Hacker News new | past | comments | ask | show | jobs | submit login
In California, giant Stratolaunch jet flies for first time (techxplore.com)
165 points by lelf on April 14, 2019 | hide | past | favorite | 131 comments

There are several advantages for airplanes as first stages. Less air drag on the higher altitude and more efficient engines, as they are optimized for smaller external air pressure. The speed of aircraft adds to the speed of the rocket. Another plus is that with aircraft you don't need to build a launch pad, the rocket is just released in mid-air. Yet another benefit is that aircraft can fly to a more preferable point for the launch - for example, the point where it is convenient to drop the first stage of the rocket, or the point closer to the equator so the satellite won't need to spend as much fuel to change the orbital plane. The airplane first stage can potentially fly to very different launch point, in many different countries, which can make logistics simpler. Those advantages all add up economically.

I've a detailed comment that addresses several of the points you're making at https://news.ycombinator.com/item?id=19660163

Is it true, strictly speaking, that you may use less propellant in the expendable stage due to air-drop? Sure. Although to be clear when talking about engine efficiency, it's really about being able to run a larger area ratio nozzle and getting a higher mission-averaged Isp. But ultimately when you consider all the "puts and takes" in launch system economics, those considerations are small potatoes (price out bulk LO2, and your choice of fuel). People (even university professors make this mistake) get so focused on minimizing propellant use that they fail to realize that complexity and dry weight are orders of magnitude more important.

The only serious advantages to air-launch are the logistical ones you noted. Also note that for U.S.-based systems, the idea that you will be able to fly down to equatorial latitudes requires you have the range to carry the laden launch system (and top-off to contend with LOX boil-off) pretty far. Strato certainly doesn't have these kinds of legs. Virgin will probably be able to get more out of their system, but it's not going to scale significantly higher.

I agree that for traditional launch systems it doesn't make that much sense -- particularly in light of the recent development of reusable (or at least, partially reusable) rockets.

That being said, I think the math changes somewhat for smaller launchers. The smaller the rocket, the worse the losses to the atmosphere get, and the more advantageous it is to escape atmospheric resistance in the soupy low atmosphere. So I think where air-launch systems could really shine, is by enabling low-cost launch of very small payloads.

I am, however, not entirely convinced that a sufficient market for those kinds of dedicated launch services will emerge (as opposed to hitching a ride in spare fairing space of a separate launch). I could see this going either way; as much as I'd like to see easier space access for eg. university-funded nanosats, I'm not sure the timing will be right for something like Stratolaunch, because the new private aerospace (SpaceX, Blue Origin, etc) is focused so intensely on pushing down the cost of larger-scale rocketry, to a point where that same market would be better served by bulk launches using some kind of standardized dispenser.

> That being said, I think the math changes somewhat for smaller launchers.

Do you have any analysis you can point to to substantiate this? I suspect any such analysis is heavily predicated on assumptions, such as whether or not the small launcher is designed around existing engines/motors.

I don't think there is a compelling case to be made that operating an air-launch carrier aircraft in addition to the costs associated with the remaining rocket stages works out economically^.

Orbital was not able to get compelling economics out of Pegasus. To be fair, they were doubly-hamstrung by their use of a carrier aircraft airframe operated by vanishingly few (L-1011) as well as the Pegasus using expensive solid rocket motors (here's another non-intuitive reality of rockets - solid motors are very expensive unless you buy in bulk).

^The most compelling test of economic viability of air-launch for small systems will come with Virgin Orbit/VOX space that is using a carrier aircraft airframe which is still in significant commercial use (quite a few 747 air freighters) as well as modern-yet-relatively low complexity LOX/RP liquid rocket propulsion. And if they want to say double their payload capability, they will not be able to do that with the 747. Maybe at the extremely low-end, GOLauncher will also be similar.

> Do you have any analysis you can point to to substantiate this?

Unfortunately I don't, no. I thought about running such an analysis as part of my grad studies (this was quite a while ago; I've switched careers since then) but I didn't get much farther than filling up a napkin before I decided I was already stretched too thin in my coursework.

Ultimately the premise is based on cubed-vs-squared relationships in rocketry (both in terms of aerodynamic forces and structural ones, though through a neat trick of math the mass of your fuel tanks actually scales linearly with their volume [1]). Like most other things in engineering, there are also economies of scale at play (for example, avionics mass consumes a smaller mass fraction of larger rockets), but my hunch -- and this is, as you say, fairly unsubstantiated -- is that the aerodynamic effects alone are sufficient. Cubed-vs-squared in aerodynamics is really just incredibly punishing. At 35kft (Stratolaunch's altitude) it's not as good as at 100kft or so (like you might expect with a weather balloon), but it's still a pretty big difference [2] -- basically allowing you to halve the radius of your rocket compared to an equivalent aero loss at sea level.

To take this to a really absurd level, I can imagine a 100kg rocket -- large by amateur model rocket standards, but beyond tiny compared to consumer rocketry -- with sufficient mass ratio to make it to orbit in a vacuum, but I can't possibly imagine the same rocket making it from the earth's surface.

[1] https://en.wikipedia.org/wiki/Pressure_vessel#Design

[2] https://www.engineeringtoolbox.com/standard-atmosphere-d_604...

> reusable (or at least, partially reusable) rockets

No such thing. A "reusable" rocket means you waste capacity carrying fuel for the return trip, and so raise the costs.

It's not like you can refuel the rocket once you get where you're going.

This is the reason why nobody uses "reusable" rockets today, not tech. (The tech was available back in the '60s.)

Fuel is almost irrelevant from a money perspective. However, two rockets both cost the same to build and one needs less fuel to orbit that directly translates into more payload to orbit. It’s a question of mass not money.

Alternatively, you can spend that budget on something else like a ‘single stage’ to orbit system which more directly lowers costs. This is more useful for tiny cargo like an anti satilite weapon system where getting even a few pounds into the right orbit works.

