But that's me.

But I was told that it would save on development costs, and that the requirements "overlap" anyway?

That's the theory, but generally trying to make a hyper-optimized, bleeding edge jack of all trades, doesn't always work out as well as people expected. And now the requirements don't necessarily overlap.

~Jon

[jongoff]

While avoiding that hydrogen did allow Spacex to get flying sooner it also forced some design decisions such as densified propellants that later came to haunt them when they decided to upgrade their vehicle to EELV class payloads.

This. For all the talk of how LOX/LH2 is so much harder and more expensive to deal with, ULA had had only a few minor issues with Centaur (knock on wood), and SpaceX has had a ton of challenges with densified propellants, failures due to having to bury their pressurant bottles inside the tanks, etc. Sometimes you end up with more complexity in your attempt to avoid other complexity.

~Jon

[Robotbeat]

Perhaps that's because RL-10 on Centaur and Atlas has half a century of heritage. Its first flights had problems (only ONE of the first five flights actually fully succeeded, and most of those failures were problems with Centaur). I bet if highly densified propellants had half a century of heritage, they'd be basically problem-free, too.

Concern-trolling because SpaceX advances some new technology and it's not perfect out of the gate.

[Oli]

Perhaps that's because RL-10 on Centaur and Atlas has half a century of heritage. Its first flights had problems (only ONE of the first five flights actually fully succeeded, and most of those failures were problems with Centaur). I bet if highly densified propellants had half a century of heritage, they'd be basically problem-free, too.

Concern-trolling because SpaceX advances some new technology and it's not perfect out of the gate.

Stainless steel common bulkhead balloon tanks haven't exactly taken the world by storm.

[savuporo]

If you just count how many earth departure burns LOX/LH2 upper stages have delivered over decades, i think it's clear its worth it.

[sdsds]

I very much admire how F9 has gotten so much performance without hydrogen. (Just a few years ago I would have described as "essentially impossible" what now looks to be almost routine.) The payoff from the techniques providing that performance will be worth the bumps they introduce along the way!

But we know there are also techniques (e.g. IVF) which could improve the capabilities of hydrogen stages too. I acknowledge my perspective is shaded by being a Moon First thinker. But a stage with a lifetime so long it could provide propulsion for lunar orbit insertion? That kind of hydrogen upper stage is definitely "worth it!"

Oh. And providing that would be "essentially impossible" for the F9 upper stage.

[pippin]

Perhaps that's because RL-10 on Centaur and Atlas has half a century of heritage. Its first flights had problems (only ONE of the first five flights actually fully succeeded, and most of those failures were problems with Centaur). I bet if highly densified propellants had half a century of heritage, they'd be basically problem-free, too.

Concern-trolling because SpaceX advances some new technology and it's not perfect out of the gate.

Well, the track record for the first Atlas flights with kerolox wasn't exactly better in those days so I don't think attributing that to hydrolox isn't exactly fair, that's how first flights were in these days.

And densified propellants do have half a century of flight history, just not in the US

[TrevorMonty]

Both Methane and Hydrogen upper stages can use IVF to elimate He (SpaceX curse) and hydrazine. Both fuels would allow for long lived stages capable of multiple refuelling from ISRU fuel. This thread is more about Methane vs Hydrogen than RP1.

For stages that use aerobraking to reenter and land then Methane is superior as stage will be smaller and therefore lighter than Hydrogen. Large light weight stainless tanks are great in vacuum but don't work to well when they have to be integrated into a airframe with heat sheilding.

For lunar missions hydrogen gives better performance plus long term there is option of ISRU fuel. Mars missions long term have option of either ISRU fuel but aerobraking favours Methane.

[Pipcard]

For stages that use aerobraking to reenter and land then Methane is superior as stage will be smaller and therefore lighter than Hydrogen. Large light weight stainless tanks are great in vacuum but don't work to well when they have to be integrated into a airframe with heat sheilding.

