Over on twitter today, Jeff Foust (@jeff_foust) reported on a new reusable rocket project that will be starting in DARPA’s Tactical Technology Office (TTO) next month. Rumor has it that Jess Sponable, the former Program Manager for the DC-X project, and long-time advocate for reusable launch vehicles, will be leading this new project.
According to Jeff’s twitter notes from Pam Melroy (DARPA TTO’s Deputy Director) talk at SPACE 2013:
- The goal is for a reusable first stage that can fly 10 times in 10 days, at speeds up to Mach 10
- The first stage would be sized big enough that with an expendable upper stage, it could place 1-4klbs in LEO
- The target launch price (not sure if this is for the 1st stage only or for both stages) is $5M/launch
- While the concept art showed many winged vehicles, they’re open to other approaches, so long as they can do staging at Mach 10
- There will be an Industry Day for XS-1 sometime early next month–this is an opportunity for interested parties to learn more about the program, and ask questions of the Program Manager and Contract Management team there at DARPA
- A Broad Area Announcement (BAA–a form of solicitation DARPA uses for most programs) is planned to be out sometime next month
That’s all Jeff shared via twitter, but I wouldn’t be surprised if he did a blog post with more details tonight on Space Politics or over on his NewSpace Journal blog. I’ll link to them if he does. I also created a discussion thread over on NASASpaceflight.com for this topic.
A couple of thoughts I mentioned there:
- Mach 10 is a pretty high staging number. Most of the RTLS First Stage concepts I discussed here on Selenian Boondocks optimize at staging numbers in the Mach 3-5 range. In fact, other than a hypersonic flyback stage (which I doubt is the best choice technically or financially), the only approaches that would yield anything near optimum performance with a Mach 10 staging requirement are the two Lift-Assisted Boostback options I discussed in my Boostback post. Knowing Jess, I’m sure this requirement is based on some sort of analysis, so I’d like to see where it came from. I worry that by picking such a high staging speed, he may be eliminating many promising alternative approaches. Though DARPA TTO has had a strong interest in hypersonic vehicles for a long time, so it could just be pressure from inside the office. Hopefully this still allows for a range of creative options. I guess it might be possible to do this boostback style with a slightly-less optimized VTVL vehicle of some sort.
- A reusable first stage with an expendable upper stage that can put 1-4klb into LEO could likely put 300-500kg into LEO with a reusable upper stage. This was in an interesting sweet spot that I was discussing with Jeff Greason at Space Access a few years ago.
- It’s important to remember that historically, most DARPA missions get canceled before flight. Some of this may be due to the high planned turnover at DARPA–program managers get hired for something like 3yr stints, after which time they go back to industry or elsewhere in government. The good news is if they can make enough progress, a lot of times technology from a mission that gets canceled can get infused into a newer mission with slight different focus (like the FREND robotic arms from the cancelled SUMO project getting baselined into the Phoenix program). I’m still super-excited that Jess was able to talk DARPA into funding something like this, and he does have a good shot if he has adequate funding. I just don’t want people acting as though this program has already succeeded. I sure as heck hope it does, and Jess is a sharp program manager with experience running projects like this before, so I’m hopeful. It just seems like cool space ideas get jinxed by premature overoptimism.
So, as I said, I’m really excited to see how this turns out. I wish I could attend the Industry Day unfortunately, but since Altius isn’t really a rocket company anymore, justify the travel costs may not be feasible. But if someone else can go, I’d love to get a report.
And good luck to DARPA!
[Update #1 2:59pm MDT: Over on my NASASpaceflight.com thread, yg1968 just posted some slides from the presentation, and a link to a video of the presentation–the relevant part starts just after the 20min mark, and confirms my concerns about them conflating hypersonic testing and low-cost access to space.]
[Update #2 3:05pm MDT: Jeff Foust just did a Space News article on the topic.]
Jonathan Goff
Latest posts by Jonathan Goff (see all)
- NASA’s Selection of the Blue Moon Lander for Artemis V - May 25, 2023
- Fill ‘er Up: New AIAA Aerospace America Article on Propellant Depots - September 2, 2022
- Independent Perspectives on Cislunar Depotization - August 26, 2022
Excellent news! Perhaps an impetus came from the celebration this year of the DC-X.
I think we already have, effectively, a vehicle that could perform this role: the X-33. Whether you blame NASA or Lockheed I think the cancelling of the X-33 was premature. If you run the numbers you find even if you replace the composite tanks that failed with aluminum-lithium ones, the X-33 could still act as a first stage that with a cryogenic upper stage could send a few metric tons to orbit.
I’ll write about this in an upcoming blog post.
Bob Clark
Jess has been pushing for something like this for a long, long time. I’m not going to speculate on where the Mach 10 business came from, I haven’t spoken with him in some time.
I don’t know if it’s of any relevance, but recall that the reusable winged rocket segment of Swiss Space Systems’ design also releases the upper throwaway stage at Mach 10.
