YHABFT: The Danger of Insider Euphemisms: “Shooting the Puppy” Edition

I’ve noticed that business people and engineers tend to use a lot of jargon and insider euphemisms in their conversations, sometimes with rather humorous results. I had a recent example occur that was too funny not to share (but too long for Twitter).

Back at the end of 2011, my startup hit a really bad dry spell with no contracts for about 6 months, and we were really close to running completely out of money. I went to ask a local space entrepreneur for some advice, and he he used the analogy of “shooting the puppy,” to metaphorically describe the painful decision of whether it would be more prudent to shut a company down intentionally and mercifully rather than letting it die slowly and painfully. Fortunately, we decided not to “shoot the puppy” in this case, and a few months later landed two contracts with the DARPA Phoenix program. Ever since then though, we started using the phrase “shoot the puppy” as a euphemism for making a tough business decision to end some project we were emotionally attached to.

Which brings me to the humorous situation. I was recently talking on the phone with a colleague who was flying back from a business trip. He was going to help me with some artwork for a long-shot proposal that I was really excited to bid on. Unfortunately, we ran into some snags and I was starting to wonder whether it would be more prudent to “shoot the puppy” and focus on finishing up a few other proposals instead that I felt were more likely to win even if they weren’t as big and awesome. I still needed one key data piece before I could make that decision, so I told him that I’d figure out whether or not I was going to shoot the puppy. You know how sometimes when you’re talking on the phone you can start speaking rather more loudly than you would otherwise intend? Well my friend was in that mode when says something to the effect of “Ok, cool. Go and figure out if you’re going to shoot the puppy. But let me know what you decide. It’s probably better for me if you shoot the puppy, but let me know one way or another.”

Mind you, he’s saying this loudly in the gate area of a major international airport waiting for his flight. I can only imagine what sort of sociopath the people sitting around him must have thought he was. Fortunately, nothing worse happened than him being really embarrassed when I pointed out that that euphemism probably wasn’t the best one to use in a public place where people don’t know the context. But it just goes to show that we should probably be more careful about our euphemisms and jargon, and try to think about how they might sound to the uninitiated.

Posted in Humor, YHABFT | 5 Comments

Post a Day

john hare

In helping Jon complete the post a day for a month celebrating 10 years of selenian boondocks, I found I’m not very good at nonspontaneous writing. When I have an idea that I’m passionate about and I write it up during the enthusiasm of developing it, it’s a lot of fun. Trying to create a post a day is the first time I’ve ever written to a deadline. It’s very definitely a different atmosphere. It seems to me the posts are less convincing and probably less fun for the reader.

One reason for less enthusiasm now is I’m not really doing anything physical to move spaceflight forward. I tend to be a participant in things I’m interested in and hardly ever a spectator. Like I tell people about watching porn or sports, that ain’t you so why bother. I want my epitaph to read participant not spectator, and I tend to of been more of a spaceflight spectator than not over the last several years. Finances tend to control ability to participate in space as well as anything else, and since the recession my finances have been quite limited for anything other than keeping my business operating and living my life.

Of the many ideas that I have about propulsion, spaceflight, and making it all happen, virtually all of them require seed money that exceeds what I have readily available. So my focus has been on my business with the idea that if I develop the financial capability I can push some of these ideas forward. Is not a tremendous amount of money to develop demonstration prototypes in the aerospace world, but still considerable to an individual.

I will still write up a few ideas when they occur to me and I’m in the mood and have the time to write, but it is unlikely that I will try to maintain any type of set pace for publication.

Thank you everybody that put up with my latest fumbling around.

Posted in Uncategorized | 2 Comments

The Best of Selenian Boondocks’ First 10 Years

June 16th this year was the 10 year anniversary of starting Selenian Boondocks. It’s been a great 10 years. We’ve had a lot of good contributors, including myself, Ken Murphy, Kirk Sorensen, and last but definitely not least, John Hare. We’ve discussed a lot of fun topics. We’ve even started, built, or perpetuated several space policy or technology memes. And we’ve had a lot of fun without taking ourselves too seriously along the way.

In celebration of this anniversary, we did a blog post a day over the past month. In a way that was unintentionally “meta”, this started out with me blogging up a storm, and then when my life got too busy for a few days (I’ve been on a family vacation in Yellowstone and the Pacific Northwest), John Hare picked up the slack with several days of his blog posts.

Tonight I just wanted to post links to several of my favorite Selenian Boondocks posts or series from over the years. This is far from an exhaustive list, but these are the blog posts I think of the most when I think of what we’ve done over the years.