For lunar missions hydrogen gives better performance plus long term there is option of ISRU fuel. Mars missions long term have option of either ISRU fuel but aerobraking favours Methane.

The question is whether it is better to spend money on developing and having an extra production line for an optimized lunar lander, or use the same methalox Mars vehicle for your lunar missions, even if only the LOX is available for ISRU (but it would be great if the carbon seemingly found by LCROSS was verified). I guess it would depend on the scale of the operations.

________________

some more food for thought

You don't make a a single engine and shoehorn a way oversized vacuum version of it onto a second stage that's made with identical tooling to  your first stage (when you could lighten it by using other tooling) if you're not driven by producibility/commonality over performance.  You don't use a kerolox upper stage that gets neither long coasts of hypergols nor the performance of hydrolox unless you're not driven by optimization.  The mantra is three fold and clear as day: lower cost, lower cost and lower cost.

[jongoff]

For stages that use aerobraking to reenter and land then Methane is superior as stage will be smaller and therefore lighter than Hydrogen. Large light weight stainless tanks are great in vacuum but don't work to well when they have to be integrated into a airframe with heat sheilding.

For lunar missions hydrogen gives better performance plus long term there is option of ISRU fuel. Mars missions long term have option of either ISRU fuel but aerobraking favours Methane.

Yes and no. If MSNW's magnetoshell aerocapture technology pans out (I'm bullish but you probably already knew that), hydrogen could actually have an advantage over methane for aerocapture/aerobraking, and maybe even aeroentry.

~Jon

[jongoff]

For stages that use aerobraking to reenter and land then Methane is superior as stage will be smaller and therefore lighter than Hydrogen. Large light weight stainless tanks are great in vacuum but don't work to well when they have to be integrated into a airframe with heat sheilding.

For lunar missions hydrogen gives better performance plus long term there is option of ISRU fuel. Mars missions long term have option of either ISRU fuel but aerobraking favours Methane.

The question is whether it is better to spend money on developing and having an extra production line for an optimized lunar lander, or use the same methalox Mars vehicle for your lunar missions, even if only the LOX is available for ISRU (but it would be great if the carbon seemingly found by LCROSS was verified). I guess it would depend on the scale of the operations.

I'm pretty confident the answer is an optimized lunar lander. The amount of benefit you get by being able to do full ISRU refueling instead of having to still carry your own fuel is huge. LOX ISRU helps, but full ISRU shines especially with reusable vehicles. And the benefit of full ISRU is added to the top of the benefit from using LH2 vs Methane in the first place.

Both ULA and Blue Origin will have LOX/LH2 upper staged vehicles flying, so developing a kit to enable lunar landings for one of those stages shouldn't break the bank. It might be harder for SpaceX to compete for lunar missions, but that's why it's good to have an industry with multiple providers taking multiple approaches.

~Jon

[Lars-J]

While avoiding that hydrogen did allow Spacex to get flying sooner it also forced some design decisions such as densified propellants that later came to haunt them when they decided to upgrade their vehicle to EELV class payloads.

This. For all the talk of how LOX/LH2 is so much harder and more expensive to deal with, ULA had had only a few minor issues with Centaur (knock on wood), and SpaceX has had a ton of challenges with densified propellants, failures due to having to bury their pressurant bottles inside the tanks, etc. Sometimes you end up with more complexity in your attempt to avoid other complexity.

~Jon

Let's not pretend that ULA developed Centaur from scratch here... Because that would be very much inaccurate, since the first Centaur variant flew over 50 years ago. If Centaur is available, that is one thing. But if you are working on a clean sheet design, that changes things.

[Pipcard]

For stages that use aerobraking to reenter and land then Methane is superior as stage will be smaller and therefore lighter than Hydrogen. Large light weight stainless tanks are great in vacuum but don't work to well when they have to be integrated into a airframe with heat sheilding.