Charles,
I’d love to find out. Could very well be that someone did a study that I haven’t read about where Mach 10 makes sense for a reusable first stage–maybe something clever like the second lift-assisted boostback concept. If you do get a chance to speak with him, I’d love to hear what his thoughts are. I’ve got a ton of respect for Jess with DC-X and what he’s been trying to do for the industry over the years.
~Jon
The S3 guys say they chose Mach 10 separation to reduce the size/cost of their expendable upper stage, which they believe will be the most expensive element of routine operations.
Also, given that they can fly up-range to launch, down-range booster recovery may be somewhat less demanding on glide/boosted fly-back.
Yeah, if you could use an airlaunched first stage, the Mach 10 staging plus RTLS would be *tons* easier, even for a VTVL style first stage. It’s if they insist on a ground take-off that it becomes really challenging.
~Jon
I think some features of this program are a result of the government funding paradigm that creates the “valley of death”. There’s funding for the R part of R&D, but no funding for the D part. If you really wanted to launch 1-4klbs into orbit with a high flight rate for low cost the best way to get that would be to give XCOR, SpaceX, and Blue origin indefinite delivery indefinite quantity contracts to do that and let them stage at whatever mach number they want. But government research money can only be spent on things that are “new”.
Why now? Could it be that commercial efforts are causing NASA some discomfort since commercial is actually showing signs of successfully moving ahead with the holy grail of space launch from Earth. And if that’s the case then I wholeheartedly agree with Bob, give it to commercial. NASA’s only going to be playing catchup and at the rate certainly SpaceX is moving, they won’t win that game.
Neil,
This is a DARPA program, not a NASA one. With Jess Sponable behind it, I think its an honest attempt at getting more money to seed reusable space vehicles. No conspiracies needed.
~Jon
I worked for Sponable on DC-X back at the Pentagon in the 90’s. This is exciting news. Hope to read more soon.
I too noted that MACH 10 is a very high staging number. Interesting. I remember having discussions with Sponable and others about a “generic” flyback booster that could handle a variety of upper stages.
Note on the Mach-10 part, what I read on the DARPA site says that the stage is required to reach Mach-10 “at least once” during the program and my thought is that the performance difference could bring it well into the range of using that “extra” for Boost-Back purposes.
The also state that NOTHING at this point is to be taken as a “given” even though the illustrations show a winged vehicle they are putting “everything” on the table at this point.
A question/concern though, they mention that the program is compatible with the Airborne Launch Assist Space Access (ALASA) program (launching a multistage expendable rocket from a “conventional” sub-sonic carrier aircraft) which as I understand it was canceled earlier this year and the funding pulled. Is that still an “ongoing” program or not?
Randy
Randy,
Re: the DARPA website’s clarification on the “Mach 10” requirement, I had seen that yesterday, and mentioned it on the NASASpaceflight.com forum, but forgot to mention it here. The clarification definitely gives me more hope, though a lot still depends on caveats that probably won’t be clear until the industry day is over. But it does make me more optimistic about things.
Also, regarding ALASA, it’s still an active program. It was SeeMe, which was a program to develop a type of payload for ALASA that got canceled.
~Jon
Jon. Sorry, I knew DARPA wasn’t NASA. Must be tired. Wife’s gone back to her home country for a holiday and you don’t know what you’ve got ’til it’s not around – huh, not much of a one, holiday that is, here I can tell you!
Hmm. There might be a unstated, underlying reason for why DARPA wants the stage capable of Mach 10. I found this reference while doing a web search in regards to the Air Force’s Reusable Booster System program:
Future Launch Systems.
Robert Hickman and Joseph Adams
Crosslink, Volume 5, Number 1 (Winter 2004), p. 42.
“Reusable launch vehicles are commonly proposed as responsive and inexpensive alternatives to expendable rockets. Analogies to aircraft systems suggest that reusing flight hardware should substantially reduce cost.
“According to Aerospace analyses, reusable launch vehicles that have been optimized for minimum dry mass have staging velocities (that is, the velocity at which the second stage deploys) roughly between Mach 10.5 and 11.5. In this case, the orbiter will be about half the dry mass of the booster. The mass of the reusable launch vehicle will grow steadily as the staging velocity deviates from this range.”
…
“Assuming optimal staging, at about Mach 7, hybrids expend about 35 percent of the hardware a comparable expendable rocket would expend. Thus, their recurring production costs are much lower. Also, the mass of the reusable booster stage for a hybrid is about 45 percent that of a fully reusable launch vehicle. Thus, development and production costs are significantly less. For these reasons, even relatively low launch rates could economically justify their development.
“The hybrid vehicle also carries less risk than a fully reusable launch vehicle—primarily because it does not employ a reusable orbiter. Reusable orbiters present a difficult technical challenge, as they must survive on-orbit operations and reentry through Earth’s atmosphere without significant damage. The reusable booster experiences a much less severe environment, resulting in fewer technical challenges and less risk.”
http://www.aerospace.org/wp-content/uploads/crosslink/V5N1.pdf
The analysis in this article concludes that while both reusable upper and lower stage systems would optimize at about Mach 10 staging, a hybrid system, meaning a reusable lower stage and expendable upper stage, would optimize at Mach 7.