  • Orbital Access Methodologies: This was probably our most popular blog series, where I discussed a range of approaches for doing reusable launch vehicles, including Air-launched SSTOs (ala Dan DeLong’s Orbital Spaceplane), and a range of various TSTO options including “pop-up TSTO” (ala The Rocket Company or many of John Carmack’s old concepts), “glideback TSTO“, and my two current favorites: “Boostback TSTO” (similar to what SpaceX is trying to do with F9R, and what Masten, Blue Origin, and several others have looked at for reusable orbital vehicles), and “Air-Launched Glide-Forward TSTO” (first suggested to me by John Hare, and then expanded upon in my still uncompleted Boomerang TSTO RLV series).
  • Venus ISRU Series: This was another popular series, which is also unfinished. Venus just doesn’t get much love in space settlement circles, and this series was my attempt at trying to discuss the potential of Venus as a destination for human settlement. My favorite posts in this series were: this post where I describe what materials we have to work with, Venusian Rocket Floaties where I discuss the counterintuitive realization that most rockets would actually float like dirigibles in the Venusian atmosphere, these two posts describing ways of extracting and separating condenseable species and gas-phase species from the Venusian atmosphere, and one of my all-time favorite humor posts about Venusian Acid-Cooked Turkeys (thanks to George Turner for restoring some faith in the value of having a comments section).
  • xGRF (Variable Gravity Research Facility) Series: This was a series of posts discussing what I still think is the best approach to answering the question of how much gravity humans need to live and thrive. The first post describes the concept (initially conceived by coblogger Kirk Sorensen, while he was at NASA). The latter two posts describe ideas for how to implement this for less cost using commercial crew assets such as Dragon V2, and how to retire technical risk for the tether portions of the concept using a series of low-cost cubesats. I’ve been coming around to the idea that something like Cygnus might be a better platform–I think the key to making this happen though is finding some way to do this experiment for low 8-digit costs, leveraging ISS assets without unduly impeding other research on ISS.
  • RLV Markets: Another uncompleted series about different aspects of markets for low-cost RLVs, and how they might differ from the markets for ELVs.
  • My Top 10 Technologies for a Spacefaring Civilization I still agree pretty strongly with most of these items.

A few other more minor posts of mine that I think are still interesting (I could probably list 20-30 of these, but will only list a few):

Sorry if that list is almost entirely my own posts. John, Kirk and Ken have all done many great posts, I just have an easier time remembering my own posts. In the comments, I’d love to see recommendations for other good posts we’ve done, including ones done by John, Kirk, and Ken Murphy.

Looking forward to continuing interesting discussions during the second 10 years of Selenian Boondocks!

Posted in Administrivia | 4 Comments

Lap Construction

probationary second string substitute apprentice relief bloggers’ helper in training john hare

Everybody that reads my posts knows that I think most people get way too complicated with launch assist platforms. In my post last week I suggested a really inexpensive platform that was a flying wing powered by a pair of very large cage Jets. I didn’t justify how I felt it would be such low-cost. I’m going to try to do that in this post.

The cage Jets I suggested would be on the order of 60 feet in diameter. This large, they would have an RPM on the order 600. This large, they would be of a size useful to power plants. If they were useful for power plants, then there would be enough production to get the cost down. It is ironic that the larger the engine of this nature, the easier it is to maintain clearances and margins.

The blades would all be a single profile which could be extruded, cut, and locked into the wheels. By making the blades and all the other parts very simple, cost comes down. By making it very large, inspection is by people walking around inside the engine checking for problems. Maintenance is mechanics with large wrenches and not technicians with superhigh tech computer-controlled gadgets.

Other than the high thrust to weight of these engines which is critical, and potential very low cost, the major advantage a for launch assist platform is that power can be pulled from any part of the circumference of the engine. This means that during takeoff and landing some air can be bled into the plenum chamber of the hover landing system. Some thrust can be straight down from the nose when rotating for liftoff. During low-speed flight a large quantity of air can be directed from the wingtips for an air curtain virtual winglet of very large size. This is to allow the low aspect ratio wing to operate with any efficiency approaching that of a medium aspect ratio wing.

Construction of the aircraft that is the launch assist platform was more or less hand waved in the last post. I do not specify exact construction techniques other than to say it should be something simple cheap and easy for the available construction force and supply chain.


In the cartoon above is a quick sketch of the vehicle from the rear. I see it as a truss layout in both directions big enough to allow construction workers to walk inside and out of the vehicle while under construction. The circle on top being the launch vehicle.

My mental picture was of it being an aluminum structure assembled in a manner similar to the steel buildings we see go up every day. Semi trucks would deliver large truss assemblies that cranes and forklifts would place in the designated area to be bolted together by a construction crew. With good design a construct this size could be assembled by modest crew in well under a month. Then I see sheeting coming in again on semi trucks, with cranes and forklifts lifting it up to be riveted or screwed onto the main structure.

As I said in the original post however, construction should be whatever is most comfortable and familiar with the designer and crews available. George suggested wood and fiberglass. That is certainly feasible as the Mosquito bomber of World War II was made of wood and capable of well over 400 miles an hour. We’ve had 70 years to improve on that wooden technology. It is possible the vehicle would be best made out of plain steel with steel trusses and sheeting assembled by workmen used to dealing with that material. The skin could even be corrugated metal as in the Fokker Tri motors and most famously by the Junkers JU52. When we’re only looking for 400 to 500 miles an hour, there are many construction techniques that might work. The controlling factors not get so high tech and overdesigned that the cost goes through the roof.

The direct material cost using aluminum, steel, wood, or even fiberglass would be under $1 million for this very large vehicle. With proper design, labor cost for assembly could be under a million also. The cost of the basic structure would be dwarfed by the cost of the engines, crew modules, and control systems. The way to hold the cost down would be to build a subscale model first, and test it out not only for flying capability, but also for ease of construction and the cost thereof. If there is a glitch in the small UAV, you address it by building another UAV of similar small-scale to address the problem. Once problems are handled is that scale, you can build a quarter scale model that could possibly launch a rocket vehicle capable of one ton payloads to orbit. Only after that works out the you begin construction of the larger vehicle.

