This third installation in my Orbital Access Methodologies series (parts I can be found here, and part II here) has been a long time in the coming. It has taken so long, not because I’ve been spending months researching and analyzing the topic (I knew most of what I wanted to say back in January), but mostly because I was surprised by how much favorable attention the first part received, and I’ve been worried about not meeting expectations. A good part of the reason why that first article was so good was that I was able to lean heavily on help provided by Dan DeLong and Antonio Elias, both of who had been analyzing air-launched orbital access methodologies since I was still in gradeschool. I now have a bit more empathy for movie directors trying to make a sequel or a prequel to a first movie that had been far more successful than they had ever thought.
In the previous installation, I discussed approaches to incrementally make ELVs more reusable (or at least recoverable/refurbishable). I discussed why I think that while making ELVs recoverable will be an improvement over the state of the art, such incremental improvements may actually be on a different evolutionary path from high-flight rate capable, truly reusable launch systems. I then discussed the key challenge for TSTO RLVs: how to get the first stage back after a mission, and I outlined the benefit of having the first stage be able to return itself to the original launch site without having to land downrange. This article and the next several in the series will focus on TSTO approaches that provide for return to launch site capabilities.
The first of these approaches, what I like to call “Pop-up TSTO”, has gained quite a bit of attention over the last several years, particularly due to Patrick Stiennon and David Hoerr’s book “The Rocket Company” (which they had me review here, and here). The basic concept is to have a TSTO vehicle, where the first stage flies up purely vertically (John Carmack, who is a fan of the approach has likened the first stage in this concept to a freight elevator) with an apogee of around 100km, the second stage separates from the first stage, and then the second stage provides all of the horizontal acceleration to reach orbital velocity. The first stage reenters and lands vertically like the suborbital vehicles that we at MSS, as well as our friends at Armadillo Aerospace, TGV, and Blue Origin are trying to do. The upper stage after delivering its passengers or payload, reenters and also lands at the launch site.
Benefits of the Pop-Up TSTO Approach
There have been several benefits posited for this TSTO approach:
- The vehicle is very operationally simple. The first stage goes straight up, the second stage straight over. You have at most four important engine ignition events (liftoff, 2nd stage ignition, 1st stage landing, and upper stage landing if the upper stage uses powered landing).
- If the upper stage T/W ratio is high enough (approximately 1.4) or if the first stage staging altitude is high enough, the first stage ends up soaking up most or all of the typically 1600m/s of losses that an SSTO design would face. This means that the upper stage only has to provide the ~7800m/s needed for orbital velocity, minus ~325-465m/s for the rotational velocity of the earth depending on launch site latitude, yielding a required delta-V of around 7400m/s for most US launch sites.
- The upper stage main propulsion system only has to operate in vacuum, so all of the engines can be vacuum optimized, giving much higher mission averaged-Isp.
- The upper stage also doesn’t operate for the most part inside the atmosphere, so it might not need slosh baffles (or if it does, they probably don’t have to be as heavy as baffles needed on a lower stage). It also probably doesn’t need anywhere near as much gimbal authority as a 1st stage would.
- Staging can be done at high enough altitude that it is a very low dynamic pressure event. Part of what caused the loss of the last Falcon I flight was that the staging ended up occurring at a lower altitude than planned, which imparted higher aerodynamic forces on the stages, which caused a collision between the 2nd stage nozzle and the first stage.
- The first stage ends up having performance requirements more like a suborbital launch vehicle than a typical orbital first stage. This means that it’s easier to make it robust and simple, costs can be lowered at times by throwing weight at problems (since the first stage is very weight insensitive). This also means that the first stage could be either evolved from a future suborbital launch vehicle, or at least could possibly be developed by a team that has worked out the challenges of a VTVL suborbital vehicle.
- Since the upper stage has such a high delta-V requirement, it will end up having a relatively high propellant mass fraction, which means that when it reenters, it will be mostly empty and will thus be very fluffy. Having a low ballistic coefficient (ie a low mass per unit frontal area) means that you decelerate quicker, higher in the atmosphere where the density is lower–this yields both a lower peak g-loading, but also a lower heat flux, thus making the TPS material challenge somewhat easier than for a dense reentry vehicle like the shuttle or most capsules.