Then perhaps DARPA’s goal is to get to a fully reusable system. And rather than having to redesign such a lower stage, start from the beginning with a stage with this staging velocity.
Bob Clark
Bob,
Interesting article. Maybe that’s where they get it from. That said, I’m glad that as Randy pointed out, the Mach 10 doesn’t look like it’s the staging velocity, but its likely the “we want to fly this first stage up to this velocity at some point” velocity. That may potentially make things a lot easier.
Normally staging at Mach 10 drives you pretty quickly to jet flyback stages, and they’re annoying, IMO.
~Jon
I suspect they’re focused on a horizontal takeoff and landing because it would be almost impossible for a tail-sitter (conventional rocket) to be turned around in anything remotely close to a single day. A year or so ago I ran some numbers on launching and landing a booster horizontally, and it actually looks very workable, with a low mass penalty, greatly reduced operational costs, and quick turnaround.
George,
What is so special about HTHL that makes it uniquely suitable to fast turnarounds? It’s not like we have a wealth of data from previous attempts at high-flight rate TSTO rocket vehicles of any sort, let alone data points that show that a VTVL “tailsitter” is somehow uniquely harder to turn around than an HTHL spacecraft.
Could you explain your logic for why you think HTHL lends itself more readily to rapid turn around, when you’re talking about a TSTO rocket that has to attach an additional stage somewhere that doesn’t interfere with takeoff aerodynamics for the HTHL?
It’s possible that you’re right, but I think you’re stating your opinion more strongly than you ought to with how little you’ve offered other than assertions.
~Jon
Oops. Sorry about that. I was typing, got distracted by the neighbors, and since I was at the end of a thought I hit the comment button. I had been writing up a description of the concept, and here it is.
The concept is to take off and land oriented horizontally, but vertically under rocket power (like a Harrier). This gives the ground handling convenience of a conventional aircraft without requiring heavy, expensive landing gear and wings, or having the take-off weight determined aerodynamically. The mass penalty is relatively small.
The problem with trying to launch a large booster horizontally like an aircraft is the extremely high required take-off weights, inducing a large mass penalties for landing gear and wings, along with the wings’ high drag penalty as the aircraft goes hypersonic. The fuselage must be rather heavy to handle the horizontal bending loads of the high fuel fraction, since the applied forces to the fuselage are concentrated at the main landing gear and wing spar. This all limits performance, fuel fraction, and ultimate size. By most rules of thumb, the landing gear alone will weigh more than the rocket engines, and as vehicle size increases the wing will represent a greater percentage of the empty weight. There’s also the unavoidable problems of development cost and schedule with a large re-usable hypersonic flight vehicle.
The largest supersonic aircraft built (the XB-70, B-1, and Concorde) had maximum takeoff weights less than 600,000 pounds. The stainless steel Valkyrie had an empty weight that was about 47% of maximum takeoff weight, and both aluminum aircraft had an empty weight that was about 41% of maximum takeoff weight. Taking those examples as starting points, and given that a booster capable of Mach 10 is going to devote a fair percentage of dry mass to thermal protection, the mass ratio is not going to be very high. Even if you get the dry mass down to 30% of max takeoff weight, unless you push the bounds of historic aircraft size the vehicle probably won’t carry half as much fuel as a Falcon 9. Given the inherently low mass ratio, unless it uses very advanced air-breathing engines the payloads such an architecture could deliver into orbit could never be very large.
Turning to more mature, conventional space launch vehicles with high mass ratios, the problem is size and vertical orientation. As such rockets get large, assembly, transport, and launch support systems become large and prohibitively expensive to build and operate. Commercial providers like SpaceX, along with the Russians, try to avoid vertical assembly because of the difficulty and cost, assembling and moving the rockets horizontally and then raising them vertically at the pad. Yet even doing that requires a lot of fixed and expensive launch infrastructure, and it’s also slow and complicated, eating into the 24-hour turnaround XS-1 is supposed to achieve. The turn-around on a large launch pad like NASA traditionally uses takes at least several days since it has to be inspected for damage from the just departed launch, then readied for the new launch.
Landing vertically as a tail sitter, like the SpaceX grasshopper, requires a shock-absorbing landing gear that does not also allow the very tall, light structure to tip over, especially in moderate to high winds. The gear almost needs to be a giant oleo strut with wide splayed feet. Once the booster has landed vertically on its tail, there’s still the problem of attaching a new second-stage to the top, since the booster obviously can’t land inside a vertical assembly building. The booster has to either be lowered back to horizontal for stage attachment (which might take the better part of a day) and then re-erected, or it has to be lifted, secured, and transported back to the processing facility, where the second stage can be mated, then transported back out to the launch area. That alone almost rules out anything close to a 24-hour turnaround time. Tall things are expensive and difficult to work on. Long things (like ships and airliners) are easy to access, maneuver, and transport.