It may turn out that the smaller vehicle is profitable enough that you don’t need the larger vehicle for quite some time. It also may turn out that you become internally funded to the point that you do not have to go begging for vulture capital funds. It is also possible, that the entire concept is found to be flawed, in which case you can stop before investing enough money to sink your company.

A follow on for later that would be a bridge too far today, would be a supersonic launch assist platform. If the cage Jets have the capability that I suggest that they do,Mach 3 to Mach 5 is eventually attainable. If it becomes desirable to have this capability, then it may be that an incremental upgrade is a way to attain it.

Posted in Uncategorized | 5 Comments

SpaceX Downselect by Senate?

Second string substitute apprentice relief blogger in training john hare

The various antics of our elected officials in Washington tend to bring into question their motives and loyalties. The commercial crew  cuts by the house that were cut even further by the senate seem to be a gambit to downselect to a single provider. Boeing being the likely selectee.

One problem with this is that commercial crew is heavily politicized. A down select to one will cause outcry from the other. So there will likely be some form of compromise unless the full senate increases the item toward the original request. If the commercial crew budget stays low however, and they try to split it up in some seemingly fair manner, Boeing will push the schedule out by years. If pushed out enough years, their participation will become visibly non viable. Somewhere around that point, Boeing could be encouraged to walk away from commercial crew as too much trouble. That would be more or less a voluntary downselect. If there is a  functional downselect to SpaceX this way, seemingly by accident, the senate and house could in good conscious reduce item funding to $600M or so and brag about the savings even if it basically happened by accident and opposite intent.

Could be quite entertaining across the next several years.

Posted in Uncategorized | 4 Comments

Custom LAP

Second string substitute apprentice relief blogger john hare.

In my last post I described a new type of turbojet based on my cage jet of years ago. The engine I described has the capability of good thrust and good fuel economy which is ideal for launch assist platforms. Launch assist platforms want to have the capability of lifting very heavy loads off runway, and taking them to very high altitudes, and pitching up in a gamma maneuver that allows near vertical launch of the orbital vehicle. Sometimes they want to cruise to a particular launch location and cruise back to base.

In order to reduce upfront investment, most of us start by looking at existing aircraft that can be modified for our purposes. That is almost certainly the way to get started, but has the problem of limiting our capabilities to that of whatever carrier aircraft is selected. The problem with designing our own Launch Assist Platform Aircraft, is that it adds a tremendous amount of expense to a project is almost always funds limited. To date, the launch assist platform aircraft that have been designed have been designed by aircraft designers that are going for extremely capable aircraft, but don’t seem to have much input from the launch industry. The White Knight series of lifters exemplifies this.

I suggest what should be done is design the aircraft around the launch vehicle, instead of vice versa. We should also design around available finances, skill sets, and available ground facilities.

First thing is the launch vehicle payload required, which defines the rocket vehicle, and dictates the capabilities of the Launch Assist Platform Aircraft. It is necessary that we make an assumption about the maximum payload that this system will will want to place into orbit. For the purposes of this blog post, I am going to make the assumption that it is desired to place 25 tons in orbit as a maximum payload. While this is much less than the heavy lift vehicle’s several companies are considering along with NASA and the United States Congress, it is quite sufficient for almost any mission we see in the next decade, as long as we assume orbital tugs and propellant depots. By developing  the launch assist platform now with its attendant launch vehicles, a revenue stream can be developed first, which can then be enhanced by the orbital tugs, and the propellant depots.

Designing the launch assist platform aircraft, is much like designing the foundation for a multi-story building. When designing a building you do not start with the foundation, you start with a roof. Then you design the top floor which also carries the loads of the roof, then the second from the top floor, all away down to the basement. Only after all that do you design the foundation of the whole building. In a similar manner we have to work backwards from the payload to the launch assist platform aircraft. If we assume a basic launch architecture of launch assist platform, and single stage from there to orbit, the mass ratio can be on the order of 12 with high-performance kerosene engines. The dry mass would be on the order of 4% each for payload and vehicle.

4% net for a payload of 25 tons gives a rocket vehicle of 625 tons. This becomes the desired payload of the launch assist platform aircraft. This is clearly beyond the capability of any existing aircraft including the White Knight 3. This is the technical requirement based on my assumptions.

Available finances dictate the actual capabilities we will end up with. Trying to design a conventional aircraft with the capability of 625 tons in external carrying capacity is not going to work. There’s not enough work for that vehicle to use on other projects which means that the launch assist platform aircraft must carry the entire burden of cost simply on launch revenue. Available finances are the funds that can be spent on the vehicle considering ROI, and not based on some percentage of a billionaire’s net worth, or how much money can be conned from the United States Congress. The 25 ton payloads, at the pricing that can be expected a decade from now when the system would hit its’ prime dictates how much money could be spent now if we assume that the system is flying at least daily. Since it could be competing against $500 a kilogram or less from other companies, the finances suggest a gross revenue of about 12 1/2 million dollars per flight. Subtract fixed and marginal costs from that number, and multiplied by the number of flights expected annually, and we get a number 10 years out that we can work backwards to find the amount of money available today. Since the LAP is only one component of a two unit system, it is only worth  a percentage of the total. The rocket stage will get the lions share of the costs and investments leaving perhaps 2 million per flight available to service the debt on the LAP after its’ own fixed and marginal costs. Assuming a flight rate of 250 per year, and revenue available for debt interest is $500M per year. A high risk debt can be expected to have an effective interest rate on the order of 25%. So the vehicle debt at that price range and interest can be no more than $2B.