As to long term costs, having already built the aircraft that becomes a sunk cost. It might have been a poor choice, but it represents actually value before you can get satilite stock orbit. Which then allows a company to raise more money on better terms.

> However, two rockets both cost the same to build and one needs less fuel to orbit that directly translates into more payload to orbit. It’s a question of mass not money.

The way you state this is insufficiently constrained as to make the comparison I think you are trying to make. I think what you are trying to suggest that is that if two launch systems have the same per-launch recurring costs and the same gross mass the one that has higher payload mass fraction is better. Well, I can't disagree with the rocket equation.

If you are trying to relate this back to air-launch, my point is that for equivalent technology levels (meaning the same propulsion system other than nozzle area ratio, tankage with similar structural efficiency, etc.), the benefits you get in terms of recurring cost per unit mass from air-launch (by reducing rocket propellant consumed and dry mass of expended hardware) pale in comparison to the operating economics of having to additionally operate the air-launch platform. In other words, the recurring cost per unit mass to orbit, for the equivalent mass class, will be higher for air-launch than ground launch. And that doesn't even take into account the non-recurring costs associated with the air platform and need to amortize them.

> Alternatively, you can spend that budget on something else like a ‘single stage’ to orbit system which more directly lowers costs.

I'm not even sure where "directly lowers costs" applies in the SSTO case. I'm guessing you mean the case of a fully-reusable SSTO vehicle? At the end of the day, your per-fligt costs will reflect the need to amortize the non-recurring costs you've incurred. SSTO systems will tend to very high non-recurring costs, and the extreme sensitivity to dry mass means you will pay more per pound of dry mass for an SSTO vehicle than a multistage vehicle. The only people that think SSTO with today's technology will lower costs are the ones still hoping for the VentureStar to become operational.

> As to long term costs, having already built the aircraft that becomes a sunk cost. It might have been a poor choice, but it represents actually value before you can get satilite stock orbit. Which then allows a company to raise more money on better terms.

I'm not sure what you're trying to say here. If Strato can find a net cashflow-positive way to use their carrier aircraft, great. I think it's unlikely their launch economics will beat other systems in the marketplace so it won't win on pure cost short of Strato taking a loss on launches. However, in terms of launch, they may get some business if a customer really needs the logistical flexibility mentioned in the root post of this thread.

> pale in comparison to the operating economics of having to additionally operate the air-launch platform.

That’s likely true, but depends on how many launches you do. An air launch platform could likely handle multiple launches per day, but finding customers for thousands of launches per year is not going to happen.

> I'm not even sure where "directly lowers costs" applies in the SSTO case.

I am very specifically talking about a non reusable, anti satilite weapon. You can find several such weapons launched from aircraft. In that specific case using far more fuel and a larger rocket is potentially worth it to have dramatically fewer moving parts. Actual designs may be multi staged to increase capabilities to reach geostationary satilites etc, but it’s far less nessisary for LEO.

How does a rocket deal with those side stresses while it's attached to the airplane? AFAIK rockets generally do not deal well with sideways forces, because a rocket is a giant can.

It's clear that solid-fuel rockets are OK with sideways forces, because they're transported that way. The can may be weak (except re internal pressure, obviously) but the fuel/oxidizer is ~like rubbery plastic.

It's also clear that liquid-fuel rockets are OK with sideways forces, before they're loaded with fuel and oxidizer. Because they're so light. But once they've been loaded, they're very fragile re sideways forces.

I've watched lots of rocket-failure video. A ~common scenario is loss of attitude control, for whatever reason, followed by crumpling and fireball.

So I don't see how this thing could launch liquid-fuel rockets. Unless it could fly straight up, for long enough to transfer fuel and oxidizer. That'd be insane.

A liquid rocket vehicle would have to be designed to be carried and launched from a carrier vehicle like Strato has, but that's eminently doable. You design in the load paths suitable for carriage and also for engine-on ascent loads. There will be more structure attached to the cylindrical sections. But this added structural mass further detracts from the delta V advantages of being released from a "Stage 0" air platform at altitude and speed. And yes, you also design for the bending loads associated with acquiring the ascent flight path angle you need. Those bending loads will be more severe than what a well-guided ground-launched system will be designed to endure. Strato's aircraft undoubtedly pitches nose-up pretty hard before dropping the rocket, but it's definitely nowhere near vertical.

Solid motor casings are inherently quite stiff, to deal with combustion pressures. There would be less attachment/carriage hardware mass penalty than with liquids.

Bottom line, there is nothing inherent to Stratolaunch that makes it incapable of launching liquids, and even with the additional structural mass penalty, it's entirely possible for Strato to have lower cost per unit of mass to orbit via a liquid rocket stack than with solids (likely, in fact).

OK, I'm certainly no engineer.

But it strikes me that one could get a sense by looking at LPG and LOX truck tanks.

For LPG tanks, I found one source:[0]

    Material: 16MnR
    Yield strength (Mpa): 345
    Permissible stress (Mpa): 170
    Comparison: The wall thickness is ... 10-14mm.
I found that SpaceX is using 301 stainless.[1] But I haven't found anything about thickness. Anyone know?

0) https://www.anstertrailer.com/lng-lpg-tank-trailer-guide/

1) https://www.popularmechanics.com/space/rockets/a25953663/elo...

> Another plus is that with aircraft you don't need to build a launch pad

That one doesn't seem like an advantage, a launch pad is much cheaper than the world's largest really odd-looking and unique airplane.

Yes, but for a launcher you may want to have several launchpads for different kinds of missions, say GEO and SSO launches. Strato can serve them all. And if something happens to launch pad after rocket ignition, you have to rebuild the pad, while with Strato if you released the rocket and flew away, rocket problems don't affect the plane.

I hadn't thought of the fixed launch pad situation.

Do they expend a lot of energy / cost moving a satellite from where a fixed launch pad might easily place the satellite vs it's proffered orbit?

Mostly they have to pick their launch site so the flight path, headed into the right orbital plane, doesn't pass over populated areas. It would be rude to drop boosters and stages on people. https://history.nasa.gov/EP-165/p25.htm

Well, while possibly rude it still happens sometimes: https://www.extremetech.com/extreme/263732-china-keeps-dropp...

