I want to clean out some of the Skribit suggestions that I’ve been neglecting for a while. Some of these may end up somewhat brief, but I wanted to at least try to be responsive.
Drop Tanks to Enable SSTO?
I don’t like drop tanks. While they do definitely make the rest of the SSTO easier, they come with several significant drawbacks:
- Large expendable hardware on an otherwise RLV somewhat misses the point.Â While rocket tanks are relative cheap compared to the rest of the stage, they’re still pretty expensive compared to the propellant cost, or even the likely maintenance cost of the rest of the vehicle.
- Dropping stuff on people is usually considered somewhat anti-social.Â That means you’re stuck launching over water like the ELVs (there might be one or two over land paths where the drop zone would be in a safely unpopulated area that isn’t a national park, but it’s unlikely).Â Launching overwater means aborts have a much higher chance of costing you the launch vehicle.Â It also likely limits you to existing ranges, and ties you in with their practices that may not be very conducive to RLV operations.
- You still get some of the drawbacks of TSTOs.Â You still have to design and analyze two vehicles instead of one.Â You still have to design and analyze a separation system.Â You still have to integrate two stages together at the pad instead of dealing with just one.Â You still have to figure out how you handle aborts (if possible) while the drop tank is on.
- Drop tanks reduce the “fluffiness” of the orbiter stage, which may complicate reentry TPS considerations (though maybe not enough to matter).
- The dry weight to orbit savings might not be as much as you think.Â Propellant tanks are pretty lightweight for pump-fed vehicles.Â Mechanical connections, separation systems, plumbing, pressurization systems, quick disconnects all add quite a bit of mass.Â Â How much mass do you really save compared to just building slightly bigger tanks on the SSTO side?
- Drop tanks probably complicate the aerodynamics.Â If you have one drop tank, you now have something like a Biamese vehicle (I’ll get to those in a minute), which has much more complicated aerodynamics, harder abort environments, etc.Â If you go with a lot of smaller tanks, the aerodynamics becomes easier, but your scar weight to structures, plumbing, etc becomes a lot heavier.
That’s not saying they don’t have some advantages.Â Any reduction in dry mass can go a long way towards making an SSTO shift from marginally infeasible to marginally feasible.Â Expendable tanks can be designed to lower safety factors since they don’t need to consider fatigue issues like RLV tanks do.Â Less dry mass means a lighter landing weight, which decreases the amount of landing propellants required and the weight of landing gear (or the wing and landing gear weight if you’re doing an HTHL design).
It’s not that they don’t have advantages, but the disadvantages are enough that I’m not convinced it pays for itself.
Biamese or Parallel Staged TSTOs
Even though my boss Dave is fan of Biamese approaches, I’ve never been.Â He gives me crap about far-out stuff like FLOC, I give him crap about Biamese.
While once again there are enough benefits for Biamese vehicles to make them sound interesting, I think the drawbacks once again win out:
- The engines have to be able to operate from launch altitude all the way to vacuum.Â While altitude compensation can often help TSTO RLV designs, and while there are some techniques like TAN and Flow Separation Control that might make it easier to have an engine operating over that wide of a regime, you’re still making some pretty big compromises to both stages (or you’re making compromises to the biamese concept and losing some of the benefits).Â The upper stage now has to have far more thrust than it would’ve wanted anyway.Â That means more weight that the upper stage didn’t need.Â It also needs some way of operating its space-rated engines at low altitudes.Â This means altitude compensation (which tends to be heavy and complicated), high pressures (which makes the engine harder to develop and more complicated, or something like TAN or Flow Separation Control, which it might not have otherwise needed.Â If the booster and upper stage engines are the same (as they would be in a purist Biamese design), the booster engines now have to have a much higher expansion ratio than they would have otherwise.Â Once again this leads to higher pressures or other complications that they might not have needed otherwise.
- A nearly 50/50 mass split is not very optimal for staging.Â Especially when you figure in that the first stage wants to get back to the launch site.Â It isn’t heinously bad, but it does mean you end up having moderately high delta-V requirements on both stages (since the first stage is likely going to be far enough downrange that you’ll need to do some sort of boostback).
- TPS requirements for the two stages are vastly different.Â For a purist Biamese vehicle this means the booster is lugging around a lot of weight and complexity it really doesn’t need.
- If you don’t do a purist Biamese design (where the two stages are really identical and interchangeable), you’re back to designing two separate stages, but now with all sorts of unneccessary constraints.Â The more divergences you make to simplify things (like going with a lower expansion ratio on the booster stage, or going with a lighter TPS on the booster stage), the more the two designs become different, and the more and more you start losing any real benefit from the process.
- Parallel staged vehicles have uglier aerodynamics.Â Aerodynamic design and analysis for supersonic vehicles can be very complicated and expensive.Â I’ve never seen a TSTO Biamese design that didn’t look like it would be a bear to analyze and design the control system for.
- Biamese RLVs tend to lead to compromised structural design.Â Rocket vehicles are most weight efficient (and easiest to design and fabricate) when they are bodies of revolution.Â In order to get good mechanical connections, most Biamese vehicles I’ve seen end up being lifting bodies, which starts driving either really weird propellant tank shapes (with added weight and fabrication complexity) or really inefficient structures (where the propellant tanks fit inside a more complicated shell.
- Because of the different operating modes of the stages, you’re really stuck still designing and analyzing three different vehicles (the two together, the first stage independently and the upper stage), not just one stage.