For lunar missions hydrogen gives better performance plus long term there is option of ISRU fuel. Mars missions long term have option of either ISRU fuel but aerobraking favours Methane.

The question is whether it is better to spend money on developing and having an extra production line for an optimized lunar lander, or use the same methalox Mars vehicle for your lunar missions, even if only the LOX is available for ISRU (but it would be great if the carbon seemingly found by LCROSS was verified). I guess it would depend on the scale of the operations.

I'm pretty confident the answer is an optimized lunar lander. The amount of benefit you get by being able to do full ISRU refueling instead of having to still carry your own fuel is huge. LOX ISRU helps, but full ISRU shines especially with reusable vehicles. And the benefit of full ISRU is added to the top of the benefit from using LH2 vs Methane in the first place.

Both ULA and Blue Origin will have LOX/LH2 upper staged vehicles flying, so developing a kit to enable lunar landings for one of those stages shouldn't break the bank. It might be harder for SpaceX to compete for lunar missions, but that's why it's good to have an industry with multiple providers taking multiple approaches.

~Jon

I like where this thread is going. So this is the real dilemma for the future of space development if it gets to the point of colonization and/or large scale industrialization: is it worth it to have methalox ISRU only, hydrolox ISRU only (if zero boil-off systems are practical), or both?

[Robotbeat]

With hundreds of thousands of people living in space, there's plenty of room for both. Doesn't mean that one might not become dominant, but if they still trade as closely as they do today, there's enough room in such a market for both solutions.

[Pipcard]

With hundreds of thousands of people living in space, there's plenty of room for both. Doesn't mean that one might not become dominant, but if they still trade as closely as they do today, there's enough room in such a market for both solutions.

What about the near-term future? (including the decade following an ITS Mars landing, assuming that the program is successful)

[Robotbeat]

With hundreds of thousands of people living in space, there's plenty of room for both. Doesn't mean that one might not become dominant, but if they still trade as closely as they do today, there's enough room in such a market for both solutions.

What about the near-term future? (including the decade following an ITS Mars landing, assuming that the program is successful)

Assuming ITS is fully successful, almost nothing even on the drawing board could compete with it. Possibly something Bezos is working on (Blue Origin uses both methane and hydrogen) or, on the smaller end, an evolution of Masten Space Systems' XS-1, which uses methane (and, I believe, an ITS-like launch cradle... It was actually Masten Space that pioneered that idea before SpaceX got interested in it).

[Pipcard]

Assuming ITS is fully successful, almost nothing even on the drawing board could compete with it.

Even for lunar missions, in the case that the LCROSS data showing carbon on the Moon were erroneous?

[Robotbeat]

Assuming ITS is fully successful, almost nothing even on the drawing board could compete with it.

Even for lunar missions, in case the LCROSS data showing carbon on the Moon was erroneous?

Easily. A fueled up ITS can go to the Moon and back with a huge amount of payload. No refueling required.

People underestimate the cost of getting lunar ISRU up and running. The environment is much harsher than Mars and the water much rarer. And it won't be exactly easy on Mars, either. (Also, mining water for export from the Moon will be easily out competed by $9/kg in LEO by at least one variant of ITS.) Even a space elevator couldn't compete.

This is all /assuming/ ITS is fully successful. I /don't/ think that is currently most likely to happen.

[jongoff]

I like where this thread is going. So this is the real dilemma for the future of space development if it gets to the point of colonization and/or large scale industrialization: is it worth it to have methalox ISRU only, hydrolox ISRU only (if zero boil-off systems are practical), or both?

Let's not pretend that ULA developed Centaur from scratch here... Because that would be very much inaccurate, since the first Centaur variant flew over 50 years ago. If Centaur is available, that is one thing. But if you are working on a clean sheet design, that changes things.[/quote]

I'm torn to be honest. For the Moon, my guess is it'll mostly make sense to go with LOX/LH2. I'm just skeptical that there really are enough hydrocarbons relative to water to enable LOX/Methane on the moon. And a full-ISRU system will always beat an oxidizer only ISRU system. Any relative higher cost of LH2 launch infrastructure and propulsion hardware is swamped IMO by this consideration. At least for cislunar space once ISRU is operational.