Will likely no longer be an issue once the new chinese spaceport in the south fully comes online.

Interesting fact, Israel is one of the few countries on earth that regularly launches into retrogade LEO (first stage flight path over ocean), because of the extreme political considerations around lobbing something that looks like an ICBM over their eastern neighbours. And the problem with dropping spent first stages on everywhere east of them.

Oh great, so if ASAT activity blows any of their stuff up, it'll have even more velocity relative to the other denizens of similar orbital planes.

If they were the self landing type and it landed in the local wall-mart parking lot.... would be kinda cool.

The uncontrolled versions, less entertaining.

I asked the engineers a while back, and they said the biggest gain was in potential energy that the rocket no longer has to make up.

They’re going to need to tighten up the landing. After touch down and braking, it was swing left and right. Seems kinda worrying

I'm puzzled by what added benefit being 10km up in the air and going at 300-400kmph gives over regular rocket launches since speed and altitude are a very small fraction of what's required to orbit.

The article says it will be less affected by weather windows but (1) would a normal LEO commercial launch be so highly time bound to find this of value? (2) if weather was bad for a rocket launch on surface, would you fly a heavily laden giant plane into it?

Leads me to wonder if the real customers of this is the military who do value very short turnaround launches.

Pegasus launches occur at closer to 900km/h.

It takes a Falcon 9 about a minute and roughly 150,000kg of propellant to get to 10km altitude, at which point it's going to be travelling closer to 1200km/h. If you can move that part of the mission to a carrier, you can substantially reduce the size and fuel capacity of the stage or maybe even entirely skip a staging operation (edit: not skip, carrier becomes first stage, thanks DenisM!).

Back to the Falcon 9 at mid-first stage, from 10km to stage separation you're running on motors that had to be built to operate at sea level. This means that they are 'underexpanded' for operating at 10km altitude and are therefore not as efficient as they could be. If you know your engines will always start off at 10km, you can build bigger expansion nozzles and improve the overall efficiency.

You’re not skipping staging, your carrier is your (reuseable) first stage.

And, I presume, much much more efficient

Why is it much more efficient? Doesn't it still have to carry the same mass up to the same altitude and velocity?

No, because most of a rocket's propellant mass is the oxidiser. With a jet first stage you get the oxygen from the air up to the point where you run out of air.

Yes but turbofan engines have a much higher thermal efficiency than rocket engines. Specific impulse is one metric used to describe efficiency, it's units are seconds and the value basically translates to how long a motor could maintain one pound of thrust with one pound of fuel. (Alternatively, how many pounds of thrust could the engine generate if it burns one pound of fuel in one second)

The specific impulse of a Merlin engine is on the order of 280s at sea level. The specific impulse of the PW4000's in Stratolaunch is closer to 10,000s. Now the Stratolaunch 'cheats' a little by not carrying onboard oxidizer, but that's almost three orders of magnitude more efficient.

Put another way, the Falcon 9 and Boeing 747 carry about the same amount of fuel onboard. However, that same fuel can push a Boeing 747 for 16+ hours while the Falcon 9 burns it in ~2.5 minutes.

If you accelerate your car 0-60MPH in 3 seconds vs. 30 seconds, which uses less energy?

Assuming you're talking about energy transferred through the drive wheels, I'd think the 3 second case would use less energy since less goes in to moving air.

Should the energy change be roughly the same = 0.5mv^2

The power would be different (energy is used up much faster in the 3 second case)

Indeed! Added a correction, thank you!

I wonder, why isn't possible an engine that can use air ( for oxigen ) then later use/switch another oxidant, when the air oxigen level drops?

edit: just found this https://en.wikipedia.org/wiki/SABRE_(rocket_engine). Seems that there is a solution.

Because fuel and oxidizer flow in a rocket engine is an incredibly difficult affair. When you add 'piping air-intakes' as a into that requirements list, you get an engineering problem that has stumped the people working on engines like the SABRE for decades.

The sad thing is that the development is quite underfunded, while VC & co pour money in all kind of internet start-ups...

Yeah, but that seems pretty rational. VCs are only slightly less short-term than public markets. While rockets are awesome, it is not clear that there can be a 10 year positive return on them. It could be possible that a startup with a bunch of VC could develop a ton of useful rocket-related technology and still not see a financial return for 50 years.

By contrast, internet businesses return or bust in that timeframe.

It does appear that we don't have a great funding mechanism that is expecting long-term returns vs short term. Past the 10 year expected payoff mark, there's debt, and there's grants. VC is a possible solution, but, there is very little willingness to stretch funds out past 10 years, since the 10 year mark is kind of a benchmark that is marketed to investors in the funds. When you are playing with other people's money, 10 years is already a really long time.

Maybe not space but a plane even faster then the Concorde without the noise issues.

VCs invest other people's money, who want to see a clear path to profit. They can't afford to invest in long term risky bet like space or fusion energy. VC is not the right model for space investment. Nothing sad about it.

It's also hard to crowdsource a space venture. Government's funds such as DARPA's or billionaire's such as Elon Musk's or Bezos's, or Paul Allen's in this case, make more sense. So our best bet to advance science is to have science friendly government in place. That or inspire rich people to be interested in science.

As an exception, there are segments that receive substantial investment, such as small rocket launch. Many many companies emerged in recent years such as Rocket Lab, Firefly, Relativity, Vector, Astra. Investors clearly see the market in near term for small satellites.

From Wikipedia [1]:

"40,000 feet (12,000 m) is only about 4% of a low earth orbital altitude, and the subsonic aircraft reaches only about 3% of orbital velocity, yet by delivering the launch vehicle to this speed and altitude, the reusable aircraft replaces a costly first-stage booster."

[1] https://en.wikipedia.org/wiki/Pegasus_(rocket)

It could replace an extremely weak, low capability first stage. In comparison the F9 first stage throws the second stage up to an altitude of 65km at 2km/s, or 26% of orbital velocity. So this air launch approach only achieves about one fifth as much as an F9 first stage.