- Since the first stage has no downrange velocity, it’s Instantaneous Impact Point stays right around the launch site throughout the flight. This makes it easier to launch over land, out of more populated areas (instead of having to launch along the coasts or from islands or sea platforms out in the ocean). Most of the high-risk phases of flight (ignition, max-Q, staging, upper stage ignition, etc.) happen when the IIP is within spaceport grounds, and thus away from the uninvolved public. This should make it easier to get licenses for the vehicle to operate out of less traditional launch facilities, which may be a key to lowering some of the cost of space access–and to being able to get more customers for said vehicle.
Now, there are probably other advantages, but those are some of the primary ones as I see it.
Challenges, Constraints, Limitations and Drawbacks
Like with the Air-Launched “Assisted SSTO” concept I discussed in Part I, the Pop-up TSTO approach does not come without its own set of problems. There are always both pluses and minuses to all approaches, and the key to good engineering is to make sure you understand what those limitations really are so they can be dealt with properly. Here are a few of the main drawbacks that stick out to me:
- Much like the air-launched SSTO rocket stage, the upper stage for a Pop-up TSTO vehicle still faces a nearly-SSTO level of delta-V requirements. Due to the non-linearity of the rocket equation, knocking off 1600m/s vs. a ground launched SSTO makes a huge difference, but providing 7400m/s in a single, reusable stage is still challenging.
As an aside, many commenters on my air launched SSTO concept seemed to think that such a vehicle wasn’t really technologically doable, but that a Pop-up TSTO stage would be relatively easy to build. I stayed up till 2am doing the math last night, and the fact is that the two are not as different as you might think (I can provide some of the math and explanations if people are interested). The Air-launched SSTO stage needs about 8000m/s (maybe 100-150m/s less for a stage using a more dense propellant combination, or one that has a high thrust to weight at ignition due to using Thrust Augmented Nozzles), compared to 7400m/s for the Pop-up TSTO upper stage. What this equates out to is that for two stages using similar propellant types and similar propellant loads, the pop-up upper stage would only have 20-25% more mass to play with than the air-launched SSTO stage. Specifically for a LOX/LH2 upper stage, you’re talking about propellant mass fractions (the propellant mass divided by the stage plus payload mass) in the range of 0.81-0.82 for the pop-up stage, and around 0.84 for the air launched stage. For LOX/HC, the numbers are around 0.89-0.91 for the pop-up stage, and and 0.9-0.92 for the air launched stage. While that 20-25% more dry mass is nothing to sneeze at, it’s a lot closer than most people would seem to believe.
- The upper stage needs a relatively high stage thrust to weight ratio at ignition in order to avoid incurring drag losses (around 1.4 being ideal according to The Rocket Company). While you could theoretically loft the first stage a bit higher to give more time, this quickly starts putting your abort g-loads in the range that is problematic for manned flights. So, you either end up taking a small delta-V hit (thus pushing you closer to the air-launched SSTO case), or you end up taking a mass ratio hit for larger engines.
- The upper stage ends up being very sensitive to weight growth. Adding 1 pound to the upper stage could require an additional 20-30lb worth of hardware and propellants on the first stage. This either means designing in lots of performance margin on the first stage, taking a hit to payload, having to spend a lot more money on weight control on the upper stage, or possibly all of the above.
- The high delta-V requirements, and the sensitivity of first stage weight to upper stage weight growth push you towards LOX/LH2 or at least LOX and one of the lighter hydrocarbons (cryogenic methane or subcooled propane) for the upper stage. This is typically done by the ELV people as well, but the complexity of adding a cryogenic fuel on-board is annoying.
- The typical configuration for a pop-up TSTO is going to be two serially stacked stages, which now requires ground handling equipment for stacking stages. This costs money and makes it harder for a given location to setup a launch site.
- Because the delta-V split on the stages is less than optimal, this results in very big first stages (depending on the achievable propellant mass fractions). Which means that as you scale up, at some point you’ll wind up with a stage that’s too big for normal ground transportation. And because RLVs will typically have a much lower payload to GLOW ratio than ELVs, you’ll run into this road/rail transportability limit at much smaller payloads than ELVs do. For instance, if you don’t go with a LOX/LH2 upper stage, even a very light RLV (1-2klb payload) could end up having a first stage that’s as big as a Falcon IX first stage.