If we assume that the initial investment covered a development time of six years, and a further four years was spent ramping up business, and the interest on the development money was at 25%, then there would be something on the order of $200 million available to develop the launch assist platform. The only way this can possibly be done for that number is if the vehicle though very large is very very simple.

The second requirement is to design the launch assist platform around the available skill sets of the people available to the project. Since this is a blue sky concept, I am going to assume that reasonably competent but not brilliant designers are available, along with a workforce that is motivated and experienced at the construction method under consideration. This requires that the construction method under consideration be very simple.

The vehicle must also be designed around available facilities. This is fairly simple, runways and available hangers will limit the design. Runways have length, width, and weight limitations. Hangers have length and width limitations unless you build a fancy and very expensive new structure. Since the 625 ton upper stage will probably be matched by a 625 ton launch assist platform, the runway must have the capability of handling 1250 tons. Since this exceeds any aircraft ever built the weight must be distributed over wider areas that any aircraft landing gear has ever experienced before. 1250 metric tons is 2,750,000 pounds. 2,750,000 pounds can be accommodated by using a hovercraft undercarriage of the type that was experimented with 50 years ago. If we assume a very high wing loading, there will be something toward 30,000 ft.² of wing area. It will take a very low aspect ratio wing to fit in the available facilities. The aspect ratio will probably actually be around 1.5.


In the cartoon you can see the hammerhead shroud hanging over the front of the vehicle. The cage jets are inside of the wings. And instead of wheels underneath there are hovercraft skirts to spread the load across the whole runway.

The way I see it this launch assist platform will be a flying wing with a wingspan of about 200 feet and a length of about 200 feet with sweep to wingtips are still 100 feet long. The launch vehicle will ride on top of the wing centerline. The hammerhead shroud will protrude in front of the vehicle. There will be a huge cage jet mounted inside each wing. Each cage jet will mass about 40,000 pounds and have a thrust of 1,000,000 pounds. The airframe should be around 10% of takeoff mass and will be about 125 tons for airframe. With engines and airframe at 165 tons, and other required systems at 35 tons, there will be about 425 tons of fuel available to cruise and accelerate. Enough fuel will have been burned by the time of the gamma maneuver, that the vehicle can accelerate at a fairly high rate during the gamma maneuver on cage jets alone. The rockets on the launch vehicle can be lit before separation allowing rocket systems checkout during the maneuver. When the vehicle separate the launch assist platform will have a higher thrust to weight ratio than the rocket, which will allow it to accelerate away without worrying about rocket plume impingement.



Posted in Uncategorized | 3 Comments

Cage Truejet

Second string substitute  relief blogger john hare.

Some years ago I did a few posts about an air turborocket with the bladeing based on the squirrel cage fan. October of 2008 if you’re interested.  Some varieties of the squirrel cage fan have blade geometries that are simultaneously useful as compressors and radial inflow turbines. By using the blades as compressors on 75% of the cycle and as turbines on 25% of the cycle, 100% of the incoming air regeneratively cooled the blades so they could run a considerably hotter turbine inlet temperature than normal. The higher the allowable temperature, the higher the available performance. The other benefit of this blade geometry is that all moving components were on a single wheel, which allows for massive weight reduction compared to conventional turbine based engines.

The downside of the concept is that the cycle doesn’t close. Using the same blade for outflow compressor and inflow turbine means that the turbine inlet pressure must be considerably higher than the compressor will deliver. It was only as an air turborocket that the concept works as originally conceived. As an air turborocket though, thrust/weight ratios of 25 are quite attainable with specific impulses of nearly a thousand. By adding multiple wheels the specific impulses were somewhat closer to that of turbojets, though not turbofans. Adding extra wheels was still like adding epicycles to make the concept attractive.

The cagejet turborocket will always be a niche engine if it ever gets built. Thrust /weight will be far less than rockets, while fuel economy will be worse than turbojets and far worse than modern turbofans. Attractive for Launch Assist Platforms that want high acceleration for limited time in the atmosphere, but not for the long cruises some of these platforms want. Also attractive for some limited military applications.

It was called to my attention a few years back that perhaps I was too focused on reaction turbines when impulse turbines were useful in some applications. An impulse turbine can be a bit less fussy in the bladeing in exchange for considerably more critical nozzles to drive them. If I could use the impulse turbine concept, it might be possible to work a radial outflow turbine with the same blade that is a compressor on the rest of the cycle.  If this can be done, the cycle might close an allow an engine that doesn’t require a rocket to drive the engine. The elimination of the oxidizer turns it into a very light turbojet with high thrust/weight.

A second thing pointed out to me was that the squirrel cage blades were speed limited by the mach number at the leading edge of the blade. This is the same problem of centrifugal compressors. The speed limit forces the inlet area down in relation to the wheel diameter. It also puts a limit on available compression ratio.  A compression ratio of 2 is respectable for an air turborocket, and insufficient for a turbojet. By putting a fan in the inlet plane of the cage, it is possible to power prewhirl the incoming air so that the cage blades can run faster. Double the possible compression ratio faster. So I added a fan that prewhirls the incoming air, but also creates some compression in its’ own right. Compression ratio of 4 is now possible which is barely in turbojet country.