On Mars... I can see good arguments for either LOX/Methane or for "both". For landing and ascent, LOX/Methane probably wins. But for Earth Return, LOX/LH2 might not be out of the question, if your earth return vehicle uses LOX/LH2. I guess I just don't think that SpaceX will hold a monopoly on in-space transportation, and if you can make LOX/CH4, you can also make smaller quantities of LOX/LH2. By definition.

~Jon

[jongoff]

Easily. A fueled up ITS can go to the Moon and back with a huge amount of payload. No refueling required.

I strongly disagree. Even if ITS really has the magical price/kg people believe, a similarly reusable system that gases up on the moon is always going to be cheaper than hauling propellant up and down the gravity well. The delta-V from LEO to the lunar surface and back (assuming aerobraking at earth) is about 9km/s. With LOX/Methane that's a lot of propellant per unit payload compared to something where you can refuel along the way.

People underestimate the cost of getting lunar ISRU up and running. The environment is much harsher than Mars and the water much rarer. And it won't be exactly easy on Mars, either. (Also, mining water for export from the Moon will be easily out competed by $9/kg in LEO by at least one variant of ITS.) Even a space elevator couldn't compete.

First off, most of the cost of setting up ISRU is driven by the high cost of transportation to and from the Moon. In a world where that's dropping because of RLVs, the cost of setting up ISRU will plummet too. And I don't buy for a second that Elon has an approach that'll get anywhere near $9/kg in LEO. Yes, if you assume magic on your side and no magic on the other side, the magic side is always going to look better.

This is all /assuming/ ITS is fully successful. I /don't/ think that is currently most likely to happen.

And even if it is "successful" does that mean it'll be anywhere near $9/kg in LEO successful?

~Jon

[notsorandom]

Let's not pretend that ULA developed Centaur from scratch here... Because that would be very much inaccurate, since the first Centaur variant flew over 50 years ago. If Centaur is available, that is one thing. But if you are working on a clean sheet design, that changes things.

Historically hydrogen has been a bit harder to tame. The Centaur and other LH2 stages around the world had plenty of early difficulties and it is harder to design an LH2 engine. However Blue Origin went ahead and developed the BE-3 and New Shepard under a commercial, non-government sponsored program. The system so far has a good if not short flight record.

The Centaur may be a legacy from a time when these things were more difficult. Blue has shown that a company can develop and operate a LH2 rocket economically with a stable budget for a fraction of the cost of the Centaur. Its still harder to develop and it might not be a good idea for a company with a more constrained budget but it doesn't seem to be an impossible thing anymore.

[TrevorMonty]

Blue didn't quite start from scratch they hired a lot of people with LH2 knowledge. In a way it works out better for them as clean sheet design can be more efficient than modifying legacy systems.

[Robotbeat]

Let's not pretend that ULA developed Centaur from scratch here... Because that would be very much inaccurate, since the first Centaur variant flew over 50 years ago. If Centaur is available, that is one thing. But if you are working on a clean sheet design, that changes things.

Historically hydrogen has been a bit harder to tame. The Centaur and other LH2 stages around the world had plenty of early difficulties and it is harder to design an LH2 engine. However Blue Origin went ahead and developed the BE-3 and New Shepard under a commercial, non-government sponsored program....

...That is actually an urban legend. BE-3 was actually a paid milestone under CCDeV. They directly got NASA funds for it. Additionally, they got a lot of free access to NASA test facilities and/or personnel as part of an unfunded SAA like the one SpaceX is using for Red Dragon.