That sounds like an 80/20 rule application.

You get to skip the densest part of the atmosphere, and you get to use an air-breathing jet engine (instead of a rocket engine that carries its own oxidizer) to get up there.

There's three advantages you get from air launch. One is that your first stage rocket engines can use nozzles that are optimized for vacuum or close to vacuum resulting in better performance through the entire first stage burn. (The RS-68 used on the Delta-IV, for example, has a sea-level ISP of 365s and a vacuum ISP of 420s, which is a pretty substantial difference.[1])

The second is that a launch aircraft at 30,000 or 40,000 feet is providing not just forward velocity but has also allowed the first stage to avoid some gravity loss (to accelerate at 1G at launch you need 2Gs of thrust). It's also allowed the first stage to side-step a major portion of the aerodynamic resistance.

Finally, an air-launched system is not as strongly constrained to specific launch windows. When targeting a specific orbit, you don't have to wait for a specific launch window at the primary launch site (the airfield), you can simply wait until one is relatively nearby and fly to it. This might be a big operational advantage even when it's not a big performance advantage since you can launch your aircraft when you have decent weather over the airfield and the fly to a more ideal spot to launch the rocket. If the radius of the loaded launch airplane is wide enough, this could be a very big advantage.

Even if you add all of these effects up, it's still a pretty small fraction of the 9,000 m/s of delta-v or so that you need to reach orbit. But keep in mind that rockets only deliver a few percent of their launch mass to orbit. A relatively small reduction in the delta-v needed to reach orbit might result in big gains in payload into orbit. In extreme cases, maybe it could double it.

The biggest disadvantage of air launch is that you are seriously constrained in the maximum size of the rocket. The Stratolaunch aircraft can carry about 230 metric tons of payload, whereas a Falcon Heavy weighs in at something like 1,400 metric tons[2].

[1] http://www.astronautix.com/r/rs-68.html

[2] https://en.wikipedia.org/wiki/Falcon_Heavy

Comparing stratolauncher payload to Falcon Heavy’s gross weight at takeoff is not really a helpful comparison. FH has a 60 metric tons to LEO max payload, roughly 5% of its total weight. I’m sure someone has done the calcs on the theoretical potential of the Stratolauncher, but it looks like a Pegasus can get 1000 lbs to LEO at a total weight of 40,000 lbs, so same 5% ratio. Seems to be in the same ballpark, and Stratolauncher (would?) use a fraction of the resources.

It depends on what you're trying to launch. It's kind of looking like the next big thing in the launch market is going to be a lot of small satellites in low orbit. This is a market that Stratolaunch might perform well in. It's not the right approach for manned space exploration and probably not for unmanned deep-space either, however.

I'd personally like to see Stratolaunch succeed, but without a dedicated rocket designed to exploit its scale, that's going to be hard. I think Rocket Lab's Electron rocket (a small rocket and a relatively small payload fraction on top of that) would likely benefit greatly from air-launch. On the other hand launching one from the Stratolaunch would be complete overkill, unless you stuck 5 or 6 of them on a rotary launcher or something. An aircraft like Virgin Galactic's White Knight Two would be a much better match for Electron.

The amount of rocket you need to reach a given speed is exponential with respect to that speed: https://en.m.wikipedia.org/wiki/Tsiolkovsky_rocket_equation

Shaving off a small amount of speed can have a large payoff. This is why most launch sites are near(ish) the equator.

That's not actually a very big effect compared to mid latitudes. The main reason to launch near the equator is because a lot of payloads go into equatorial orbits and plane changes are expensive.

From the equator you can go into any orbit equally easily (because you just wait for the earth to rotate until you’re on the desired orbital plane), but the only way to go into an equatorial orbit without a plane change is from the equator.

That’s a big effect, certainly, but the payload boost from the additional speed is still significant.

You can’t really get into any orbit equally easy when launching from the equator. The higher the inclination, the more delta v you need. For example, the Shuttle lost several thousand pounds of payload capacity when launching to the ISS compared to what it would have been able to carry to the Space Station Freedom, which would have been at an inclination matching KSC’s latitude. It’s just that the penalty is vastly worse for launching into a low-inclination orbit from high latitudes (thus why we didn’t tell the Russians to suck it and deal with an inclination better suited to KSC).

That’s a good point, the higher the inclination the more of the equatorial rotations Speed you need to cancel out. I’d still argue it gives you more flexibility overall compared to a higher latitude launch site.

It's definitely a very small effect. Equatorial launch sites are almost always preferable in terms of the launch itself, and the fact that most launch sites aren't smack on the equator is due to logistical or political issues.

Vaguely related, Baikonur is an interesting example of political issues. It's at 45.6 degrees latitude, but the minimum inclination it can reach is 51.6 degrees, because anything less would involve dropping boosters on China, and China is not enthusiastic about the idea.

The real answer, which is somewhat saddening given the immense amount of effort put into Stratolaunch, is that air-launch, when all is accounted for, has negligible benefits from a performance perspective when you are deploying expendable stages.

AIAA paper 2004-872 "Air Launching Earth-to-Orbit Vehicles: Delta V gains from Launch Conditions and Vehicle Aerodynamics" gives a decent and readable treatment of this scenario.

The advantages air-launch inarguably has are letting you pick the location and azimuth at which the high-performance launch vehicle stage begins its run as opposed to being constrained by your launch site and populations under the launch vehicle trajectory. If those are your most important metrics, air-launch may "win". If you are releasing a reusable stage from the air-launch platform, that could also be a unique value proposition as you could do a downrange landing of the stage (reserve no propellants for boost-back) to any desired site. SpaceX accomplishes this with a landing barge for their system.