The third modification to the concept is the fuel handling. By using blades with fuel passages and film cooling holes that are also fuel injection holes, it is possible to get very fast mixing, while also using the fuel to regeneratively cool the blades as well as supply film cooling to them. With the blades being cooled by 100% of the air during 75% of the cycle, and simultaneously cooled by 100% of the fuel during 100% of the cycle, it is possible to run this turbojet at stochiometric mixtures without damaging the blades. This allows a high thrust/weight ratio from the turbojet even with a compression ratio of 4, and eliminates fuel hungry afterburners.

cagetruejetThis side view shows the incoming air in light blue  that goes through the compressor/turbine blades into the volute for pressure recovery. From the volute into the burner that is mostly not shown except for the section close to the turbine nozzles. After the burn the hot gas is through the turbine nozzles to the turbine blades. Into the thrust nozzle after driving the turbines to produce the thrust.


This is a side view of the engine. The air enters through the prewhirl fan on the left. It enters the compressor blades on 75% of the cage interior perimeter shown here on the bottom. Leaves the compressor blades into the volute shown on the very bottom. Hits the flameholders in the hot section. Burns and enters the turbine nozzles and turbine blades. Leaves the turbine blades into the thrust nozzle. Provides thrust.


By running the liquid fuel through blades as shown here, the blades are both regeneratively  and film cooled by the entire fuel flow. The compressing air strips the fuel film from the blades and mixes with it in the volute even as it is doing pressure recovery. The air and fuel will be well mixed before the mixture hits the flame holders.

What results from all this is a true turbojet capable of a thrust/weight of 25 at sea level with a specific impulse of over 2,000. In cruise considerably more. The double regenerative blade cooling along with the fuel film cooling means that this engine could run stochiometric up to around Mach 5 without throttling down. It also means that the use of liquid oxygen for mass injection precooling would allow even higher thrust at any altitude or airspeed. Enough fuel could be used to burn all the oxygen for high thrust even at high altitudes.

An auxiliary rocket could be used in the burn chamber in air turborocket mode for extreme altitudes and to guard against flame outs if there is an inlet unstart.

This type engine could be used as an add on for a conventional air launch aircraft. Use the normal engines for cruise with the cagejets at take off and the gamma maneuver. Mach 0.9 in a near vertical climb at 50,000 feet would seem a good place to be compared to the drop and light of most air launch concepts.  If the cagejets are wanted to cruise also, then they could be used as jets during cruise and turborockets when a lot of extra thrust is needed.

If it was desirable to modify this engine further for turbofan class fuel economy, a second smaller cage could be used to bring the total compression ratio to 16 for the burn in the second spool. Specific impulse to several thousand with the possibility of getting extreme thrust levels with the flick of the switch.

Posted in Uncategorized | 4 Comments

Goff Family Vacation 2015

We’re about to head out on a two week road-trip/vacation–the longest vacation we’ve done as a family since I started Altius in 2010. Here’s our planned route map:

GoffFamilyTripRouteWe’re talking about 3500 miles, not counting any driving around at our destinations and visiting 9 states1. We’ll be visiting Yellowstone, Tiff’s family in Oregon, the Redwoods, Silicon Valley, and my family in Utah.

Overview of our schedule:

  • Day 1: Leave Lafayette and drive up to Dubois, WY.
  • Days 2-3: Camp in Yellowstone.
  • Day 4: Break camp, catch church near Yellowstone, then head toward Eugene (via Montana and Spokane, WA).
  • Day 5: Arrive in Eugene, OR at Tiff’s dad’s place.
  • Days 6-9: Have fun near Eugene, including probably a trip out to the Oregon Coast.
  • Day 10: Drive down to the Redwoods in California, then from there down to San Francisco.
  • Day 11: Spend time in San Francisco with a friend’s family.
  • Day 12: Taking a one-day break from my vacation to visit some various groups in Silicon Valley for Altius, and then hanging out with the Traugotts2 in Livermore.
  • Day 13: Drive to my sister’s place in Eagle Mountain, Utah.
  • Day 14: Visit my family in Utah, including my parents, several of my siblings, my grandmother who is turning 95 later this month3, and a physicist friend of mine.
  • Day 15: Drive home to Colorado.
  • Day 16: Hopefully be alive enough to go back in to the office…

It should be epic. And hopefully fun. But definitely epic. I’ll post pictures as I get to places with internet access.


Posted in Administrivia, Excuses for Light Blogging, Family | Leave a comment

Boomerang Air-Launched TSTO RLV Concept Part III: Carrier Plane Support Subsystems

Before getting into my thoughts on potential options for the carrier plane itself, I wanted to mention a few nice-to-have options for the carrier plane itself. I don’t know that any of these is strictly required, but all potentially help:

Top-off Tanks
For many reasons cryogenic propellants would be the best option for truly competitive air-launch. But both for boiloff reasons, and for providing cross-fed propellants during the gamma-maneuver, having some smaller propellant tanks on the aircraft itself could be useful. These tanks could be insulated more thoroughly than flight tanks, since the carrier plane is the least weight-sensitive part of the system.

One clever option that Doug Jones mentioned to me at Space Access if you have such tanks is to fly up to 30kft, vent the launch vehicle propellant tanks (one at a time)1, let the tank vent until it is at the now much lower ambient pressure, and then refill the tank till nearly full. Cryo propellants will boil at a colder temperature at altitude, and the heat absorbed by boiling off some of the propellants will chill the remaining propellant to this lower temperature, densifying it2, and significantly reducing the pressure needed in the tanks to suppress pump cavitation. Both of these can result in a non-trivial reduction in system mass, especially if your system is large enough that you’re not at minimum gage levels for your tank wall material.