Here's one news article on it (it'd be nice if someone found the actual contract showing it's a paid milestone: https://www.flightglobal.com/news/articles/blue-origin-completes-full-power-tests-on-thruster-for-orbital-vehicle-377721/ )

So it most certainly was done with government funds helping. Don't be surprised that Blue Origin/Bezos doesn't advertise this much. It's better PR if it seems like it's totally their own doing.

(This is secondary to your main point that hydrogen doesn't have to necessarily be extremely expensive.)

Just like SpaceX's COTS and CC, this is a good example of efficient use of government funds. The contract was competitively awarded, and the company contributed to development as well. This is EXACTLY what we want: government funds helping a transformative nascent industry get off the ground that otherwise would've taken longer or wouldn't have been able to make progress. And being competitively awarded is key to this.

[Pipcard]

Are zero-boil off systems prohibitively complex for a hydrolox-based architecture?

Also, in the SpaceX ITS presentation, hydrogen was deemed a bad choice for in-space refueling. So why is ULA planning distributed launch with hydrolox refueling for ACES?

[Steven Pietrobon]

Also, in the SpaceX ITS presentation, hydrogen was deemed a bad choice for in-space refueling. So why is ULA planning distributed launch with hydrolox refueling for ACES?

Its relatively easy for ULA to do hydrolox propellant transfer as they have a good idea on how to do it, based on the experiments they've done on the ground. For SpaceX its probably in the "too hard" basket.

[jongoff]

Are zero-boil off systems prohibitively complex for a hydrolox-based architecture?

I wouldn't say ZBO LOX/LH2 systems are prohibitively complex, but they are complex enough that they're often not worth the hassle compared to making sure your system has a very good passive thermal design (like what ULA is trying to do with ACES). In most cases the passive thermal system is "good enough", and in the cases where it isn't, having a good passive thermal design makes the active thermal systems easier to design, since you've minimized the heat leak into the LH2 tanks that you're trying to actively remove.

FWIW, by passive thermal design I mean things like good multi-layer MLI or sun-shield insulation, minimizing penetrations into the tanks, being creative about how you mount stuff that needs to be warm relative to the tanks, using tanks with low CTE materials, vapor cooled skirts, etc.

From the ULA papers I've seen, some key considerations are:
1- Location of the stage: if you're in low orbit around a planet, that's a big warm body taking up half the sky, so that's a more challenging thermal environment than say L2 or deep space. IIRC they were saying boiloff rates at EML2 would be 1/10th what they are in LEO.
2- Duration: if you're talking missions measured in days or weeks, passive is clearly the way to go, but as you start getting into the months/years range, active cooling starts looking better.

Also, in the SpaceX ITS presentation, hydrogen was deemed a bad choice for in-space refueling. So why is ULA planning distributed launch with hydrolox refueling for ACES?

The ITS presentation was SpaceX's perspective on the problem, not some independent statement of objective truth. As others have pointed out ULA and others have a lot more experience with LH2. So some of it may just be due to them being more comfortable with other fluids. Some of it stems from being Mars focused, where the easy availability of CO2 from the atmosphere changes the equation. Some of it can be just smart people coming to different conclusions--note that throughout the space age, really brilliant people don't always come to the same conclusion, even for the same problem. I almost get more worried when I see everyone agreeing... :-)

I'm pretty sure that with ULA's decades of experience (including pre-ULA experience) in LH2, and their economic situation, that LH2 really is the best upper stage propulsion choice for them as a company.

~Jon

[jongoff]

Also, in the SpaceX ITS presentation, hydrogen was deemed a bad choice for in-space refueling. So why is ULA planning distributed launch with hydrolox refueling for ACES?

Its relatively easy for ULA to do hydrolox propellant transfer as they have a good idea on how to do it, based on the experiments they've done on the ground. For SpaceX its probably in the "too hard" basket.