By virtually any other measures, air launch from a subsonic platform does not make engineering sense over a ground-launched system. DARPA's ALASA program explored air-launch from an F-15, and was struggling to get ~100 lbs. to orbit. Even if you use a "big" supersonic platform like a B-1, by the time you take into account the volume associated with a launch vehicle, you quickly realize that at absolute best, you are talking a low supersonic staging Mach number. It would require a new high-thrust aircraft platform if you wanted a supersonic air-drop. And regardless of staging speed, if the flight path angle ("gamma") at release is too low, the deltaV advantage significantly diminishes (I recall the cited paper above shows this).

Consider the following issues:

* The Strato carrier aircraft is (unless even more megadollars are spent) is one-of-a-kind. If anything happens to it, no paying customer will be able to launch.

* Any stage stack that is launched from the carrier aircraft has to be rated for manned safety. This is its own expensive undertaking that autonomous ground-launched systems do not have to contend with. This is in addition to any public safety issues associated with population overflight (to be fair, air-launch may be able to diminish those concerns, but compliance with the former is much more onerous than the latter).

* The carrier aircraft presents very significant structural and volumetric constraints to the launch vehicle. The rumored lift capability is in the neighborhood of 500 klbs. For something like a hydrogen-fueled launch stack, you can do some rough sizing and realize that geometry will constrain you faster than mass. Don't forget to include flexible body effects, as that will be very significant with this airframe.

There is definitely an engineering accomplishment in the flight of this aircraft. However, unless Strato does something that they can uniquely do, or they lose money hand-over-fist on each launch (Pegasus is very expensive already, and I think SpaceNews reported that Strato shut down their own internal launch vehicle development efforts), they cannot be commercially-competitive. That launch point and azimuth control has to be critical, or air-drop just doesn't make sense.

As much as Burt Rutan has to be admired for his ability to think creatively, Stratolaunch was an insufficiently vetted architecture sold to an enthusiast billionaire who was probably delighted with the result of their SpaceShip One collaboration. And yet you see that in SS1, Scaled basically designed a one-off experimental vehicle architecture that fared very poorly when it was scaled into what was intended to be a sustainably-operated commercial system. Strato is another example of this kind of mindset.

Launches are very often cancelled due to weather, and launch windows are usually limited due to the physical location of the launch pad.

Stratolaunch solves both those problems.

Agree that picking the launch point and azimuth could address these issues. Of course, you'd have to be clear of ground-level weather enough for the Strato stack to take off, and the upper level winds associated with the trajectory would also need to be within the vehicle constraints, so it's not like you're free of thinking about weather.

While that's all true of air-launch, saying "Stratolaunch solves both those problems" is overstating the case. Right now, the only stages they have after their air carrier "Stage 0" are economically uncompetitive Pegasus XLs, and it's not clear what the weather capability is of their carrier aircraft.

Rockets waste a lot of fuel punching through the atmosphere at lower altitudes.

Planes can fly above (some) bad weather.

I think it's just a matter of needing less fuel, which saves weight, further lowering fuel requirements. So it just makes launches cost less.

Using the rocket to get off the launchpad is wasteful. It's basically the wrong tool for the first stage but the simplicity of the approach is nice.

It's probably a more enjoyable launch experience for passengers.

I thought speed was one of the most crucial elements of achieving orbit. Granted the extra 3-400kmph you mention is basically negligible for that.

A rocket engine nozzle produces thrust only for the part where the pressure is greater than atmospheric pressure. (Makes sense, right?)

Launching at lower ambient pressure (greater altitude) makes the engine work better. It gives more thrust for the same fuel flow (higher ISP). You can also make the nozzle bigger, getting even more thrust and ISP.

This is a significant advantage AFAIK. Also, AFAIK, if your rocket engine is low-tech, you will likely have low chamber pressure, meaning the nozzle is small. The beneficial effect is proportionally larger then.

> A rocket engine nozzle produces thrust only for the part where the pressure is greater than atmospheric pressure.

A rocket engine produces thrust when its chamber pressure exceeds ambient pressure, which is trivially true for any rocket engine. The flow coming out of the nozzle can have lower pressure at the exit plane than ambient and still make plenty of thrust. This is called "overexpanded" flow, and is the regime that produces shock diamonds in rocket and afterburning jet engine plumes at low altitudes.

Reduced to quasi-1D form, thrust is given by:

  T = mdot*(delta V) + (p_e - p_a)*A_e
Where mdot is mass flow coming from the nozzle, delta V is change of velocity of that efflux with respect to the vehicle reference frame (in a rocket, it is just exit velocity, in an airbreathing system, you have to take into account the incoming momentum), p_e is static pressure at the exit plane, p_a is ambient pressure, and A_e is the area of the exit plane.

In an overexpanded scenario, p_e < p_a , but the majority of thrust is coming from the momentum change term, which is the one containing mdot.

Nozzle performance is optimal when pressure at the exit plane matches the ambient pressure. This is when the exhaust velocity is maximized and the pressure thrust term vanishes. It can be shown (undergraduate aeropropulsion course homework) that you get more thrust from ideal expansion than from pressure thrust (underexpanded regime).

The point you are trying to make is that an air-launched system allows the rocket to start with a higher area ratio than if it were ground-launched. In the ground-launched scenario, a large nozzle area ratio will suffer overexpansion losses and how large you can go will also be limited by structural implications from the internal nozzle separation that will occur with gross overexpansion.

While this is all true, the reality is that the additional rocket performance you can get from a larger nozzle area ratio is not that compelling in light of building, operating, and maintaining an entire air-launch aircraft platform. You're better off putting that money into incrementally larger propellant tanks due to the lower mission-averaged Isp associated with the ground-launched vehicle.

I think you're arguing against something I didn't say.

You said:

> A rocket engine nozzle produces thrust only for the part where the pressure is greater than atmospheric pressure.

This is trivial at best and wrong at worst. When you say "the pressure", what do you mean? If chamber pressure, this is trivial and also not relevant to air launch vs. other modalities. If you mean pressure at the nozzle exit, the statement is false. Most of what I wrote above addressed why the latter is false.

It's the nozzle, so how could it be chamber pressure? I also don't mean nozzle exit, and I didn't write such a thing.