Crossfeed Pumps and T-Zero Disconnects
A lot of ground launch vehicles have propellant, pressurant, and electrical umbilicals/quick disconnects that only separate right as the vehicle is taking off the pad. You might not want to cut it quite as close here (T-1 disconnects would be fine too), and you’ll definitely want features for retracting the hoses out of the air-stream after they’ve disconnected, but making these types of hoses and disconnects shouldn’t be that hard. In the case of crossfeed pumps, you only need a pump that can keep up with the flow-rate of the engines operating at whatever throttle level they’re running at during flight operations, and push against the backpressure from the main propellant tank pressurant, since you’re feeding in through the same fill ports the vehicle uses for ground filling. The power levels required would be low enough that an electric powered pump would make a lot of sense–easy to control the pump flow-rate/pressure to make sure you don’t over or underfill the tank. You’re probably talking about needing a pressure of 15psi or less, which means that compared to an electropump for a main propulsion system (like RocketLabs and Ventions are using), you’re probably only looking at 1-5% of the power needed for the cross-feed pumps. I see how cross-feed for dual strap-ons on a tri-core rocket stage like Delta-IV or Falcon Heavy might be hard, but this seems relatively straightforward by comparison.

Propellant Umbilical Reconnect Mechanisms
A slightly harder task would be designing the disconnects in a way that they could be in-flight reconnected. This might involve some level of robotic or mechanism hardware to make the reconnection, but could be handy in case of a last-second abort. Also, this same sort of hardware would likely be exactly what you’d want for refueling or detanking the upper stage at an orbital propellant depot.

Emergency Detanking Hardware
In addition to cross-feed/tank-up pumps, it might be good to have a way of detanking the propellants from the rocket in case of an aborted mission. This could possibly use some of the same hardware, but thought should be taken on how and where you route the dumping propellant, and how you sequence them, so you avoid building up hazardous concentrations of flammable materials near spark sources during an emergency propellant dumping operation.

In-air Propellant Transfer Hardware
I wasn’t thinking about this in my baseline Boomerang system, but having the ability to transfer propellants to the carrier aircraft in-flight might enable launch vehicle performance enhancements without requiring a bigger carrier aircraft. While kerosene transfer is routinely done by the military, LOX and cryogenic propellant transfer should also be technically feasible, but would require some demonstration3. I’d probably have prop transfer go into the holding tanks on the aircraft, and from there into the launch vehicle (that way you minimize the odds of damage to the vehicle, and reduce the dry mass impact on the launch vehicle itself).

Most aircraft mass limits are due to take off thrust and abort considerations. If you could launch with the rocket empty or mostly empty of at least one propellant type4, you could carry a much bigger rocket and payload at takeoff with the same carrier plane. This would allow growing to a larger system over time if desired without requiring a new carrier vehicle design. Depending on the range of the tanker aircraft, this might also give the carrier airplane more flexibility on how far it flew prior to launch operations.

Tow Cable Reattachment Hardware
At least some carrier plane options would use a glider-based carrier plane towed by a larger, more traditional aircraft. It might be challenging, but in the case of an abort with a glider based carrier aircraft having a mechanism similar to the in-air-fuel transfer mechanism that allows reconnecting a tow cable to the nose of the carrier aircraft might be valuable. It might even be useful for normal operations, where the towing aircraft could maneuver out of the way for launch, and then catch up with the glider and reconnect with it to tow it home to the launch site. If you’re particularly crazy/clever, you might even be able to find a way to combine this with in-air propellant transfer hardware to enable retanking the rocket in case of an abort, though I’m not sure what scenarios that capability would make sense.

Anyhow, as I said at the start of this blog post, most of these capabilities are nice-to-have, not have-to-have. Prioritywise, the top-up tanks and cross-feed pumps are the ones I think would be most worth looking into for a first generation Boomerang system.

Next up in Part IV: my thoughts on carrier plane options.

Posted in Boomerang TSTO RLV, Commercial Space, Launch Vehicles, Orbital Access Methodologies, Rocket Design Theory | 3 Comments

Boomerang Air-Launched TSTO RLV Concept Part II: Carrier Plane Gamma Maneuver Safety Considerations

In this post I would like to discuss some of the key considerations and options governing the carrier plane for my Boomerang TSTO RLV concept. Specifically I’ll be covering considerations relating to safely supporting the Gamma Maneuver, and useful support hardware. I’ll also discuss specific options for the carrier plane role, and their pluses and minuses. I had originally intended to include a lot more of the math and illustrations for this post, but I’ll have to come back and provide those in a later part of this series.