The tail sitting configuration creates almost insurmountable issues with achieving a daily flight rate with a multi-stage conventional vehicle (though a week probably wouldn’t be an issue), so a very high-flight rate vehicle should avoid that orientation. But going down a runway for a horizontal takeoff introduces performance and size issues. Leaving the rocket oriented horizontally and then performing an upward launch (like a Harrier) followed by a quick 90 degree rotation solves a lot of problems. The rotation wouldn’t require a high degree of precision (although that would be easily achievable), since any errors would be quickly corrected by the main engines once the rocket is pointed upward, just like a Trident missile that comes wobbling up out of the ocean before the main engine ignites to establish positive control. As an added benefit to horizontal lift-off, since there would be no launch tower, there’s no tower to collide with.
I ran some calculations on the rotation, and it could easily be completed in anywhere from 5 to 30 seconds, depending on what angular velocity and acceleration limits you want to set. You could do the whole lift and rotate with solids, but then you’d have to replace them every flight and couldn’t use them for landing after an abort, although you could drop them to shave weight.
The horizontal configuration obviously requires a near doubling of the total engine thrust, since the thrust of both the horizontal lift engines and the main engines have to exceed the total vehicle weight, but the mass penalty is small because many modern engines have thrust to weight ratios well exceeding 100:1, so the extra engines would add less than one percent to the gross launch weight. This could even be improved by having some of the main (rear) engines pivoting 90 degrees for takeoff (dual use), and some of the horizontal engines could pivot far over to provide some extra thrust (though with large cosine losses) during the early stages of vertically oriented flight, until the fuel mass drops enough to let the main engines alone provide the thrust.
If the horizontal take-off engines are relatively small and spread along the length of the rocket, they will support its mass along its length and won’t induce the large bending loads that a more conventional aircraft configuration would, where the loads are concentrated at the main gear and wing spars. This would help keep the structural weight fraction low. If such engines are essentially pressure fed thrusters fed from a common supply trunk, the rocket could use a lot of them and individual engine failures wouldn’t cause serious issues, while individual engine replacement would be fairly cheap.
Their are many other side benefits.
The take-off thrust, and thus exhaust, becomes spread out along the width and length of the rocket instead of being concentrated at the base. Most conventional rockets like the Falcon 9 or Saturn V have a thrust versus footprint area in the range of 6,000 to 12,000 pounds force per square foot. If they took off horizontally instead of vertically, the thrust concentration would drop to around 500 or so pounds per square foot, greatly reducing the pad’s blast requirements and perhaps allowing takeoffs and landings on nothing but bare flat concrete. From that it follows that the rocket could perform an abort, fly back, re-orient horizontally (the engines spread along the length of the rocket give excellent pitch and roll authority), and land back at the launch pad with relative ease. That also means that the system could be test launched and flown many times before attempting orbital launches.
The controllability in the horizontal orientation is very good, with roll, pitch, and yaw authorities far in excess of a tail-sitter. A pilot could probably land it like a Harrier without much computer intervention, as opposed to balancing a broomstick. Winds at the landing site would no longer be a serious issue for takeoff or landing. Once on the ground again, the rocket would be as easy to move as any airliner that airport workers shuffle between terminals on a routine basis, as opposed to having dozens to hundreds of people in hard hats watching a giant vertical rocket moving at the speed of a turtle.
I’ve looked at quite a few configurations, having started out with the idea that the rocket would be a conventional cylinder in vertical flight, with horizontal engines that pivot out from the booster’s inter-tank and interstage sections, mounted on arms quite like retractable landing gear, with semi-circular doors that would close flush with the airframe. I even thought that the arm could hinge along fuel and oxidizer ball valves, which would make it impossible to fire an engine except in the full out and locked position.
Then I imagined simplifying things by just giving the rocket a fairly flat underside (slabs added to cylindrical tanks to resemble a heavy transport aircraft) and mounting the engines in the chin area. This has the advantage of not concentrating the loads into just the inter tank and interstage sections, distributing the takeoff thrust down the entire length of the rocket, and allowing multiple sets of landing gear legs to evenly distribute and support the fully-fueled weight.
But a more useful and interesting configuration arises if the horizontal lift-off stage is used as a strap-on booster lifting a relatively unmodified conventional rocket such as a Delta, Atlas, Falcon, or smaller missiles. Yet the booster would need to lift a core stage (which would lack the mass penalty and complexity of horiztonal lift engines), and mating two cylindrical boosters to a cylindrical center core and lifting it horizontally puts a tremendous amount of torque on the mating joints (which isn’t the case with strap-ons in a vertical configuration). So, instead of building the strap-on boosters as cylinders or with a slab side like a cargo plane, build them with chines like the SR-71 fuselage, so that the two strap-on booster’s edges almost touch at lift off, with the core stage nestled on top of the mating chines. That allows the core stage to be completely and evenly supported along its entire length, and the chines provide space for the lift engines to transfer their thrust almost directly to the core stage, a very structurally efficient placement. That also allows the chines to shield the less robust core stage from lift off blast effects.