- You do potentially reduce the number of engines you need to make, and may allow you to design some subsystems only once, but now they’re being designed to meet more constraints.Â Many times it’s easier to design two slightly different subsystems with 10 constraints each than one with 15.Â You can still reuse a lot of the design and analysis work if you do things right, but each of the two designs are easier.
I guess to me it boils down to the fact that jacks of all trades really tend to be compromised kludges by the time they make it into operations.Â In a Biamese system, both stages are carrying stuff they don’t need, and are being designed to more constraints than were necessary.Â I really don’t see how that will lead to a cheaper system than one that has the two stages scaled the way that performance and operations want them to scale, and that can be more custom-suited for the task they’re being asked to perform.Â Hybrids tend to give you the worst of both worlds.
For those of you not familiar with this concept, the Sea Dragon was an old Aerojet design by Bob Truax for putting 1,000,000lb of payload into orbit on a single TSTO launch vehicle (whose first stage might be recoverable). The design was a Big Dumb Booster, with pressure fed tanks made of maraging steels, built more like a submarine than a rocket vehicle. The first stage engine would’ve been something like 70x higher thrust than the F-1 engine on the Saturn-V. You can get more details here.
I’m a tiny bit more torn on this one than the others, but I still think it makes sense in today’s world.Â It might make sense at some future date, but not right now.
Here’s my big concerns:
- Where’s the demand?Â I don’t think we currently as a species launch a million lb or payload into orbit in a year.Â Until other systems like RLVs get the cost down and the flight rate up there’s never going to be enough demand for more than one or maybe two of these per year.Â While the marginal cost of one of these would be pretty low, the fixed costs and development costs aren’t going to be trivial, and they have to be amortized over those flights, cutting into any cost advantage the design might have.Â Now, if RLVs do get the cost down to the point where you start having enough demand where Sea Dragon could make sense, you run into a different problem–the Sea Dragon is no longer competing against expensive existing ELVs, it would be competing against RLVs.Â Sea Dragon may get stuck only launching payloads where the integration costs of launching them separately and putting them together in orbit outweigh the cost diffrence between the two.Â Now, we live in a world where even though most stuff gets shipped in tiny intermodal containers, there are still Super Guppies and Belugas that get used occasionally.Â In an RLV centric world, there may still be situations where a Super Heavy Launch Vehicle might be useful enough frequent enough to justify its existence.Â But we’re nowhere near that point in time.
- What’s it going to be like developing and testing a 70 Mlbf rocket engine?Â Pintles are a pretty cool, pretty scalable combustion system, but will they really scale up to something 70x bigger than has ever been built before?Â We have no idea what unknowns lurk between here and there.Â Maybe pintles will turn out to work fine without any problems, but we’re pushing far past what has ever been done in the past.Â Â But pintles tend to get worse c* as they scale up, will they still have adequate performance at those scales?Â Nobody knows, and nobody will know until they start.Â That’s scary.
- Testing an engine this big is going to be mindbogglingly expensive as well.Â Every second the engine would be going through about 5 Falcon-1’s worth of propellant.Â That’s only $110k/s of propellant (the upper stage uses much more expensive propellants, so even though it’s only 8Mlbf, it’s still likely going to cost a lot), but that’s not counting anything else.Â You’re talking about $25M per full-duration burn test.Â With how expensive the payloads would likely be for a vehicle this big (see below), you’re likely going to need to do a lot of tests.Â Just the injector testing alone for something like this would likely run you into the $1B+ range.Â If you did even a fraction of the number of runs typically done in a rocket engine project, you’d be talking about billions of dollars up front.Â And where are you going to test a monster that big?Â You’d pretty much have to do it out at sea a long ways.Â How are you going to vacuum test an engine the size of the upper stage engine?Â I guess you can get away with not doing the full nozzle extension tests, but that’s still putting a lot of risk into the first few flights.
- Development costs would be insane.Â Â Between testing the huge engines, and doing at least one or two flight tests, you’re likely talking several billion dollars to develop–if it’s done on a commercial basis!Â The marginal cost of one of these things in 2009 dollars is likely going to be in the $1B range, so that starts adding up fast.Â It’ll be a long time before there’s enough demand to justify putting up that kind of money.
- Payload costs per launch would likely be very high.Â While I full-heartedly agree that relaxing mass constraints can reduce the cost of space payloads, it’s only one part of the cost involved.Â Being able to go with welded steel and an FOS of 3 may reduce some design and fabrication costs a lot, you’ve still got the fact that unless you’re launching bulk commodities, you’re designing hardware that has to operate in a very harsh environment, that still needs to be fairly complex, and which very few will be built.Â People talk about stuff like being able to launch an ISS in a single launch.Â While avoiding all of the EVAs and integration stuff would take a lot of work out of ISS, you’d still be talking about a several billion dollar payload.Â How would you get launch insurance?Â How often can you afford to fly a vehicle that costs $500M-1B to launch and has payloads that will tend to cost several times more?
Unlike the other two ideas, I’m not convinced that super heavy launch vehicles will never have a place in the rocket world.Â I’m just not convinced we’re even within visual range of such a time where they make sense.Â We’re still back not too far past the Wright Flyer stage of launch vehicle design.Â We’re nowhere near the point where a Beluga or Super Guppy makes sense.
Latest posts by Jonathan Goff (see all)
- An Updated Propellant Depot Taxonomy Part VI: Roving Depots - February 22, 2021
- An Updated Propellant Depot Taxonomy Part V: Human Spaceflight Fixed Depots (Low-Orbit) - February 16, 2021
- An Updated Propellant Depot Taxonomy Part IV: Smallsat Launcher Refueling Depots - November 14, 2020