Gamma Maneuver Considerations
As discussed in Part I, one of the key enabling concepts for Boomerang is the use of a Gamma Maneuver, where the rocket engines are ignited while still attached to the carrier airplane, enabling the airplane to pull up into a steep flight-path angle prior to rocket separation. For review, some of the key reasons why the Gamma Maneuver is worth considering even though it is scary include:

  • Performance: The traditional straight-and-level drop-and-light approaches that have been used for most air launches to-date1 all result in a losing most of the performance benefit from air launching. The 3-5 second delay between separation and ignition results in a significant negative component to the flight path angle (velocity vector) that now has to be made up either propulsively or via a wing. If you use the wing route, you end up driving the structural design, making liquid propulsion harder, and still taking a decent performance hit. Without the wing, you lose a significant amount of altitude, require a higher T/W ratio on the stage, and wipe out most of the saved gravity loss. With the exponential nature of the rocket equation, doing air-launch in a way to gain the full 1km/s of saved drag and gravity losses makes a huge difference, potentially allowing you to not have to push the design as hard in other areas. Often it’s best to pick one or two hard things to do that make everything else easier.
  • Engine Operation Verification: Performing the Gamma Maneuver enables you to verify nominal engine operations before separation with the carrier airplane, similar to using a hold-down mechanism like SpaceX and several other ground-launch operators. Think of how many times SpaceX has decided to abort a launch due to an issue during engine startup prior to lift-off. How many launches would they have lost by now if they didn’t get to light the engines until they were already in free-fall? Drop-and-light greatly increases your odds of losing launch vehicles and payloads due to no-lights. That might be acceptable for expendable “artillery rockets”, but isn’t a good risk for RLVs.
  • Eliminating or Minimizing Bending Loads: The Gamma Maneuver enables you to do your rocket vehicle in a way that avoids bending loads and large wings that are typically required on air-launched vehicles. The need to aerodynamically change the vehicle’s flight path angle using wings induces significant bending loads on the launch vehicle, which significantly decreases the achievable mass ratio of the launch vehicle, partially erasing some of the benefits of air-launching.

The problem is that most pilots consider lighting a rocket attached to their carrier plane to be extremely anti-social behavior. Specifically, from my conversations with several aircraft companies about the Gamma Maneuver, I’ve heard several concerns voiced:

  1. Carrier aircraft controllability
  2. Assumed risk to the carrier aircraft pilots
  3. Rapid Unplanned Disassembly of the Launch Vehicle taking out the carrier aircraft
  4. Separation dynamics issues resulting in post-separation recontact
  5. How loads are reacted through the attachment structures/release mechanisms
  6. Structural loads induced on the whole aircraft during the gamma maneuver
  7. Plume heating on the carrier aircraft
  8. Plume impingement on the carrier aircraft

Thrust Related Concerns
My initial plan for solving most of the thrust related problems (specifically Items #1, #4, #5, and #6) was to have the aircraft begin pitching up, and then light the rocket throttled way down, and gradually increasing it to cancel out the drag on the rocket and the component of gravity acting on the rocket parallel to the thrust of the aircraft and rocket2. By doing this, you actually relieve stresses on the attachment mechanism/structure because the only net force will be one normal to the attach structure3, you shouldn’t have significant off-axis thrusts to cause control issues, and separation dynamics would be very similar to a drop-and-light, except the separation acceleration would be decreased by the higher flight path angle (possibly necessitating the use of some separation springs).

The problem is that for a non-winged VTVL rocket vehicle, you need to be in a flight path angle >60 degrees above horizontal. I thought this wouldn’t be a problem–after all, WhiteKnightTwo has an unloaded takeoff T/W ratio > 1, so I figured it shouldn’t be a problem. Unfortunately, when I ran the numbers on how much thrust you lose by the time you get to 30kft/9km, it turns out you only get ~33% of the thrust at that altitude. Neglecting drag, and assuming the rocket is offloading all the drag and the parallel component of the rocket’s weight, you need a T/W ratio greater than ~0.85 to get to a 60 degree flight path angle4. With the data I have for WK2, it can only get up to a ~25 degree flight path angle using a Gamma Maneuver that only cancels out the drag and weigh components of the rocket. There are several ways to deal with this:

  1. You could use a winged RLV first stage. By getting up to a 25 degree flight path angle, you’re already near the optimal flight path angle for a lightly-winged rocket stage. For instance the SpacePlane concept Dan DeLong came up with at Teledyne Brown would’ve used separation at a 25 degree flight path angle, with the RL-10 engines lit to compensate for drag losses. Unfortunately this trick doesn’t work so well for VTVL designs that tend to need much higher separation flight path angles (60-75 degrees).
  2. You could settle for only a 25 degree initial flight path angle. This would likely require a turning wing, big gravity losses, or a very high vehicle T/W ratio combined with high maximum dynamic pressures5. Without running the numbers more I don’t know how bad this.
  3. You could do a “zoom climb”. This is one where your vehicle is not able to maintain speed at the higher angle, but you pull up and then separate before your aircraft has slowed down to too close to its stall speed. This has the penalty that your velocity is now lower, erasing some of the benefit of a gamma maneuver. However, I’d have to do a lot of additional analyses to figure out if the losses are showstoppers or nuisances.
  4. If the aircraft’s launch vehicle support structure can support enough net thrust from the rocket to enable reaching the desired flight path angle/velocity, but the separation mechanism can’t handle large net shear forces, you could throttle up the rocket enough to get to the right flight path angle and velocity, and then throttle down to just drag/weight makeup shortly before separation, and then re-throttle up after departure. This is more complex on engine operations though, and might not be possible with COTS carrier vehicles like WK2.
  5. If the aircraft’s launch vehicle support structure can support some net thrust, but not enough to maintain velocity at the desired flight path angle, you could do a hybrid between #4 and #3, where you pull up to the desired flight path angle, providing enough net thrust to minimize the velocity loss before separation. You might or might not have to throttle down prior to separation, but this option would likely work on existing COTS carrier vehicles, and would minimize the velocity losses from option #3.
  6. The best option would be to modify the carrier aircraft to have a support/release mechanism that can handle the rocket operating at enough thrust to maintain speed at the desired flight-path angle. The control, structural design, and separation analyses would all be more complex, but this would give the best performance, with the least complexity on the rocket side.