A side benefit of horizontal launch for the flyback boosters is that if they need to brake prior to re-entry, they only have to rotate about 90 degrees instead of 180 because they have thrusters along the bottom, not just in back, and they can hold that orientation through most of the re-entry. The thrusters also inherently give tremendous pitch and roll authority that’s virtually absent when re-entering with a conventional strap-on booster. Then the chines both shield the entire upper side from the peak re-entry temperatures, allowing for reduced TPS mass, and the chines provide a much higher lift coefficient for an atmospheric turnaround a high angles of attack. They’d also provide an ideal location for swing wings or deployable air-breathing engines if the fly-back range requirements are excessive.
If you pushed that idea further, the center core station could share the chine configuration as it moves toward its own re-usability. Oriented upside-down on the ground so that it’s chine stretches across the top of the two adjacent boosters, it would actually reduce wetted area. For ground handling and mating, the core section would be supported, unfueled, from both ends and the strap-ons would roll in from each side for attachment. Such strap-ons could be made for almost any launcher in the inventory, and the only major modification to existing rockets would be air starting their main engines. The strap-on system would give a payload increase to existing rockets that don’t currently use strap-ons, while eliminating much of the expensive ground handling and infrastructure required to support vertical launches. This would allow the program to produce tangible payload benefits even with early, small scale prototypes lifting and boosting solid-fueled military missiles, proving the overall concept, producing a wealth of operational and aerodynamic data, while generating revenue.
Unlike the hypersonic aircraft configuration, the horizontally oriented vertical launch and landing architecture scales up well, with much larger practical size limits. Conventional vertically launched rockets also suffer severe scaling problems in their support equipment. The SLS is already pushing the maximum height allowed by NASA’s VAB doors, and going any taller would require a new building even bigger than the old one, along with new launch pads, and possibly all new crawlers. In contrast, a horizontal takeoff vehicle could be stretched without impacting hardly any infrastructure.
So my proposal would be to start with a small model rocket demonstrator using some Estes engines to show the lift and flip-turn in action, using timed solids, and then scrounge some RCS thrusters sitting on a shelf to build a working liquid-fueled model to test the concept more thoroughly, developing the control software and exploring flight modes much as was done with the DC-X, and continue the project if it looks promising.
***
I’m sure I’ve missed some things in all that, such as making sure the fuel for the landing phase is in small side tanks instead of the main tanks, because fuel sloshing (front to back) as the rocket approaches horizontal and tries to stabilize for touchdown would be an irritating issue.
Also, considering that the weight will be vastly reduced for landing, is the option of just a conventional aerodynamic landing. One of the reasons I tried to avoid that is that it would require the development of something like the Shuttle’s APU system to drive the hydraulics for the control surfaces, whereas everything needed for a pure-powered descent is already included for the take-off phase.
The concept is so simple that anybody could run with it and produce some good numbers on mass and performance.
Darn it. I forgot a section of thoughts on how you’d develop this with something like a Falcon 9, which I used as a representative platform for early analysis.
First, for the cross sectional profile, draw three equal-diameter circles along a line so their sides to touch (the standard strap-on configuration), then draw a line along the bottom The area above the line and under the circles is the chine profile. If you draw a line across the top you get the core stage’s upper chine profile. Of course, you might want the lower section to be a semi-circle to even out the aerodynamic heating during re-entry. That also makes more space for engines and landing gear.
Take a Falcon 9 booster and stretch it to the length of the upper and lower stages of a Falcon 9 v1.1 and include extra anti-slosh baffles. The boosters need to be as long as the main stage so they can fully support it along its length, otherwise the bending loads on the second stage attachment would be high during the lift-off. A useful addition to the boosters would be a long needle nose that would support the large payload shroud from both sides, and in the case of the existing large-diameter Falcon shroud, these supports would be tubes about one foot in diameter.
The chine sections will include landing gear struts, possibly with metal casters (rocket exhaust and rubber tires don’t get along well) so the stage can be rolled around on the ground. The lift-off thrust for early experiments could be provided by pressure-fed Super Dracos, but ideally by a small, highly reliable LOX/RP-1 engine, possibly up-scaled XCOR 5K18s. The number of engines (and thus the required thrust per engine) is a function of how evenly you want to distribute the thrust along the length of the rocket.
That vehicle is then flown around their facilities for a few months like it was a helicopter, testing hover, touchdown, horizontal flight, roll, yaw, and pitch control, plus the pitch-up with main engine ignition, followed by a pitch down like it would perform during an abort-to-landing. This would also allow development of ground handling and rapid refueling via service trucks. Then the booster could be flown to explore the flight envelope, along with test flights all the way through re-entry and return, treating it like an X plane project. At the end of that phase, the returnable horizontal-launch booster becomes a working reality.