They've also done a lot of post-mission in-space experiments, like that DMSP launch from a few years back. SpaceX wasn't the first company to realize you could do useful experiments after the main customer is away, at least on mission with spare performance (that DMSP one left over 11000lb of prop in the tanks after spacecraft separation, so it was an unusually good one).

As for the in-space transfer part (the one part they haven't done), LH2 isn't really that much harder in space than transfering any other propellant. It's bulky and cold, but LOX and Methane are already cold enough that you have to use cryogenic seals. It's just a matter of picking the right seal, selecting the right coupling materials, and the right surface treatments and finishes. Altius has some recent SBIR work related to this.

~Jon

[TrevorMonty]

Just to add to Jon's point. ULA distributed launch paper gives a 30 day wait in LEO (180km) of 30t drop tank while it waits for LV it is refuelling. The wait assumes a certain amount of boil off (0.9t LH) and doesn't use sunscreens. Move that same tanker to L2 with sunscreen and months for same boil off rate become possible.

A quickly deployable small LV like XS1, used as topup tanker (1-2t LH) could  extend stay of main drop tanks to months for an additional cost. Good insurance to have just in case main payload LV is delayed.

[Jim Davis]

A quickly deplorable small LV like XS1...

Deplorable? Ah, autocomplete!

[notsorandom]

Let's not pretend that ULA developed Centaur from scratch here... Because that would be very much inaccurate, since the first Centaur variant flew over 50 years ago. If Centaur is available, that is one thing. But if you are working on a clean sheet design, that changes things.

Historically hydrogen has been a bit harder to tame. The Centaur and other LH2 stages around the world had plenty of early difficulties and it is harder to design an LH2 engine. However Blue Origin went ahead and developed the BE-3 and New Shepard under a commercial, non-government sponsored program....

...That is actually an urban legend. BE-3 was actually a paid milestone under CCDeV. They directly got NASA funds for it. Additionally, they got a lot of free access to NASA test facilities and/or personnel as part of an unfunded SAA like the one SpaceX is using for Red Dragon.

Here's one news article on it (it'd be nice if someone found the actual contract showing it's a paid milestone: https://www.flightglobal.com/news/articles/blue-origin-completes-full-power-tests-on-thruster-for-orbital-vehicle-377721/ )

So it most certainly was done with government funds helping. Don't be surprised that Blue Origin/Bezos doesn't advertise this much. It's better PR if it seems like it's totally their own doing.

(This is secondary to your main point that hydrogen doesn't have to necessarily be extremely expensive.)

Just like SpaceX's COTS and CC, this is a good example of efficient use of government funds. The contract was competitively awarded, and the company contributed to development as well. This is EXACTLY what we want: government funds helping a transformative nascent industry get off the ground that otherwise would've taken longer or wouldn't have been able to make progress. And being competitively awarded is key to this.

Should have been more specific. Pretty much any engine developed by a commercial company has been directly funded by government money or can trace a substantial heritage to a previous government program. The difference I was trying to emphasize was between the BE-3 and engines like the RL-10 and J-2X. Those being the result of a government run program.

[john smith 19]

As for the in-space transfer part (the one part they haven't done), LH2 isn't really that much harder in space than transfering any other propellant. It's bulky and cold, but LOX and Methane are already cold enough that you have to use cryogenic seals. It's just a matter of picking the right seal, selecting the right coupling materials, and the right surface treatments and finishes. Altius has some recent SBIR work related to this.

IIRC ULA also mentioned that a settling thrust of about 10 x 10^-6g is enough to cut boil off by 50%. One of the reasons for their interest in IVF.

There is also the NAIC funded work on a surface coating  blocking solar absorption and allowing a tank to continue emitting heat down to an equilibrium temperature of about 47K, significantly cutting the heat load you'd need to expel long term.

But the joker in the pack remains no demonstration of LH2 (or AFAIK any cryogen) transfer on orbit. 