My preference would be to check the numbers on  #3 first, followed by #2, and then #5 because those could likely work without much/any modifications to the carrier vehicle. It would also be worth seeing if the shear load capability of the pylon and release mechanism are sufficient to support something like #4, but the odds of it working for #6 with COTS aircraft is small. Only if none of #2-5 work would #6 be that interesting, because it likely implies either a clean-sheet carrier aircraft design or significant modifications to the carrier plane, which might be hard to get approvals for on a COTS carrier plane.

The best way to retire risks for many of these concepts, after picking the approach that at least appears to look best on paper, would be to test in subscale using something about the size of the TGALS system developed at NASA Dryden. This is a scale small enough that the rocket can be low-cost, and many different ideas can be tried out quickly, and verified in a way that doesn’t risk a pilot or an expensive, full-scale operational aircraft.

Towed Glider Air-Launch System and its Towing Plane on the Ground with Ground Crew for Sense of Scale

Towed Glider Air-Launch System and its Towing Plane on the Ground with Ground Crew for Sense of Scale

Other Gamma Maneuver Safety Concerns and Mitigations:
Relative to thrust, separation, and control related concerns, the others seem a lot more easily addressable. Specifically:

  • Assumed risk to the carrier aircraft crew can be mitigated either by eliminating the hazards or by going with an optionally piloted or completely unmanned carrier aircraft design (or preferably both). The main fear most of the pilots seem to have are the rocket exploding, structural failure of the aircraft due to rocket thrust, or loss of control due to the coupled behavior of the aircraft or plume impingement after separation. Most of those can be tested-out in subscale, as mentioned in the last section.
  • Engine explosions taking out the aircraft can be mitigated primarily by providing shrapnel protection “armor” around the engine hardware. While hybrids rockets, pressure-fed liquids, and solids all tend to have a large pressure vessel storing large amounts of pressure energy, pump-fed liquid rockets tend to have their high-pressure parts all down in the engine area. For an RLV, you already want to protect engines from each other, and the stage itself from the engines. Also, there are many directions shrapnel can be allowed to travel in that do not endanger the carrier aircraft, so the blast shields only have to divert the blast, not completely absorb it. While all of this takes some engineering, a lot of it are things you’d already want to do for an RLV, and doesn’t necessarily need to take a lot of weight. Making armor so that a worst case engine failure doesn’t damage the aircraft, and doesn’t propogate to a total launch vehicle failure is probably achievable. Though you’d likely want to do a few engine fraticide tests to be really sure. More benign engine cycles like expander cycles and electropumps tend to be less likely to explode in heinous ways than cycles like staged combustion or high-pressure gas generator cycles. Fortunately, air launch makes expander and electropump cycles more feasible than for ground-launched rockets.
  • Separation dynamics risks could also be potentially mitigated (if necessary) by doing some form of separation thrusters on the launch vehicle. If you’re planning on doing a DTAL/Xeus style lander you might already have the plumbing for rockets in the right class out where you’d want to apply separation thrust. Though there are obvious plume impingement issues you’d need to look into–for instance, the separation thrusters could be timed to operate for a short enough duration that the rocket hasn’t had time to move forward enough relative to the carrier plane for the separation engines to have their plumes impinge on the aircraft. Obviously you’d only do this if the separation looked marginal or unsafe without them, as that’s an extra piece that has to work correctly for a safe separation.
  • Plume heating from the rocket can be mitigated by either picking a propellant combination whose exhaust is relatively low luminosity (LOX/alcohol, LOX/Methane, LOX/Propane, and LOX/LH2 are all pretty good), or by making the aircraft more tolerant of radiative heating, or both. Fortunately, at realistic air-launch altitudes, the engine won’t be so overexpanded that the plume actually contacts the aircraft prior to separation, so you’ll have cool air flowing over the surfaces, and heat transfer primarily by radiation heat transfer, so this problem might not be as scary as it sounds. SS2 had to do some modifications to its tails because they operate fairly close to the rather luminous hybrid exhaust. The solution appeared to be adding a layer of metalized mylar or kapton to the skin of the tails to reflect the radiated heat away. The structure of the carrier plane are likely going to be much further away, and if you use a propellant combo that isn’t very luminous, you might not even need something as fancy as that.

    SS2 with Reflective Coating on Wings Clearly Visible (Credit Virgin Galactic, Jan 2014)

    SS2 with Reflective Coating on Wings Clearly Visible (Credit Virgin Galactic, Jan 2014)

So to recap, the biggest adaptations that might need to happen to a COTS carrier plane to enable the Gamma Maneuver include: potentially going with an unmanned or optionally-piloted configuration, potentially needing to strengthen the aircraft-to-launch vehicle support structure, potentially needing to redesign the release mechanism to operate under load, and potentially needing to add a metalized layer to the carrier aircraft’s tail sections. Most of these aren’t potential show-stoppers, but they do point to the trade between using a COTS carrier aircraft, and a Gamma Maneuver-Optimized clean sheet design.

I’ll continue this in Part III where I’ll discuss additional carrier plane support subsystems, and my thoughts on the best type of carrier aircraft for a Boomerang-type air-launched RLV.

Posted in Boomerang TSTO RLV, Commercial Space, Launch Vehicles, Orbital Access Methodologies, Rocket Design Theory | 3 Comments