Once the second such booster is built, you start over with three mated stages, flying them around like a helicopter and testing ground handling, lift offs, hovers, aborts, and pitch ups. Then you do an all-up test and put something in orbit, throwing away an ordinary V1.1 core in the process. The payload will necessarily be somewhat smaller than a regular Falcon 9 heavy. Then you work on returning the core first stage to a horizontal touchdown, mating it back to the strap-ons on the ground, then sliding in a new payload and second stage from the front. If you can do it once a day you’ve met and exceeded all the goals for the XS-1 project.
I suppose the way to summarize the idea is to note that we’ve always launched rockets vertically because it was the method with the lowest parts count, the absolute minimum engine requirements, and personnel and launch support weren’t considered a concern because we were throwing money and people at the problem. With rockets about the size of short or medium range military missiles, the personnel and equipment required for a vertical launch are trivial.
As the rockets start topping 150 feet, they become serious, and as rockets pass 300 feet the equipment and personnel requirements become extreme, with launch facilities worth billions and support staff the size of a small army, riding up and down elevators all day or operating giant cranes to carefully position multi-million dollar pieces of delicate hardware. Add a little more complexity to the rocket so that it achieves vertical orientation as a brief flight phase, and then on the ground the rocket can be treated and serviced like an airliner.
One of the reasons we haven’t ever tried it is that our engines are still a bit trouble-prone and extremely expensive and we’ve generally relied on expendable stages. If the engine represents most of the cost of a rocket and you’re going to throw them away each launch, and rockets are already too expensive, why on Earth would anyone suggest doubling the number of engines, and thus the expense? But as the ratio of touch labor cost to engine costs shifts, the idea gains merit. If you push for really high flight rates, it makes even more sense.
George,
Re: Landing gear.
If you can’t protect wheels from the heat/backwash of the landing rockets, then use flat feet. When on the ground, use a pair of carts on each foot to move the stage around.
Re: Horizontal VTOL.
I have no idea if the maths works for Earth launch, or the engineering (particularly chined tanks), but the horizontal/vertical thruster system would be useful for technology to develop for reusable DTAL lunar landers.
Is there an actual requirement that the booster return to the launch site? Or can it land downrange, be refueled (&re-oxydized?) and then return to the launch site for turn around?
Rodney,
Is there an actual requirement that the booster return to the launch site? Or can it land downrange, be refueled (&re-oxydized?) and then return to the launch site for turn around?
Not explicitly, but I’d say it’s a strongly implied requirement. I’m assuming that Jess won’t go for a demo where you launch and land downrange to the east, then turn around and do the next flight as a launch and land downrange to the west maneuver, as I expect that would be seen as gaming the system. If you’re talking about landing downrange and then self-ferrying back before the next flight, that’s definitely possible. But then you’d need to effectively fly 20 times in 10 days to meet the flight rate requirement (10 sets of an actual launch and a self-ferry home), and that seems really, really aggressive.
~Jon
Re: Landing gear
Well, one of my thoughts was that if there aren’t any wheels, trucks could drive directly underneath the bottom of the vehicle, perhaps lifting it with airbags since the bottom will be pretty wide and flat. Other trucks could be used as convenient work platforms for accessing all the rocket’s subsystems, which would all be located along the underside and accessible through panels (although unfortunately also open through the heat shield). You could come up with some interesting configurations where the heat shield is on top, though.
Here are some rough calculations on a Falcon 9H version of a horizontal lift implementation. Just quickly based on the Merlin 1D’s thrust, specific impulse, 150:1 thrust to weight ratio, a 180 second first stage burn time, and stated 30:1 mass ratio in the strap-on configuration, I’m guessing that the nine Merlins should weigh around 8,800 lbs, the structure weighs 20,000 lbs, and the first stage consumes about 838,000 lbs of fuel. Those numbers are going to be wrong, but somewhat close.
So, if the core and booster stages each have 1.32 million pounds of thrust in their main engines, the two boosters need to both lift themselves and the core (which lacks the extra lift engines), so each booster will need about 2 million pounds of thrust in the horizontal orientation. Assuming the lift engines have a 100:1 thrust to weight ratio, they’ll add 20,000 pounds per booster. If I double the structural weight to account for the small chines, engine mounts, TPS, and landing legs, the dry mass goes from 28,800 lbs to 68,800 lbs and the mass ratio drops from 30:1 to 13:1. That’s almost twice the mass ratio of the Shuttle SRB’s and with a higher specific impulse, and achieved even with a pretty fat margin on adding extra structural mass. Achieving that kind of mass ratio with a conventional-takeoff winged hypersonic vehicle would be extremely difficult, to say the least.
I mentioned that the proposed configuration doesn’t need to do a 180 to slow down prior to re-entry, only a 90, and that the inherently greater pitch authority of the horizontal lift engines makes the maneuver much easier. It also makes it more fuel efficient because a rotation performed with the main engines, given their travel limits, means most of the initial thrust from the mains would still be propelling the stage forward instead of rotating it, because only a fraction of the thrust is providing pure rotational torque.