[jongoff]

Just to add to Jon's point. ULA distributed launch paper gives a 30 day wait in LEO (180km) of 30t drop tank while it waits for LV it is refuelling. The wait assumes a certain amount of boil off (0.9t LH) and doesn't use sunscreens. Move that same tanker to L2 with sunscreen and months for same boil off rate become possible.

A quickly deployable small LV like XS1, used as topup tanker (1-2t LH) could  extend stay of main drop tanks to months for an additional cost. Good insurance to have just in case main payload LV is delayed.

It's an outside chance that Masten will get the XS-1 Phase II nod, but XS-1 and Vulcan/ACES would make a really good team...

~Jon

[jongoff]

But the joker in the pack remains no demonstration of LH2 (or AFAIK any cryogen) transfer on orbit. 

I'm not too worried about LH2 transfer on orbit. It needs to be flight demonstrated, but I don't think it's that high of a risk. If we get our SBIR Phase II, we'll hopefully have hardware that could be flight demo'd by the end of the 2yrs.

~Jon

[john smith 19]

I'm not too worried about LH2 transfer on orbit. It needs to be flight demonstrated, but I don't think it's that high of a risk. If we get our SBIR Phase II, we'll hopefully have hardware that could be flight demo'd by the end of the 2yrs.

Is that SBIR to actually do such a transfer?

This would be excellent news and long overdue.

[Vultur]

I think people ought to look at propane again... from what I've read, the Isp is only a few seconds worse than methane and the density is much better. It might be the ideal SSTO fuel.

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

[Katana]

I think people ought to look at propane again... from what I've read, the Isp is only a few seconds worse than methane and the density is much better. It might be the ideal SSTO fuel.

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

Propane has isp only a few seconds above kerosene but density much lower, near to methane.

Kerosene SSTO maybe even more close.

[Jim]

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

It isn't even possible must less practical.

[jongoff]

I'm not too worried about LH2 transfer on orbit. It needs to be flight demonstrated, but I don't think it's that high of a risk. If we get our SBIR Phase II, we'll hopefully have hardware that could be flight demo'd by the end of the 2yrs.

Is that SBIR to actually do such a transfer?

This would be excellent news and long overdue.

The SBIR Ph II proposal is to develop and qualify a small-scale coupler (initially focused on LOX but traceable to LH2). There's only so much you can realistically jam into a $750k research contract. That said, if we win the Phase II, we'll definitely be talking with as many groups as we can to see if we can push this through LH2 ground testing, and then eventually flight demonstration.

Further conversation should probably move to the Altius thread.

~Jon

[jongoff]

I think people ought to look at propane again... from what I've read, the Isp is only a few seconds worse than methane and the density is much better. It might be the ideal SSTO fuel.

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

Propane has isp only a few seconds above kerosene but density much lower, near to methane.

Kerosene SSTO maybe even more close.

Subcooled propane has a nearly identical bulk density to LOX/Kero. But I'm not sure this has much to do with whether LH2 upper stages are worth it or not...

~Jon

[Robotbeat]

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

It isn't even possible...

(gets really close to the microphone) Wrong.

[john smith 19]

I think people ought to look at propane again... from what I've read, the Isp is only a few seconds worse than methane and the density is much better. It might be the ideal SSTO fuel.

There's an old post from Steven Pietrobon on this. In truth at either NBP or a bit sub cooled all short chain hydrocarbons come out about the same with the same engine conditions except Proypyne (Methylacetylene to use its non systemic name). The strained triple bond was good for 5-10 extra seconds IIRC.

But it's horse for course. For absolute maximum Isp LH2 is  the way to go. Once other factors come into play  (IE you're not on a cost plus government contract) the choice becomes more complicated.

[Robotbeat]

For absolute maximum Isp on a chemical rocket: fluorine, lithium, hydrogen tripropellant.

[jongoff]

For absolute maximum Isp on a chemical rocket: fluorine, lithium, hydrogen tripropellant.