There is one possible work-around to that problem–and that’s having the first stage be modularly assembleable. While I think John takes the modularity concept way too far (I’d never go more than 7, and would generally try to keep it to 3-4 parallel units), and while I’d definitely go with a more aerodynamic module configuration with higher aspect ratio modules than he has, modularity could possibly help with getting around this problem. Think Saturn-IB first stage except having the separate tanks modularly assembleable, instead of preassembled. Sure, it’ll cost you a lot more integration, and a lot more mass for the mechanical, fluid, and electrical interconnects, but your first stage is already fairly weight insensitive. This would allow you to scale up by at least another half order of magnitude, and by that point you’re probably up into the light EELV range–which RLVs won’t be approaching in the near term anyway.
- You’ve still got to deal with TPS for the orbital stage.
- Because the most likely configuration for a pop-up vehicle is two vertically-stacked stages, the upper stage may need to be able to separate itself from the lower stage in some abort modes. While HTHL vehicles can more readily survive propulsion failures at most points in their flight, VTVL vehicles like the pop-up TSTO would likely be don’t have the option of just dumping most propellants and gliding down to an emergency landing. If you have a full propulsion failure of the first stage, it may require separating the upper stage in a hurry. Since this a reusable stage though, typical expendable launch towers aren’t a practical answer, which involves some sort of reusable escape engines (possibly an aggressive TAN extension to the upper stage primary propulsion system). Testing these and making these abort modes safe and graceful is going to be non-trivial.
Being a less aggressive design approach than the Air Launched SSTO, there aren’t as many enabling technologies that aren’t already on the shelf. Thrust augmentation could possibly be helpful (especially for emergency abort operations), but aren’t necessarily required. Composite propellant tanks and structures could reduce the weight of the upper stage a bit, making it easier to hit mass targets, but the upper stage is probably within the realm of feasibility even using metal tanks and existing manufacturing processes. The first stage development and operations would benefit from the existence and flight experience provided by suborbital VTVL RLVs.
The main enabling technology for this style of RLV is going to be the TPS system (and possibly the reentry technique). There are a couple of interesting options out there that might be doable with such a fluffy reentry stage, such as metallic TPS like was planned for Dynosoar or X-33. And there are some more exotic ideas I’ve heard such as Joe Carroll’s “spike” idea. But the reality is that none of these have been proven out yet, and that’s the only real enabling technology for Pop-up TSTOs that isn’t already on the shelf. It’s important to note that this is the case for all of the RLV techniques I’ll be talking about. There are tons of good ideas, but very limited flight data.
Also, looking back at what I said in Part I, all RLVs could benefit from commercially available prox-ops tugs.
Remaining Unknowns and the Path Forward
Unlike Air-launched SSTOs, there are far fewer unknowns that I can see for this approach. The upper stage is still fairly aggressive, so there’s some questions about if we can make a highly reusable stage with the required performance. There’s still the questions about the TPS. And the other big unknown is going to be how to handle aborts throughout the flight regime. In order for an RLV to make economic sense, you can’t be losing it frequently. Just getting the crew, passengers, and/or cargo out isn’t enough if you can help it. Figuring out how to design a reliable VTVL vehicle that can survive reasonable failures is going to be a challenging task. And figuring out how to perform a rapid separation in possibly adverse conditions without adding so much mass or complexity to your upper stage that you make the vehicle less reliable or unworkable is also going to take work.
The key to moving forward though I think is pretty clear. VTVL RLV companies like us at MSS and our friends at Armadillo and the others need to keep plugging along until we are actually reaching 100km on a repeatable and affordable basis. We need to keep working our way up the learning curve, and hopefully finding businesses along the way to make that possible. Once we’re there (or possibly sooner if XCOR or Virgin beats us to space–which is actually fairly likely), subscale TPS experiments need to be done using suborbital vehicles. This can be done using the “nanosat launcher” suborbital RLV upper stage I mentioned in Part I. By decreasing the cost of actually getting real flight data into the hundreds of thousands range might allow for enough iterations to work out some of the bugs on the small scale before trying to build a full-scale prototype. Also, once a VTVL suborbital vehicle is there, most of us in the industry plan on trying to use our vehicles as a first stage for launching nano-sats. This should help work out the challenges of stage integration, staging, and could even provide an environment for testing out subscale launch escape systems and techniques.
Once all of the subscale work has been proven out with suborbital vehicles, it should be much easier to start into developing a prototype orbital vehicle. There’ll still be a lot of work involved, and there will still be some scaling risks, but by using suborbital vehicles to prove out the various concepts, a lot of the important risks can be retired before its time to start work on a full-up orbital RLV.