The same would apply in the landing sequence for a tail sitter. As the stage comes in from downrange, the thrust from the mains would accelerate the stage as they try to rotate it to vertical, instead of decelerating it. So then it would have to burn even a little more fuel to stop its forward travel, then establish a stable hover for touchdown, which of course must be done at zero horizontal velocity for a tall vertical rocket. It might make sense to add some small side thrusters to a tail-sitter just to make the flips more efficient.
The horizontal takeoff/landing architecture dispenses with the final 90 degree rotation entirely. It would just need to bleed off forward velocity, then hover and touchdown (or if it has wheels just slow to a safe rolling speed). Hopefully this significantly reduces the amount of retained fuel required for touchdown, because in this thought experiment my stage has about three times the dry mass of the normal booster. Reducing the fuel requirement for touchdown reduces the wet mass during the flyback, and reducing the fuel requirement for the aero-braking maneuver also helps pay back some of the mass penalty required for the new configuration. A more thorough set of design iterations and studies could determine the more important number for the flyback booster, which is the mass ratio just prior to stage separation. Some of the extra structural and engine mass should trade off with a reduction in retained fuel mass.
Having all those extra engines distributed along the length of the rocket also improves the CG location, which is extremely rearward on a nearly spent conventional booster, so the proposed configuration should ease the aerodynamic problems on the flyback phase. With the CG more centered, and much lower, it even raises the possibility of not flying back directly but hovering over the ocean, deploying some airbags, and just setting down in a quite normal landing orientation. Since the stage will be almost empty and has engines all along the length, only a small fraction of which would be needed for touchdown, the airbags could be placed far away from the engines that are lit. It could be towed back as is, or the stage could be partially refueled from a ship, lifted off (since it would be sitting in the normal launch orientation anyway), and flown home. Salt water corrosion might be an issue, but at least it wouldn’t be bobbing vertically with its engines underwater.
As for abort options the configuration would allow, obviously it could land right after takeoff because it had just been sitting on its deployed landing legs instead of having been carefully balanced and secured next to a launch tower. Then it rises and flips and the main engines start. If there are main engine problems, it can just unflip as it shuts down the mains, then return, hover, and land again. If you included bidirectional propellant cross-feed, you could return intact right up until the moment of booster separation because the boosters could feed off the core stage, flip the entire vehicle 90 degrees for an early re-entry, and fly back and land off the core-stage’s fuel. Since the core stage is nestled in between the two TPS protected chines of the boosters, it could survive the re-entry even without its own heat shield. The extra lift engines also provide a redundant, high-authority trajectory control mechanism during the boosted flight phase, giving even larger safety margins for engine-out or main engine gimbaling problems. The ability to return a payload intact to launch pad, all the way up until booster separation, would probably make the insurers very happy.
There are just a whole lot of benefits to the scheme, and potential benefits, and just about any current launch system could be adapted to it. The downside is buying all those extra engines and components to make the scheme work. Fortunately the idea readily testable on a small scale using a sounding rocket and spare RCS thrusters, so if there are some show-stoppers that I’ve missed, the idea could be abandoned before a whole lot of real money got spent on it. I think there’s enough merit to at least advance to the Power Point stage, if anyone feels like it.
George, what’s wrong with VTHL using a gliding, winged return? The wing mass and landing gear mass only has to be for the empty weight of the vehicle.
Bob Clark
Oh, that would work too. It could also mean that there are “landing” gear with rubber tires and takeoff legs without. I think one of the best reasons to add wings (probably swing wings) is to greatly boost the glide range on the flyback portion.
Taking a page from MIT’s open courseware program on the Space Shuttle, it was suggested that in retrospect they could’ve gone with all-electric flight controls (aka the 787) powered by the very reliable fuel cells and avoided the APU route entirely.
In fact, an odd idea I had for the Space Shuttle might help the glide ratio on the flyback leg.
*** slight digression ***
In the very early design meetings for the Space Shuttle (according to one of the lectures in the MIT open course-ware on designing the Space Shuttle), some of the engineers argued that it would be much more efficient to position the SSME’s under the external tank, which would greatly simplify the stress loads and keep them more in line with the stack’s center of mass. But the engines needed to be reused to save costs, and this meant they had to return with the orbiter. So they suggested having the engines swing from the tank to the rear of the Shuttle after main engine cutoff, and the implication I got was that they would be on some extended arm that would either retract or rotate 180 degrees.
If instead you designed the body flap at the rear of the Shuttle to overextend underneath the external tank, say at a 45 degree angle, like the boarding ramp of a C-130, and the engines would be mounted on the flap. Fully extended, it would meet mating attachment points on the ET, and the fuel lines would feed directly from the ET through the body flap to the engines, holding the flap in place on the pad. The engine thrust would keep the body flap locked to the ET until cutoff (so there wouldn’t be any physical latches that could malfunction, other than for the fuel lines). After cutoff, the body flap would close like the cargo door on a transport plane. Since the engines would be mounted at an angle to the body flap, the wouldn’t be so tall in the closed position, and the closed body flap would have much lower drag than a blunt back end, greatly improving the glide ratio. If would also completely eliminate the mass of piping in the back end of the Shuttle, saving a lot of weight, and the Shuttle structure would no longer have to transfer the compressive loads from the engines over to the external tank, saving more weight.