Take some small patch of the nozzle surface. If the pressure inside is bigger than outside, that area contributes to thrust. If the pressure inside is smaller than outside, then it contributes negative thrust.

Curious to hear if you think that is wrong.

If you look at the quasi-1D form of the thrust equation, you'll note the two terms I mentioned. Yes, you can integrate the surface pressure times the surface normals to get the pressure contribution to thrust, which will be negative in the case of the specific regions of the nozzle that are at lower-than-ambient pressure. As long as the geometric throat of the nozzle is sonic (as it would be in any meaningful rocket engine) and you have some divergent portion after the throat, the whole nozzle will not be at negative pressure thrust. But to say "the nozzle is producing negative thrust" is fallacious because in any non-trivial case, the rate of momentum change term in the thrust equation, i.e., the first one, will be very significant, and hence the flow leaving the nozzle will be contributing to thrust regardless of if it was overexpanded within the nozzle or not. Even in the most overexpanded case, the nozzle directly contributed to accelerating the flow - how can you say the nozzle is not producing thrust?

You can construct some pathological cases, like a very small thruster that is highly overexpanded to get the rate of momentum change term at or below the magnitude of the pressure contribution, but that doesn't represent a practical rocket engine (maybe more like a Reaction Control System thruster fired at ambient conditions).

I suspect your notion of thrust may be based just on the pressure contribution and is missing the rate of momentum change term, which will be significantly larger than the pressure term in launch system rocket engines.

I didn't talk about the whole nozzle. The first line of my post:

"A rocket engine nozzle produces thrust only for the part where the pressure is greater than atmospheric pressure"

With the word "part" I meant "patch of the nozzle surface".

Then you went on to talk how I'm wrong and talked about the chamber and the nozzle exit area etc. All irrelevant.

I guess it was a misunderstanding but it felt quite uncharitable.

> I didn't talk about the whole nozzle. The first line of my post:

> "A rocket engine nozzle produces thrust only for the part where the pressure is greater than atmospheric pressure"

> Then you went on to talk how I'm wrong and talked about the chamber and the nozzle exit area etc. All irrelevant.

Your statement is just plain incorrect (or being charitable, misleading) as written. When the velocity of the exhaust flow is > 0, even if the pressure in some section of the nozzle is below ambient, the nozzle (including the section below ambient pressure) is accelerating the flow, so every element of the nozzle's net contribution to thrust is always > 0.

At the end, I think you are talking about overexpansion losses, but short of pathological cases, in a rocket engine the overexpansion losses will not outweigh the rate of momentum change term.

Edit: To add further, the overexpanded nozzle is producing less thrust than the ideally expanded nozzle. But the overexpanded section is still producing positive thrust, it's just less than what you would have if the nozzle had ended earlier. It's not like that section has reversed sign in terms of thrust production.

Unfortunately, you're contradicting yourself. Every element of the nozzle in an overxepanded case can not contribute to thrust.

Let's break it down, it will be interesting to see which part do you disagree with?

1. Optimum expansion ratio has exhaust at ambient pressure. (we agree on this)

2. Adding more nozzle, going to overexpansion, reduces total thrust. We can call the expanded part of the nozzle the "nozzle extension". If this wasn't true, then part 1 wouldn't be true (optimum). (we agree on this)

3. The pressure in the nozzle extension, as experienced by the inside nozzle wall, is lesser than ambient. (my claim, that you disagree with)

4. Sampling some area of the nozzle extension, it constitutes negative thrust. (my claim, you also disagree with this)

5. If the overexpanded nozzle extension contributed positive thrust (like you claim), one could just keep on adding more nozzle and get more and more thrust. (Until flow separation.)

6. In vacuum one can add a lot of nozzle (because ambient pressure is zero) to get more and more thrust for the same fuel amount - until one hits mass or size constraints, or condensation or chemical effects too?

It doesn't really matter if the exhaust is accelerated by the nozzle extension or not. If the sum of total pressure points inwards, then when integrating over the nozzle extension area, force points downwards. This is because the momentum given by the exhaust is not the only force acting on the nozzle.

Most likely the gas is just too rarefied so the total pressure it exerts is less than ambient. Even if it gets accelerated (momentum, yay), it's not enough to overcome the pressure on the other side of the nozzle wall.

I'm interested in hearing why I would be wrong.

Replying for the sake of ensuring accuracy, not for the sake of argumentation. Got busy, wasn't able to get back to this any sooner.

We agree on #3 actually. Disagree with the "nozzle extension" terminology on #2, for reasons that should be explained below.

Next, let me admit a significant error in my parent to your post. I wrote:

> When the velocity of the exhaust flow is > 0, even if the pressure in some section of the nozzle is below ambient, the nozzle (including the section below ambient pressure) is accelerating the flow, so every element of the nozzle's net contribution to thrust is always > 0.

The "so every element of the nozzle's net contribution to thrust is always > 0" portion of the sentence is just wrong. It is certainly possible for a segment of nozzle to be accelerating the flow and still have a net negative contribution to thrust based on the pressure component. Indeed as you correctly suggest, this is the overall mechanism by which overexpansion losses arise. Very basic error on my part.

The two roots of my disagreement with your sentence are a) the simplistic relationship stated between pressure in the nozzle vs. ambient pressure being the sole arbiter of thrust production or not, and b) the ambiguous nature of the words "produces thrust". With regard to b), I believe we agree that it is both possible to participate in thrust production and also have a net negative contribution. I prefer something like the latter phraseology so things are unambiguous. As you have proceeded to parse out the specifics of what you meant, you have also clarified your words.