*Shudder*

For the highest non-terrifying Isp, LOX/LH2 with TAN and either a dual-expander or a very low-end staged combustion, so you can get away with a big expansion ratio while still getting good liftoff T/W, and if you go LOX-rich on the TAN, you might even be able to claw back a not-too-crappy bulk density.

If I were still running a propulsion R&D shop, I'd be trying to pitch NASA or AFRL on funding something like this.

~Jon

[Vultur]

I think people ought to look at propane again... from what I've read, the Isp is only a few seconds worse than methane and the density is much better. It might be the ideal SSTO fuel.

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

Propane has isp only a few seconds above kerosene but density much lower, near to methane.

Kerosene SSTO maybe even more close.

Yeah, that might be better given that there's far more experience with it... OTOH for an SSTO those few seconds matter.

IMO if you could make a propane engine with good TWR, that combined with SpaceX's mass-ratio performance, PICA-X TPS, and landing technology would make a VTVL SSTO very practical.

It isn't even possible must less practical.

Impractical I can see, but impossible seems really unlikely.

 The Titan II first stage could have made it to orbit by itself, IIRC, and the Mercury-Atlas only dropped engines - the Merlin 1D's TWR and Isp are more than good enough to send a rocket with Mercury-Atlas mass fraction (or significantly worse - Merlin 1D's current TWR is really high).

Building an expendable SSTO would be completely straightforward with today's technology. For reusability, SpaceX has already demonstrated vertical landing, so the big remaining question is TPS.

(Well, and speed/cost of reuse, but that's a "practicality" question not a "possibility" one.)

[Pipcard]

Okay, it seems that this thread has strayed too far from what it was supposed to be about - upper stages, not SSTOs - so can we please return to that? Thank you.

[Vultur]

Yeah, I'll start a new thread in Advanced Concepts.

[Vultur]

For absolute maximum Isp on a chemical rocket: fluorine, lithium, hydrogen tripropellant.

Aw, can't we get ozone in there somehow?

Li/F/H2 is definitely the highest demonstrated chemical Isp (542 seconds, IIRC). Is it the highest theoretical? I've seen claims for Be/O2/H2 holding that title - mostly just online rumors, but this actual NASA paper does seem to support it:

 "Current Evaluation of the Tripropellant Concept" by Robert L. Zurawski https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19860018652.pdf
"The Be/H2/O2 tripropellant offers the highest specific impulse of any chemical propellant combination".
It also says the Isp is 69.2 sec better than H2/O2.  They're modeling 1000 psi chamber pressure though -- I wonder if the advantage would be as significant at, say, an SSME's much higher pressure?

[Robotbeat]

Definitely not the highest theoretical. There is a list of increasingly insane chemical compounds that one could use. The problem is two fold: one, is that if you try to put more energy into the propellant, you'll melt the engine. That's why extra hydrogen is used, as it keeps the speed of the exhaust high while keeping temperatures lower. Theoretically, at very high expansion ratios (such that all the chemical energy gets converted to kinetic energy eventually), the highest Isp is actually with a stoichiometric mix as it has the highest specific energy.

The second is that these higher energy propellants will kill you if you even THINK about them too long.

[WBailey]

Free radical propellants could deliver thousands of seconds specific impulse, if they could be kept from recombining instantly upon formation.
Edit: Also metallic hydrogen, which may or may not have the same problem (we'll see soon)

https://www.history.nasa.gov/conghand/propelnt.htm
https://ntrs.nasa.gov/?N=0&Ntk=All&Ntt="Free radical" propellant
edit: https://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/19910012841.pdf

[Robotbeat]

Atomic hydrogen. Will kill you if you think about it too long, even theoretically.

[Chris Bergin]

Reacting to report to mod. Don't have time to read the thread, but trust the reporter, so....back on topic from this point onwards.  This thread isn't about SSTOs or the craziest propellant combos. It's about whether LOX/LH2 makes any sense relative to LOX/CH4.

Thanks.