*****
If you used some kind of clamshell doors around the main (rear) engines on the booster, you could greatly reduce the drag, especially the important subsonic drag. At 52,000 feet you have 10 miles of altitude to work with, and somewhere around there is probably where the booster will go subsonic. If it has a blunt back end the glide ratio is probably going to be similar to the X-15 or Shuttles’s 4:1, so you’d only get about 40 miles out of the subsonic glide leg. If you clean it up the back end to get airliner glide ratios of 20 or 30 to 1 (and the booster won’t have big dead jet engines hanging off of it, so you could possibly beat an airliner if your swing wings had a pretty good span), you get 200 or 300 mile subsonic glide ranges, which are very useful.
As a side note, I mentioned having chines on a future re-usable core stage, which would stretch across the top of the two boosters and reduce wetted area (the core stage would launch upside-down while the booster launch right-side up). You could easily extend the core chine across the top of the boosters and turn part it into a small wing, and then have the wing tips drape down across both boosters, giving you winglets or vertical stabs like the XCOR Lynx or other proposed mini shuttles, without having the winglets freely out in the air stream during max-Q.
The other benefit of having an upside down core is that its TPS system will face much higher temperatures than the booster stages, and having its hot side completely free of wheel wells or access doors is bound to make it easier to engineer.
Nice article here by Jeff Foust on the program with a quote from Jess Sponable about the Mach 10 requirement:
The return of the X-vehicle.
by Jeff Foust
Monday, October 7, 2013
One key requirement of the XS-1 that raised questions is the need to fly to at least Mach 10. At Space 2013, Melroy said that that DARPA selected that versus a less challenging speed, like Mach 3, because there was already efforts underway in industry to develop Mach 3-class vehicles, an apparent reference to commercial suborbital vehicles like Virgin Galactic’s SpaceShipTwo and XCOR Aerospace’s Lynx. A Mach 10 vehicle, she said, “is the bigger reach that DARPA is looking for.â€
Sponable said the Mach 10 requirement also minimizes the size of the upper stage needed to place the payload into orbit. “It means, potentially, that your expended hardware is a very minor part of your total cost,†he said.
DARPA had planned to release the BAA for the XS-1 program this week, after an industry day scheduled for today (October 7) at DARPA’s offices in Arlington. However, late Thursday, October 3, DARPA postponed the industry day because of the ongoing shutdown of the federal government, and a new date has not been announced.
http://www.thespacereview.com/article/2379/1
Also I didn’t know the industry day for this week has been postponed due to the shutdown.
I did write that blog post about the fact that the X-33 could perform this role even if you replace the composite tanks with metallic ones:
Saturday, October 5, 2013
DARPA’s Spaceplane: an X-33 version.
http://exoscientist.blogspot.com/2013/10/darpas-spaceplane-x-33-version.html
This can reduce the cost to space to the $2,000 per kilo, or $1,000 per pound range, a major cut in launch costs. If it had been understood that the X-33 could be used in that role then instead of it just being a demonstration vehicle, then we could have a cut in launch costs to that range a decade ago.
In an upcoming blog post I’ll show the same would have been true for the planned DC-X suborbital follow-on, the DC-X2. Then we could have had a cut in launch costs to that range two decades ago.
Bob Clark
Hi George. Finally read through your posts. You raise interesting questions, some of which are likely to be answered in the next 12 months or so by SpaceX. Of particular interest will be handling the stage in it’s vertical configuration once it has returned. They must have a plan to do it since Elon wants a 24 hour turnaround. Not sure why but he will have his reasons.
Also interesting to note that he intends to develop a fully reusable system, so that’s first and second stages as well has making FH fully reusable. Can you imagine 3 first stage cores heading back to their launch area? Even one’s will be pretty spectacular.
BTW, they intend to release video of their recent attempt.
Guys, you are having an awesome discussion, but I would like to contribute a few thoughts from an armchair rocket scientist 🙂
1) It seems to me, that the Mach 10 requirement is to minimize the size of the upper stage and this would minimize the cost of the non-reusable parts.
2) Landing downrange is not such a bad idea. If you launch from Texas, you can land in Florida. Alternatively, you can take off vertically from a sea platform “uprage” and land downrange.
3) I was thinking, what if we make some superlighweight tanks? Like pressure stabilized tanks made of mylar foil with carbon fibre sandwiched in between? Then expensive parts are guidance and engines. After staging the tanks would be discarded and the propulsion core would land downrange. This a weird twist on Philip Bono’s Rombus concept.
4) Are wings and landing gear really so much of a penalty on a first stage? Composite wings sized for the empty stage will be really light.