Regarding point a), thinking of the nozzle in a quasi 1D sense obscures the fact that the gasdynamics are at least 2D, and flow is not just being expanded in the nozzle, in anything other than a fixed conical nozzle, it's also being turned. Let's say I design an ideal bell nozzle by a standard method such as Rao's. At the end of the nozzle, when the ambient pressure is matched to the design value, the flow has zero angularity. Now let's run that same nozzle at a higher ambient pressure, same chamber conditions. If I just cut off the nozzle to a uniform length at the point where the internal pressure at the nozzle surface matches the ambient pressure, there will be > 0 angularity. Correspondingly, the pressure is not uniform across the cut-off exit plane. It's not axiomatically the case that continuing the nozzle some finite distance along its original contour, where these loss mechanisms are mitigated even though the internal surface pressure on that segment is below ambient, will produce less thrust.

What is happening mentally is the misleading equivalence between this "cut off nozzle" and an equivalent area ratio nozzle that is designed to terminate at the same nozzle surface pressure. These are actually two different contours, which quasi 1D gasdynamic theory does not distinguish. If the flow was always uniform axially while expanding (not possible), then the simple criterion you stated (part of nozzle below ambient pressure does not make a net contribution to thrust) would be true. But while it is not necessarily way off (specifics depend on several variables), this is just too simple a criterion to be correct. This is why for the purposes of this specific discussion, I think your "nozzle extension" way of thinking about the problem is misleading and should be avoided.

One of the biggest challenges for acheiving orbit is the dense atmosphere near the ground. Takes much much more fuel to go the same speed than it does when higher. It is dramatically less dense and then goes away as you get higher.

Scott Manley walks through why this isn't helpful yet since they have nothing useful to launch in delightful detail (complete with Kerbal simulations): https://youtu.be/yw84qJIGZeo

That's not a good summary. Yes, the plane is not very valuable in the short term (~4 years) because the only viable existing air-launched rocket is the pegasus, which is old and can be more cheaply launched on the smaller planes for which it was designed. But as Manley emphasizes, the whole stratolaunch plan is predicated on developing a new air-launched rocket, and we just don't know what the specs on that will be.

> the whole stratolaunch plan is predicated on developing a new air-launched rocket, and we just don't know what the specs on that will be

Neither will Stratolaunch, any time soon:


Can confirm. The whole group working on that effort has been laid off.

What an embarrassingly (for him) short sighted 'walk through'. Just having more capability to perform ariel launches will encourage companies to make vehicles. The likely reason nothing exists today is because there really aren't any options today..

Who is going to pay for companies to develop these vehicles?

Developing and certifying a new rocket is an incredibly expensive affair. What are the cost savings this promises, compared to SpaceX?

> Who is going to pay for companies to develop these vehicles?

VCs and other investors who see a company that wants to build these vehicles for cheaper launches? I don't really understand the point you are trying to make. Yes, it's expensive to build X, but usually $Y motivates people/companies to invest in the initial costs to cover building X because the payoff, $Y, will be greater than the cost. The news here is that the initial cost is potentially MUCH lower because now they can use this airplane as a stage 1 and not have to develop their own.

> VCs and other investors who see a company that wants to build these vehicles for cheaper launches?

So, a currently non-existant SpaceX 'competitor' that wants to hitch itself to the success or failure of Stratolaunch?

That's an incredibly sketchy business to build. You'll be sinking a billion or five dollars before you have any hope of seeing a cent of revenue, and then having an unknown probability of discovering that reusable rockets will eat your financial launch.

Right now, $X is measured in the billions, $Y is anywhere between 0 and billions, and there's a %Z chance (Which is probably over 50%) that your complement product (Stratolaunch) will die before you get a cent of revenue.

SpaceX is not the only commercial space company, you know this right? There are others, and hopefully with this launch plane there will be even more.

How did they prevent the two fuselages from torquing the wing apart? Seems like keeping both “sub-planes” from ripping the middle wing apart would be challenging.

One of my favorite airplane-related videos is one from the old days during 747 development, where they bent the wing until it broke. Turns out it is extremely strong.

Note: I did a search for that old video and came up with a more recent 777 video instead. Still interesting to see how far they have to bend it before it will break. https://www.youtube.com/watch?v=WRf395ioJRY

This was my first thought as well. But then I remembered that modern twin-engine commercial jets can fly with one engine down in an emergency. Not an aircraft engineer, but I suspect the structure is strong enough to handle normal torques during turning.

The complications with flying a twin engine during a single engine failure have more to do with flight characteristics than structural.

Ie, does the rudder have enough authority to stop a spin or does the engine provide enough thrust to maintain level flight.

The stratolaunch looks to have 2 rudders making things even more exciting.

Heavy turbulence might be exciting, though.

I’m not sure I follow. Unless they did something bananas with the design (which I admit is entirely possible given who designed it), the lift from the wings is transmitted to the rest of the aircraft through a spanwise spar(s). The fuselages hang off of the spar. What forces would rip that apart?

I would expect the forces go both ways. The fuselages have their own forces on the wings and since they are essentially sub-planes I’d imagine the forces could both strong and un coordinated - eg one trying to go one way while the other tries to go the other way. I’m sure under normal circumstances, computers keep everything aligned, but wondering about stressful situations.

Active control systems, presumably.

the center wing structure is strong

Isn't it amazing how you can tell a Burt Rutan design just by looking at it?

This is such a cool concept, I really wish these guys would succeed. Though last time I heard they gave up developing their own rocket (as a second stage to space). Anyone who's in the loop, does Stratolaunch currently have any potential rockets lined up?

Article states that they're going to use Northrop Grumman's Pegasus XL.

They abandonned all development activity of their own rockets, shortly after Allen's death.

The concept is indeed good, which is why people have been air-launching Pegasus rockets for decades. I’m actually not sure what the new thing is with this aircraft. Just that it can launch three rockets per flight?

They have been contracted to launch Pegasus XL rockets.

I know there's a lot of skepticism about whether the plane has any likely long-term use and whether it can be successful commercially at all. I find it spellbinding to watch it fly. I'm so glad Paul Allen funded its creation.

I get the distinct impression Stratolaunch wasn't happy with the post-landing roll:


(At 3m20s.)

I think its just fighting a bit of crosswind.

The edit cut away immediately, to onlooker shots. It's an interesting flight phase, and pretty clearly one Stratolaunch did not care to feature.

Several other onlookers did catch the roll, and the aircraft slewed markedly, with several cameramen commenting on this.

I agree.

It was crabbed a few degrees to its left while aligned with the runway (3:24 - 3:33 in GP's video). That would indicate crosswind from the aircraft's left.

If they're launching Pegasus XLs I wonder if they're on Northup Grumman's future acquisitions radar.

Their competition, Virgin Orbit has it's own subsidiary Vox Space that makes the launch vehicle.

It looks like intended rocket has changed a few times:


But did eventually settle to the Pegasus XL: https://spacenews.com/stratolaunch-abandons-launch-vehicle-p...

Is anyone looking at catapults yet? Seems that would be a great technology to pour billions into. Maybe that is why Musk is proposing Hyperloop? Get the engineering done for high speed trains, then scale it up. Way up.

The Startram project estimated $60 billion for a catapult that could launch people. http://www.startram.com/

How much acceleration can you endure? 3g is typical for a rocket launch, 9g is almost deadly. Accelerating to 6000mps at 3g will require 6000^2/(23g) = 600 kilometers of track, and it will take 200 seconds. That's a huge track, which has to be exactly linear or the payload will destroy it.

Then there is also the fact that the payload will be exiting the catapult at the sea-level, thus facing the full atmosphere at 6000mps. Which is to say the payload will be instantly incinerated unless some kind of magic is invented akin to supercavitation. By contrast rockets do not even reach supersonic 300mps until 10k altitude, thus never encounter this sort of problem. Maybe you could build this 600km track thing in the Himalaya, but that much and that high in the mountains isn't going to be cheap.

You can tell I though a lot about the same thing that you did. :)

I know people have talked about this for a long time. There must be a logistical reason for it.

As soon as we figure out how to build 100+ km tall structures.

As cool as it looks, this airplane seems promised to go the way of the Hercules H4: fly once then stay in a museum forever. This recent podcast from mainenginecutoff speaks about Stratolaunch [1]

[1] https://mainenginecutoff.com/podcast/108

I wonder why this type of specific purpose airplane needs human in the cockpit? Why can't it be remote controlled?

I suspect this huge, complex aircraft with it's very expensive, highly explosive cargo isn't as straight forward as flying a drone.

I'm curious what the emergency landing options are for this thing. It looks like there would be limited places it could land with that wingspan and landing gear spread.

Edit: Curious, not critical. Maybe it can land in a lot of places, just reacting to the size.

Since its purpose is to launch rockets, I assume it lands where it takes off, and might never stray far away.

That's the sort of thing I'm interested in. Like, for example...how far away are you once you get this huge thing high enough to launch.

I'm not trying to criticize it, I'm just curious how dangerous a typical mission will be.

It doesn't have to fly in one direction, it can do s-turns while climbing, or circle.

I think clearly payload launch is the risky part. What are the kinds of launch failures? Rocket engine failure can be uncontained and spectacular.

You can make occasional turns to keep close to the airport while climbing higher. I don’t know if that’s what they’ll do, but this problem has some easy solutions at least.

> Like, for example...how far away are you once you get this huge thing high enough to launch.

It can climb in a circle so that it's never far from its base.

It only reaches 35,000 feet.

They're going to move it all over the world (or at least, over the latitude range) in order to get into the most optimal launch position.

Me too. I have serious questions in regards to maximum landing weight, will or won't the payload have to be jettisoned, and whether there will be propellant restrictions.

Imagine an emergency scenario where a high payload of hydrazine/NTO/UDMH/MMH has to be dumped at low altitude to recover the plane "safely".

With a traditional vehicle you at least will never have an incentive to take your launch vehicle anywhere where down range includes a low altitude intersect with a population center.

This on the other hand... Oof. This brings an entirely new dimension to Range Safety. I can see the lesson having to be relearned the hard way if we aren't careful. Not saying it isn't worth it, but man, those postmortem would not be fun.

Would be a hell of a thing to see go off without a hitch though.

With an option to jettison the payload I would imagine it can land relatively short. The majority of stopping distance is related to having enough rubber on pavement and keeping brakes cool enough. Stopping energy required changes dramatically as weight is reduced as it’s less energy required to stop and a slower touchdown speed.

I’m not sure how the dimensions of the aircraft affect its ability to use other airports though and exiting a shorter runway after repair can also present a major challenge.

Was there for this flight yesterday, and with someone who worked on the project.

Most runways are too narrow for both the wingspan and wheel span. It is quite limited in areas it can land and take off.

I hope they take inspiration from these folks and do a simultaneous takeoff with a smaller plane, in opposite directions:


I wonder how fast it would disintegrate in a case of asymmetric stabilizer runoff.

I guess this thing has a pretty sophisticated control system to keep it stable, similar to modern combat aircraft. I also wonder if they have any fuel dumping procedures for an aborted launch; landing with a non-empty launch vehicle still on board seems quite scary.

Why didn't they just mount the tailplanes higher and join them? (e.g. like the OV-10 Bronco tail)

I hope that center wrong is unbelievably strong to handle the tortional forces that two separate horizontal stabilizers on lengthy lever arms might exhibit in acute turbulence.

That flimsy tail design bothered me as well. Even if the center wing and wing roots are incredibly strong, there's little reason to place them under stresses that could be alleviated with a single-stabilizer tail design, as other commenters have pointed out. The fact that they did not do this suggests this aircraft is designed to launch millions in contributions into the designers' pockets, and not so much rockets into space.

Does anyone have a cost comparison between different launch systems?

Kind of a threadbare article. Not a word on why dual fuselages are desirable? Why it needs to have the world's longest wingspan? Like I'm supposed to just know all that from my aeronautics background? What if I flunked that course?

You need a large space open to attach the rocket to, and it has to be at the center of lift. You can't have one fuselage and put the rocket to the side, because then the balance would change drastically when you add the rocket, and you can't put the rocket under/in the fuselage because it's very large.

Thanks... it's the simple things!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact