Sunday, October 27, 2013

Manned Launch to LEO Using Ramjet Missile Technology

Reported here is a concept for a two-stage to orbit (TSTO) vehicle,  capable of sending two astronauts to low Earth orbit (LEO) in a minimal transfer and return capsule.  Modern ablative heat shield items are assumed to reduce the capsule weight.  The capsule has a “service module” containing on-orbit propulsion and folding solar panels for electric power. 

These astronauts arrive in LEO with a significant on-orbit maneuvering propellant budget,  and can conduct short term (one week or less) missions unaided.  Longer missions would require docking with something that had some living space and life support supplies.  The on-orbit propellants are nitrogen tetroxide-monomethyl hydrazine (NTO-MMH),  which are very storable for very long periods in space.  Total on-orbit delta-vee was simply assumed to be 1 km/s. 

These roughed-out designs are the results of bounding calculations,  not trajectory simulations.  I went through four iterations to reach a rough size-out that I trust.  The second stage was analyzed with a loss-corrected approximation of the rocket equation.  The first stage was analyzed with rocket-impulse and ramjet-energy approximation analyses.  Everything was checked with thrust versus drag maps to ensure basic feasibility,  a considerable problem reaching high speed in the thin air where staging takes place. 

For the results of a true horizontal takeoff study utilizing ramjet assist,  see "HTO/HL Launch with Ramjet Assist",  dated 11-6-13.

First Stage

The first stage takes the form of two integral rocket-ramjet (IRR) booster pods,  based on a solid-propellant rocket booster,  and a liquid-fueled ramjet sustainer,  right out of missile technology based on ASALM-PTV,  about 1980.  The same technology is in the “Sunburn” and “Yakhonts” missiles built by the Russians today.  The booster is packaged within the ramjet combustor,  requiring some sort of inlet port cover,  and some sort of ejectable booster nozzle nested within the ramjet nozzle,  just as in ASALM and the Russian missiles.  10-30-13:  also the US ALVRJ (1970's),  AAAM (1990's),  and VDR (1980's-1990's) efforts used the same basic IRR technologies (I actually worked on ASALM,  AAAM,  and VFDR).  As did the 1960's Soviet SA-6 "Gainful" (that I also got to exploit ca. 1970's).  

I have previously posted some of the ramjet thrust performance data for such a pod (see references list),  based on RJ-5 fuel,  which is a synthetic similar to kerosene,  but slightly denser than water.   This work was based on a fixed ramjet nozzle geometry,  but a variable inlet geometry to maintain shock-on-lip throughout the range of ramjet flight speeds.  This amounts to a very simple translating compression spike,  similar to those used on the SR-71 “Black Bird”,  except operated for ramjet instead of turbine.  (The two engines are quite different,  even though the inlet components are the same.)

The IRR booster in each pod is a state-of-the-art AP-HTPB composite propellant of 20% aluminum and 87% solids.  It is similar to shuttle SRB propellant,  except for the really-large thixotropic effective-viscosity that is inherent with the really-high solids content that confers high Isp (some 254 demonstrated seconds).  Not many companies had the expertise to reliably cast such a propellant,  but the one I worked for decades ago did.  This is also essentially the same as the AIM-120 AMRAAM propellant we used successfully at the McGregor,  Texas plant,  before corporate politics closed it. 

Second Stage

The second stage is essentially Centaur technology,  revised to a slender form factor.  This is stainless steel balloon-tank technology,  with one engine similar to an RL-10.  It is a LOX-LH2 system.  I assumed only about 450 seconds of Isp,  in order to do a conservative performance estimate.  (Centaur actually does better than this.)  I assumed no need for wings during pull-up from staging,  relying entirely on body lift at Mach 6. 

Partial Reusability

I made the first stage booster pods reusable as flyback remote-control (R/C) aircraft,  fitted for a runway landing on land.  I didn’t get enough reserve fuel out of this bounding analysis to support flying back all the way to the launch site.  There are two of these pods,  one on each side,  which confers plenty of body lift,  even at only Mach 1.7,  the ramjet takeover point that I used.  The ramjet design works all the way down to Mach 1.5,  but I set takeover at Mach 1.7 to get some reliability “margin”.  After all,  this is supposed to be a manned system. 

The second stage is assumed expendable,  just as Centaur is today.  Its price ought to look about like that of Centaur.  The service module is also assumed expendable.  The capsule could be re-flown many times,  given about 2-to-3 inches (50-to-75 mm) thickness of PICA-X for its heat shield. 

First Stage Trajectory

This thing launches ballistically at a “high” gee level still tolerable by a human crew:  around 5 gees.  That’s fast enough to achieve flight speed “without time to swap ends”.  It comes off a slanted rail launcher (angle not determined here,  that’s for “real” trajectory code analysis),  and reaches ramjet takeover (Mach 1.7) at the relatively low altitude of about 5000 ft (1.5 km) above mean sea level (MSL).  Control during this phase would be by relatively-small attitude thrusters.  See Figure 1.

Once the solid integral boosters burn out,  the ramjets “light up”.  Transition in ASALM was 100 milliseconds,  this should be comparable to that.  Each pod has a frangible-glass port cover protected with bonded rubber insulation,  and equipped with a destruct charge set off as booster chamber pressure tails off.  Each pod has an ejectable booster nozzle (I computed roughly a 10 inch diameter throat,  a 38 inch diameter exit,  and a 52 inch exit bell length at 15-degree half angle).   All of this is based on ASALM technology.

I guessed these ejectable nozzle assemblies at 200 lb per pod,  scaled up from 15 lb for ASALM,  more or less on diameter squared,  not cubed,  which may or may not be as realistic as it should be.  These nozzles are held in place by a circular snap ring in a groove near the exit,  just like ASALM.  That snap ring is underlain with very-high-quality detonation cord,  also set off by the tailoff of booster pressure,  just like ASALM.  Then the fast-declining residual boost pressures “shove” the ejectable nozzle clear.  The boost pressure tailoff also sets off a magnesium flare to light the fuel-air reaction,  same as ASALM.  (From about Mach 3 on up,  the fuel-air reaction is essentially hypergolic.) 

The ramjets have very significant thrust margin over drag,  although vehicle accelerations in ramjet are far less (factor 10) than in IRR rocket boost.  The trajectory assumes a constant-altitude acceleration to Mach 2 at 5000 feet,  where thrust margin is pretty much optimum.  The vehicle then pulls up on body lift for a Mach 2 climb to 60,000 feet altitude (18.3 km).  I did a spreadsheet thrust-and-drag point-performance-estimate series,  to ensure this would be feasible.  These estimates used the same thrust and overall-efficiency results reported for the pod,  just correlated versus Mach number and altitude. 

Once the vehicle reaches 60,000 feet at Mach 2,  it pulls-over horizontal and uses its thrust margin to accelerate to Mach 6 for staging.  The crucial controlling factor for acceleration to this speed is vehicle drag.  It sets the diameter ratio of the second stage to each of the two IRR booster pods.  That is where the 7 foot to 5.22 foot diameter ratio comes from.  The ramjet cycle analysis that I used is beginning to break down from ionization/recombination effects,  from about Mach 5 on up.  It begins diverging sharply from reality above about Mach 6.  That and the ASALM speed record,  are where the staging speed used here came from.  Overall energy conversion efficiencies fall in the 22-23% class. 

This relative diameter result is a function of the number of pods.  That determines how much drag must be overcome by the ramjet thrust available in the “thin air” at altitude.    I got 5.22 ft diameter for each of two pods,  carrying a minimum-credible 7 ft diameter payload.  My drag data came from that for clean projectiles in the old Hoerner “drag bible”.  (If you don’t know what that reference is and how to use it,  you have no business playing with this concept.)

The pods have to be nose inlet configuration,  in order to limit pod drag to skin friction and a little cowl lip pressure drag (that is the minimum achievable).  Side inlet configurations will be “draggier”,  and will have far more difficulty reaching Mach 6 at 60,000 feet (or any other altitude).  The single nose inlet selection here is therefore the “optimum” configuration choice.  It features a cowl area about 50% of the pod cross section area,  an inlet duct about 40%,  a ramjet throat about 65%,  and a ramjet nozzle exit pretty close to 100% of the pod cross section. 

Higher staging altitudes reduce frontal ramjet thrust,  reducing vehicle acceleration,  extending the acceleration range,  and increasing the first stage fuel and article weights.  This is based on earlier trade studies I did with similar bounding calculations.  It’s a judgment call,  but 60,000 ft seems to be a pretty good “rule of thumb”.

Second Stage Trajectory

Since the staging point is Mach 6 at 60,000 ft,  there is plenty of body lift available to pull up to about 40-degree path angle.  “Store separation” hypersonically is even more problematical than at supersonic speeds,  something the military avoids at all costs.  That is because the aerodynamic forces are so much larger than the weight forces,  even in the thin air.  That means stage separation will require small solid motors strategically placed,  on both stages.  Those details are beyond scope here,  although the solid motors would resemble scaled-up versions of the ullage motors seen in many launch vehicles. 

The second stage pulls up to a steep path angle on body lift,  then flies a simple ballistic gravity-drag turn to orbit.  See Figure 2.  I simply assumed a 10% gravity-drag combined loss,  imposed on the velocity requirement for simple rocket equation mass ratio analysis.  Mass ratio sets the propellant mass fraction.  That and the inert fraction set the payload fraction. 

If the mass of the payload is fixed (as it is here),  that sets the weight statement,  for any given propellant combination.  The combination here is liquid oxygen-liquid hydrogen (LOX-LH2),  similar to that in the Centaur upper stage in use today.  As in Centaur,  I assumed steel balloon tankage construction,  just of a slim form factor.  Inerts are 10% of stage ignition weight,  as in Centaur. 

The Ultimate Payload

What is delivered to LEO is a manned capsule with a “service module” that supplies solar electric power plus significant on-orbit maneuver capability.  I assumed (1) very storable propellants,  and (2) 1 km/s maneuver capability.  The propellants are NTO-MMH.  I assumed a conservative Isp of 300 s for the maneuver engine.  See Figure 3. 

The capsule is a minimum-credible two-man re-entry vehicle of an assumed 2000 lb,  assuming low-density ablative heat shielding that can be re-flown a few times.  That would be PICA-X for the main heat shield,  and Avcoat on the afterbody external surfaces.  The minimum credible diameter for a cramped two-person capsule would be near 7 feet. 

This capsule would contain nothing more than two persons,  their spacesuits,  and many hours of life support.  The idea would be adequate support for a few days of independent operation,  or a “lifetime” of many weeks to several months if docked to an independent habitat such as the International Space Station.  There are an entire plethora of useful missions that such a crew transport capability could support,  a subject not explored here.

Second Stage

The second stage is a simple rocket vehicle using LOX-LH2 propulsion,  with a gimballed engine.  I used a simplified analysis and a conservative Isp to estimate the stage properties shown in Figure 4.  The payload is the two-man capsule and service module already described as the “ultimate payload”.  This second stage is not assumed to be recovered and reused,  although there is always that possibility,  given the higher inert fraction required to cover the heat shield and parachute equipment,  and perhaps some additional deorbit/deceleration propellants.  Making that stage recoverable and reusable is far out of scope here.

First Stage IRR Pods

The first stage IRR pods are shown in Figure 5.  This pod concept is basically identical to that described in the earlier postings regarding ramjet pod performance.  The booster is AP-HTPB composite propellant under web fraction and volumetric-loading conditions easy to meet,  with a burn rate already well-demonstrated without resort to ultra-fine AP grinds.  The ramjet uses the dense synthetic kerosene-like fuel RJ-5,  sometimes known as Shelldyne-H. 

This is the nose-inlet analog to the chin-inlet ASALM-PTV missile configuration.  The only difference between these results and the earlier postings of ramjet pod performance is in the pod drag:  fins that fold into a wake zone,  instead of fixed fins.  Those fins are needed for stability and control,  after staging,  for recovery purposes.  They have significant drag,  left in the slipstream.

Recovery is by runway landing,  on a steerable nosewheel and fixed skids on the ventral fins.  Two pivoting wings are extended subsonically,  to form a biplane configuration capable of landing at a bit over 150 knots equivalent air speed (KEAS),  without resort to flaps and leading-edge devices.  These pods are not manned;  recovery is by remote radio control. 

They cruise back as far as fuel allows,  at Mach 2 60,000 ft conditions,  fins extended,  but wings stowed.  Once the engine is “off”,  the pod quickly decelerates subsonic,  at which point the wings pivot to extend for subsonic glide to a dead-stick landing on the runway. 

These pods are modular:  an inlet section,  a tank section,  and a combustor section.  Refurbishment of the pod requires separation of the combustor IRR section,  and its replacement by another already loaded with propellant and transition gear out of inventory. 

The inlet section merely requires minimal servicing of the hydraulics for the nosewheel and translating inlet compression spike.  The tank section merely needs fuel control checkout,  and tank refueling.  All three sections need the external cork layer residues power-washed off,  and replaced with fresh cork insulation.  This was a technique used very successfully on the Phoenix missile for hypersonic aeroheat protection.  

The combustor section needs the internal insulation residues power-washed clean.  Then it needs a new DC 93-104 insulator cast in place on retaining ribbons,  and primed for an etched Teflon separator sheet.  That sheeted insulator then needs another primer before casting the booster propellant.  This system was quite successful in ASALM,  in spite of the otherwise-incompatible chemistries of propellant and insulator.

The booster is envisioned as a simple internal-burning cylinder with both ends unrestricted,  which would be slightly progressive,  more-or-less matching the drag increase with speed.  The propellant resembles shuttle SRB propellant,  except for the higher solids content that makes it very thixotropic.  This material has to be pressure-cast into vacuum in the case,  which is fitted with cast tooling.  The pressure-casting pressure level is very significant,  or else voids will cause fatal troubles.  This was a processing technique well-proven at the old McGregor, Texas weapons plant.  Few others could do this.

The unique concept here is “circumferential folding” of the fins,  something fairly compatible with the limited-volume of the ramjet nozzle recess,  by means of circumferentially-rotating rings to which the fins are fixed.  During rocket boost and ramjet sustain,  the three fins of each pod are stowed in the wake of the second stage,  for no perceptible drag increment.  That is how power-on pod drag can be limited to skin friction plus a bit of cowl lip pressure drag.  (Power-off,  one must include the base drag,  too,  which differs between rocket and ramjet,  due to the differing exit area sizes.)    

Upon staging,  the fins are quickly moved to their proper positions.  This can be done hypersonically,  because each fin is essentially streamline to the flow,  and thus is not subjected to catastrophic forces.  The pod flies supersonically on body lift until fuel is exhausted.  Deceleration to subsonic is then quite rapid,  after which the pivoting wings are extended.  Two wings are used,  in a biplane configuration,  to reduce the wing loading under 100 lb/sq.ft,  which gets the sea level stall speed down near 150 knots equivalent air speed.  That is a very practical value,  without any flaps or leading-edge devices.  Each wing is about 35 foot span,  and constant 2 ft chord.  A flat-bottomed subsonic airfoil is the proper choice. 

Cluster Vehicle

The vehicle comprises the second stage rocket plus two IRR pods strapped-on,  as shown in Figure 6.  The strap-ons are positioned on either side for maximum body lift capability.  They are staggered axially to provide the room at the rear in which to stow the fins in the wake of the second stage.  The wings are stowed top and bottom by pivoting into axial alignment with the pod airframe. 

This is necessarily a cluster vehicle.  The axial stagger ensures that the second stage bow shockwave does not impinge upon the strap-on IRR pods.  However,  the spike shocks of the pod inlets do impinge upon the second stage.  Depending upon material selections,  the enhanced shock-impingement heating can be quite catastrophic above about Mach 2-to-3.  Even with heavy Inconel-X skins,  shock-impingement heating became quite catastrophic on the X-15 flight that carried a scramjet test device in place of its ventral fin. 

For this vehicle,  small sacrificial panels of the tough ablative silica phenolic could be located in the impingement zones on the second stage.  Although expensive and a bit heavy,  such panels would provide a reasonably-certain “fix” for the otherwise-fatal shock-impingement heating problem. 

Concluding Remarks

I did not even try to estimate costs.  But,  consider this:  this vehicle can send a crew of 2 to LEO,  yet grosses near only 120,000 lb at launch.  That is a lot smaller than current commercial satellite launch vehicles,  and more in line with the size of the historic Titans that launched Gemini in the 1960’s. 

The second stage rocket is a form-factor variant of the venerable Centaur upper stage,  and thus should be priced similarly. 

The small capsule and service module should resemble the historic Gemini of the 1960’s except that this one is partially reusable.  The price should be similar to Gemini corrected for inflation,  but significantly reduced for experience gained since then,  and for reusability of the return capsule itself. 

The first stage pods resemble nothing so much as large tactical missiles,  and should have similarly-small manufacturing and logistical support “tails”.  The prices for these should resemble the prices for large tactical missiles,  quite a bit cheaper than is “customary”,  even with commercial launchers. 

There is absolutely nothing here that requires new technology development.  There is only the application of well-established missile technology in a launch venue not accustomed to using it since about 1960. 

Just as a wild guess,  call it $25M per launch,  for 3300 lb delivered to LEO.  That’s about $7500/lb or $16,700/kg,  which is higher than commercial,  but still less than the trend of government-operated vehicles.  The amazing schedule flexibility of a missile-like launcher is preserved in this design,  a unique advantage indeed. 

My point is that simplicity,  flexibility,  and smaller launched size,  can all be had for a price not all that out-of-line with modern commercial launch. 


Sighard F. Hoerner,  “Fluid Dynamic Drag”,  self-published by the author and then his widow Liselotte,  1965.
Pratt and Whitney,  “Aeronautical Vest Pocket Handbook”,  12th edition,  December 1969 (propellant and performance data for liquid rockets).

G. W. Johnson,  “Inlet Data for Ramjet Strap-On Pod”,  posted to,  2-20-2010.

G. W. Johnson,  “Ramjet Strap-On Pod Concept”,  posted to,  2-20-2010.

G. W. Johnson,  “Ramjet Strap-On Pod Point Performance Mapping”,  posted to,  2-20-2010.

G. W. Johnson,  “Preliminary Acceleration Margins for Baseline Pod”,  posted to,  2-28-2010.

G. W. Johnson,  “More Ramjet Performance Numbers for the Strap-On Pod”,  posted to,  7-11-2010.

G. W. Johnson,  “More Strap-On Pod Ramjet Engine Data”,  posted to,  7-23-2010.

G. W. Johnson,  “Two Ramjet Aircraft Booster Studies”,  posted to,  8-22-2010.

Update 10-30-13:

The technology described here is one of several possibilities for flexible,  rapid,  and modestly-priced launch of a small work crew to orbit,  to do a specific job and then return.  The point of this article was to show that this goal can be achieved with technologies we have had for decadesAll that is needed is the will to do it,  not endless gravy-train technology-development programs,  nor politically-dictated gigantic rockets. 

The technology outlined in “End of an Era Need Not Be End of a Capability”,  posted 8-2-11,  is essentially the deployment on-orbit,  of generalized repair and maintenance facilities.  These are only occasionally manned,  when there is a job to do.  A few of these facilities would need to be deployed in a few convenient orbits. 

You send up propellants and life support supplies in small payloads when you want to use one of the repair facilities.  The propellants and some of the supplies might even be launched at high gee with a light gas gun,  an even-cheaper-to-use technology.    The crew collects these,  and brings it all to the repair station,  when they go up.  That’s why I gave the capsule a service module with a significant on-orbit maneuver capability.

If you combine a technology like this article’s crew transport capability,   with a permanently-deployed but occasionally-manned repair craft like that described in the 8-2-11 article,  then you have recreated the tremendously-successful and valuable on-orbit service and repair capability we had with the space shuttle.  

The changeand it is a huge one,  is that you no longer incur the gigantic expense of a shuttle launch (or any other giant rocket),  every single time you need to use the on-orbit repair capability. 

This approach is not the way our government space program has historically operated.  It is more like what the commercial space companies are trying to do,  but have not really accomplished yet.  Maybe there is a lesson for government space agencies heresince they still set many of the missions to be done,  and the rules by which to accomplish them.  

Use the wrong rules or goals,  and you get a bad result;  it’s just plain old common sense.  

Figure 1 – First Stage Trajectory Characteristics

Figure 2 – Second Stage Trajectory Characteristics

Figure 3 – The Ultimate Payload (Carried By the Second Stage)

Figure 4 – Second Stage Characteristics

Figure 5 – IRR Pod Characteristics

Figure 6 – Launched Vehicle (and Recovery) Characteristics

Saturday, October 12, 2013

Construction of the Plain Cactus Tool

Update 7-30-15:  The new website is fully operational.  It has all the information,  photos,  and videos anyone could ever need.  It is a turnkey site for selecting,  customizing,  and purchasing a production tool.  Shipping is available,  so sales of plans have been discontinued.  Some additional parts and labor have been farmed out to appropriate vendors,  to adjust to higher production rates,  so prices posted previously are now obsolete.  Go to

The following illustrates construction of the plain baseline cactus tool, the “Kactus Kicker”, as I have been building and selling them recently, up through S/N-054. From S/N-055 on, they are basically identical, except for a change in the snout braces, and a change to the chain towers. I have indicated these changes as an option on the last few sets of plans that I made to sell. You should have fabricated all of the piece-parts before you ever start this assembly process, of course.

One needs a reasonably well-equipped shop in which to build these tools, because the major parts and the finished tool are so bulky, heavy, and dangerous to handle. I use a set of steel fabrication horses (“fab horses”) made uniformly, and to support 1000+ pound loads anywhere along their spans. See Fig. 1, which illustrates two sets of 3 fab horses, one set of 3 per tool. These photos were taken during the construction of S/N-053 and -054.

Figure 1 – Fab Horses Upon Which Two Tools Will Be Built

Two of the fab horses are placed parallel, centered beneath the overhead hoist, about 5 feet apart. The tool chassis will rest on these. The third is initially out-of-the-way, but will support the snout as it is assembled to the center of the leading edge. Mark the leading (LE) and trailing (TE) edges on the top and bottom of the tool, as shown in Figure 2. I use a soapstone for this. Then flip it upside down for the installation of the crush rail and skids. As it shows in Figure 3, you should have ground the cut ends of the rail “reasonably smooth”. The figure shows what that means.

Figure 2 – Marking the Top Surface

Figure 3 – Rail Cut Ends Need to be Ground “Reasonably Smooth”

You skip-weld the rail to the deck plate 5 places, each side, with the rail flush to the TE, each skip weld about 5 inches long, for a total of 10 welds. If the rail has a wear lip protruding on one side of the head, that lip should face to the TE. I use 7014 rods 5/32 dia for this, DC positive electrode, about 130 amps. You weld down both sides of each skid the same way. I use a piece of ¼ flat as a shim to get the skid overhang “just right”, as shown in Figure 4. Figure 5 shows the bottom welding in progress as an overall view.

Figure 4 – Shimming the Correct Overhang for the Skids

Figure 5 – The “Bottom Welding” of Crush Rail and Skids

Then flip the assembly over for the top welding, shown in Figure 6. The ballast bar is a 1x6 flat 8 feet (96 inches) long, flush to the TE, which places it right over the crushing rail. It is skip-welded 5 places each side, just like the crush rail. This is also the time to assemble the barge front braces and weld them down to deck, although that process is incomplete in the photo. I use 7014 5/32 for all of this, as all the welds are nearly horizontal. The deck with crush rail, ballast bar, and barge front braces is the “common chassis” that I now use for both the plain tool, and for the hydraulic wheeled set of options that is the current version of the “commercial-grade” tool.

Figure 6 – “Top Welding” Begun with Ballast Bar, Barge Front Braces Are Next

Next, you put the 1.5-inch thick shims under the LE on each fab horse (see Figure 6 again, mine are made of 1-inch square tube, 1/2x1 flat, and some 1x1x1/8 angle), and position the third horse out the centerline of the LE. The shims are required to get the correct angle relationship of the deck and the snout assembly.

Figure 7 shows the snout plates, snout tube, and the older-style snout braces being tack-welded in place. These braces made of angle have been replaced by triangular gussets from S/N-055 on. You also tack the chain towers to the ballast bar, as shown in Figure 8. These are the older square tube chain towers with small gusset braces. From S/N-055 on, I have replaced these with a three-piece chain tower made of channel. You also place the barge front plates onto the overhanging skid ends, and tack them in place to the deck and the snout tube.

Figure 7 – Tacking the Snout Assembly In Place for Continuous Welding

Figure 8 – Tacking the Chain Towers In Place for Continuous Welding

Figure 9 shows the angle-type snout braces being welded in place, along with their square-plate gusset braces. This is in-close vertical-weld work. I have been using 6011 rods 1/8 dia for this, at about 90-100 amps, DC positive electrode.

Figure 9 – Welding-In-Place the Snout Braces

If you raise the back edge of the tool as shown in Figure 10, you can do easy horizontal welds of the barge front plates to the deck and to the snout tube, with the 7014 5/32 rods. The barge front is skip-welded to the deck 3 places each side, 5 inches per skip weld. The welds to the snout tube are continuous.

Figure 10 – Tipping the Tool to do the Topside Welding

If you then tip the tool LE-up with the hoist, as in Figure 11, the continuous welds along the barge front plates to the deck become horizontal enough to use the 7014 5/32 rods. So also the weld between the two snout plates becomes a horizontal continuous weld. These welds are not laying flat, but you can still do it, because you are not trying to push your puddle upward (which can be done if you have the room) or downward (which cannot).

Figure 11 – Tipping the Tool to do the Bottom-Front Welding

Update 3-21-16:  I have revised the construction process with a separate jig for plain and hydraulic snout tack-welding,  so that all the plates and gussets can be welded up before installation onto the tool deck.  I also no longer tip the tool snout-up to do the bottom welding.  Instead,  I tip it snout down until I flip it entirely upside down to do the bottom welding,  then flip it through snout-down the other way to turn it right-side up again.  This has proven to be both more effective and safer. This applies to both plain and hydraulic tools. 

From this point, there is just painting, and rigging with the tow chains, which have been covered in another article (see list below). I let the paint dry 3 days before rigging, just to let it get hard enough to handle without scarring the finish. The idea is to cover all welds on the bottom surfaces to prevent corrosion. Then paint all the upper and forward surfaces, just to make it look “pretty”.

Other Related Articles on this Site (date highlighted on this one)

2-9-17....Time Lapse Proof It Works cactus being crushed and composted
7-30-15......New Cactus Tool Website
...................turnkey site for info,  photos,  videos,  purchases
1-8-15……Kactus Kicker Development
………………production prototype & 1st production article
1-8-14……Kactus Kicker: Recent Progress
…………..….testing a revised wheeled design (experimental)
10-12-13..Construction of the Tool
………………building a “Kactus Kicker” (plain tool)
5-19-13…….Loading Steel Safely
……………….transport and storage of materials
12-19-12…Using the Cactus Tool or Tools
…………… the tool is employed (applies to any model)
11-1-12….About the Kactus Kicker
..…………….painting and rigging finished tools (plain tool)
12-28-11..Latest Production Version

………………new bigger snout and barge front (plain tool)

Sunday, October 6, 2013

Building Conformal Propellant Tanks, Etc.

For a spacecraft,  there is always a propellant tank pressurization requirement,  even if it is a very low one (1-5 psi to prevent propellant boiloff-into-vacuum).  Pressure requirements and the structural capability to meet them is where you have to start. 

Most designs must provide at least the min net positive suction head (10-100 psi) for the engine pump,  whatever that is.  A few tank designs are for pressure-fed engines,  something like 2-3 times the engine chamber pressure.  That overpressure (tank pressures anywhere from 1000 to 5000 psi) is for the pressure drop or drops through the flow regulator. 

Internal pressure is difficult to design for,  and always ends up being heavy,  for anything but simple circular cross sections,  period.  The only feasible conformal design is “lobed”,  parts of circular sections “stitched” together into an approximation of the noncircular space to be occupied.  In effect,  you build the equivalent of an air mattress. 

This sort of thing is still today far easier to do with metals and welding than with composites,  because of the inherent difficulties with composite joints (especially at higher pressures).  How to reliably design the joints between composite components,  or between composite and metal components,  is another topic for another time. 

Especially with hydrogen,  the inherent porosity of a composite requires some sort of nonporous liner,  even if it is a paint layer,  no matter how “wet” your layup is (and “wet” layups are much heavier).  This zero-porosity liner requirement raises composite panel weight significantly.  There is no way around that dilemma. 

The best structural joints with composites are fully “glassed-in” joints.  Unless you have a conformal tank with lobes big enough for a man to crawl inside,  this can be difficult or impossible to do. 

With current material technologies,  all of the preceding facts of life are why I personally favor metals and welding for conformal tankage. 

Liquid hydrogen tanks down here on Earth are usually made of 300-series stainless steel,  and for very good reasons.  These tanks can be filled and refilled for decades without cracking,  in spite of the super-cold propellant-induced thermal stress cycles,  cycling that is excruciatingly severe.  The new lithium-aluminum alloys now favored for rocket stage tank construction seem to work for liquid hydrogen quite well.  I am not familiar with that material,  since it is so new,  and I am not. 

The effects of repeated fill cycles may cause aluminum-lithium to crack from fatigue (something inherent with all metals,  and notorious with aluminum alloys).  That sort of difficulty is not something you “run into” with one-shot throwaway stages.  But,  if it doesn’t crack with repeated use,  or there are lots of cycles available before it does crack,  then reusable tank structures are possible with it.  We’ll see. 

Lobed construction with metals does not require the extensive use of doublers (except near ends where the strain mismatches),  unless  you badly design the shapes.  Ideally,  you join segments of cylinders together,  with a properly-perforated (again,  a whole other topic) linear web-wall at the joint.  The only “trick” is to eliminate all bending by your chosen geometry.  These panels are butt-welded to a three-way joint piece at every joint line.  That joint piece has a cross section that looks like a three-prong grass burr,  radiused down in the “groins” at the base of each prong.  That radius need only be at most a little larger than the panel thickness. 

Done successfully,  you have a tank only a few percent heavier than a cylinder of the same volume,  but not heavier by factors.  It will be at least a little bit heavier,  that is inevitable.  That’s simply the price you must pay for the shape you want.  Update 10-7-13:  for the same panel thicknesses and weights as cylindrical construction,  a lower-bound estimate of the weight growth factor is the perimeter length ratio,  computed from cross-section views. 

That joint piece could be made by extrusion.  Each leg of the joint piece is the same thickness as the panel that joins to it,  and must match the tangent at the edge of the panel.  Only through the cross section of the three-way joint piece is the effective thickness about twice that of the panels.  This shape’s stress distribution has been checked with 3-D finite element analysis:  it works fine.  There will be a tiny amount of shear yielding down in each “groin” line,  but only on the first pressurization cycle. 

You could use wire-feed welding to assemble the tank,  but you have to put a slightly-bulging weld bead on both sides.  That means a man must be able to crawl inside each lobe, and be able to weld inside there.  Wire feed welding works very well with aluminum,  though. 

If the tank were a stainless steel like D6ac or 4130,  you can electron beam-weld right through from one side,  with nearly-perfect weld strength efficiency.  Weld the joint pieces to the web walls,  then weld the outer shell panels to the joint pieces,  all from the outside.  Then proof test.  The reject rate should be low,  once your process is defined.  One-side electron beam-welding of steel is a well-proven industrial technology for the mass production of solid rocket motor cases. 

About a quarter century ago,  I proposed exactly that electron beam-welded,  stainless steel lobed design,  for a small conformal-case solid rocket motor case,  to meet a seemingly-idiotic shape requirement for a weapon project (that ended up never flying).  The customer expected to see some version of elliptical designs proposed,  which simply do not work unless they are extremely heavy.  Preconceptions clouded his ability to see a lightweight solution that would work. 

Questioning the assumptions you would otherwise start with,  has been the most powerful tool in my engineering arsenal for nearly 4 decades now. 

Mastering non-conformal tank technology is required for any winged or lifting-body spaceplane,  whether it be one stage or two.  Chemical,  nuclear,  or something not yet invented,  this requirement still holds for best-storing whatever propellant is required,  within the odd spaces inside the vehicle. 

Metals we know how to handle in this application,  composites not so very much.  And that plus politics is really why the X-33 program ended up unsuccessful.  It really should have started with the metal tanks in the first place,  but the folks working on it were seduced by the higher strength-to-weight ratio of composites.  In a pressure tank situation,  those advantages tend to evaporate in the harsh light of all the other design issues.  That picture of things hasn’t changed,  and probably never will.

After reading this,  some of you may wonder why I haven’t been “snapped-up” by a Boeing or a Lockheed-Martin.  The answer is simple:  I’m old,  and old guys are more expensive. 

Having the wide-ranging cross-disciplinary experiences of an old guy on your team,  may well help guide you very cost-effectively to the “right” solution for your project.  But,  the way R&D is funded by the government in this country,  project success is not required.  So labor cost is the only factor considered in government contractor hiring.  Few-to-none of us old guys get hired. 

That has led to a widely-unrecognized lost-art problem.  The engineering project team is supposed to consist of a mix of old guys and young guys.  The old guys pass on to the young ones that engineering art that was never written down.  It wasn’t written down,  primarily because the company didn’t want to pay for writing it down.  That art is about 40% of engineering practice in aerospace work. 

If there are no old guys on the team,  no art gets passed down.  Which lack neatly explains why different outfits keep reinventing all the same wheels,  and why progress with flying machines has slowed in the last few decades. 

Few are learning from industry history anymore,  because those who knew that history are largely no longer there.  You cannot get that kind of knowledge from a college classroom,  it is dirty-fingernails workplace experience,  pure and simple. 


Wednesday, October 2, 2013

Budget Moon Missions

My friend Bob Clark has a good site named “Polymath”.  He has been looking recently at less expensive ways to mount return missions to the moon,  far less expensive than anything NASA has proposed in several years.  Bob’s slogan of late is “Free Your Mind and the Rest Will Follow”.  Here’s my two cents’ worth,  along those same lines.

I freed my mind of two preconceptions:  (1) one mission/one launch,  and (2) brand new technology needs to be developed,  both of which cripple most of the NASA concepts.  I took advantage of on-orbit assembly by docking,  and rendezvous,  both in low Earth orbit (LEO),  and low lunar orbit (LLO). 

I also freed my mind of the preconception that the transfer stage needs to be a single-piece item.  Instead,  it can be modular,  with a “service module” of tanks and engine,  plus additional modules that are tanks and interconnect piping. 

By using the same engine for the landers and transfer “service modules”,  and the same tank set in the “service modules” and additional propellant modules,  costs can be drastically reduced.  There are two types of transfer vehicles:  the manned craft with a crew return capsule as its payload,  and the unmanned craft that sends landers to the moon one-way.  One could send one or two unmanned lander transfer vehicles with the manned vehicle,  rendezvousing in LLO. 

I tried sizing this stuff to fit existing commercial launchers with LOX-RP1 propellant technology,  but the performance was not adequate to do this conveniently.  So,  I sized with LOX-LH2 technology,  but at modest size and modest chamber pressure.  This engine is to be much simpler and less demanding than the very high-pressure space shuttle main engines.  Much of this “service module” hardware would resemble very strongly a variation on the hardware in the current Centaur upper stage. 

This is ballpark bounding calculation stuff,  not precision design and performance calculations.  Any real design based on these concepts would be different in detail (and more likely to use real variants of Centaur equipment),  but broadly similar to what I computed here. 

The first thing to consider is availability of commercial launchers capable of lofting over 10 metric tons to LEO.  I restricted my view to existing rockets,  or rockets that would be flying within a year or so.  That list is summarized in Figure 1 (below at end of article),  along with launch prices I found on the internet and corrected for inflation.  All the payload and cost data are vintage 2012,  not 2013.  The Falcon-9 shown is the older one,  not the new configuration that launched a few days ago.

I roughed-out some rocket engine ballistics from characteristic velocity (c*) data I found in a 1970-vintage edition of the venerable old Pratt & Whitney “Vest Pocket Handbook”.  Thrust coefficient I read graphically off an old chart for gas specific heat ratio 1.20,  at an arbitrary pressure ratio,  and already corrected for an effective bell half angle loss factor of 0.983.  Those data are given in Figure 2,  and lead immediately to estimates of engine specific impulse and exhaust velocity.     ---  updated 10-4-13 for clarity,  not fact

I see no reason to develop a new crew capsule for this application.  And for a lunar trip,  no habitat module is needed.  There are 3 capsules currently in development:  a manned version of Spacex’s Dragon,  the Boeing CST-100,  and NASA’s Orion.  The one closest to flying is probably Dragon,  since it is already operational as an unmanned cargo carrier.  So I went with manned Dragon,  which should fly for NASA by 2017,  and could probably really fly as early as 2014 or 2015. 

The lander need not be very large.  The old Apollo LM was about 15 tons.  NASA had proposed a new 45 ton lander named Altair,  for its cancelled Constellation lunar program.  We have learned a lot since Apollo about constructing landers,  and they need not be staged,  although that can reduce lander ignition weight.   A one-stage design is potentially reusable,  although I did not require reusability here.  I just picked two arbitrary weights:  20 and 30 metric tons.

These considerations lead to a set of modules to be built,  proof-tested,  shaken-down in tests,  and then produced for the mission.  Note that there is nothing in the way of new technology here.  This is just hardware designed with already-existing technologies.  The necessary modules are shown in Figure 3,  with the lander still a rough concept. 

One can string together a service module,  a propellant module,  and a manned Dragon together as the manned transfer vehicle from LEO to LLO,  and back with free return.  This is shown in Figure 4.  This does not even require assembly by docking in LEO,  since one Falcon-Heavy has more than enough capability to orbit this entire vehicle as its payload. 

If one designs to a 20 ton lander,  the unmanned one-way lander transfer vehicle should look about as depicted in Figure 5.  This is a second service module,  two propellant modules,  and the lander itself,  totaling about 59 metric tons.  You could launch this into LEO in two shots,  either one Falcon-Heavy and one Atlas-V/Delta-IV/Proton,  or two Falcon-Heavies.  Then dock them and make connections. 

If instead one designs to a 30-ton lander,  the unmanned transfer vehicle would look about as depicted in Figure 6.  This is a service module,  three propellant modules,  and the bigger lander,  totally about 82 tons in LEO.  This would require on-orbit docking and connections of payloads sent up by two Falcon-Heavies. 

In Figure 7,  I have summarized the performance of both the 20 ton and 30 ton landers,  in two modes.  These are a two-way trip manned,  and a one way trip unmanned as nothing but a cargo delivery vehicle.  I worked out what the total payload could be,  for both sizes and both modes,  a total of four cases. 

I also worked out a guess for the weight of 7 crew with two weeks’ supplies.  The difference is the cargo in addition to the crew.  For the two-way manned mission,  the same cargo allowance coming down can be carried going back up.  For the one-way unmanned mission,  the entire payload can be cargo ferried down.  Given one manned and two unmanned vehicles,  it is possible to send several tons of cargo down with a big crew. 

As already discussed,  this is not a “gravy-train” new-technology development project.  Those rarely produce anything that actually flies.  A summary of what is needed is given in Figure 8,  along with a wild guess for how much it might cost,  if done by the “right team”.  I didn’t consider production costs for the various modules,  but at several million dollars each,  they would be “small change” compared to the uncertainties in the rest of this. 

I used these numbers to estimate costs for the 20 and 30 ton lander cases,  and for the 2 vehicle vs 3 vehicle cases.  I summarized costs and cargo deliverable in Figure 9,  as a sort of “bang for the buck” chart.  In all four cases,  the entire crew of 7 goes to the lunar surface for about 2 weeks.  The two-vehicle cases are around $0.8 billion,  and the three-vehicle cases are around $1.1 billion. 

These figures are very crude,  but the point is,  a big crew and a lot of cargo can be sent to the moon for a price that is a lot closer to 1 or 2 billion dollars than it is to 10-50 billion dollars. 

That makes possible the start of a permanent lunar base of some kind!  And for a lot less money than most folks would believe! 

This same kind of thinking,  free of crippling preconceptions,  is what must take place to design an affordable manned Mars mission,  or any other deep-space mission. 

It’s just that flying into deep space is a lot different from flying to the moon,  because of the travel times. 

Figure 1 – Commercial Launchers Existing Now or by 2014, 10+ Tons to LEO-Capable

Figure 2 – Rough-Out Ballistics for a Small LOX-LH2 Vacuum Engine of Modest Chamber Pressure

Figure 3 – Modules Needed, with Maximum Commonality of Tanks and Engines

Figure 4 – The Manned Modular Transfer Vehicle

Figure 5 – Unmanned Modular Transfer Vehicle for a 20-Ton Lander

Figure 6 -- Unmanned Modular Transfer Vehicle for a 30-Ton Lander

Figure 7 – Estimated Performance of the Lander Designs

Figure 8 – Summary of Needed Developments

Figure 9 – Summary of Mission Costs and Returns, Versus Options

Updated 10-4-13:  if the common engine is in fact at least semi-reusable (meaning capable of a few dozen ignitions and burns),  there is no reason the lander engines cannot push propellant modules to the moon,  without a service module at all.  Initially,  I failed to free my mind enough to notice this. 

This option opens up the possibility of operating the lunar lander more than once,  as a ferry or landing boat,  given an on-orbit refueling capability from tanks.  A fourth and subsequent shot could send a service module and tanks plus payload-to-be-landed one way to the moon for LLO refueling. 

How big a base do you want to build?  How much do you want to spend? 

It's still in the few $B,  not several 10's to 100 $B!  Two Falcon-Heavy launches to LEO per vehicle sent to the moon only cost about $230M in launch price. 

Tuesday, October 1, 2013

“Calibrating” the Ballpark Estimating Method

As stated elsewhere,  the methods I have been using are only ballpark bounding calculations.  To “calibrate” how “good” these techniques really are,  at least for launch to low Earth orbit (LEO),  I ran a generic two-stage rocket broadly similar to the Spacex Falcon-9,  before they upgraded it.  This older version of Falcon-9 was rated at 13 metric tons to LEO,  at low inclination out of Cape Canaveral.  Falcon-9 was a little heavier than 500 metric tons at ignition. 

The key assumptions were 5% gravity loss,  and 5% drag loss for the first stage,  and 5% gravity loss only for the second stage,  on a fast ascent trajectory,  plus 4.5% stage inert weights,  stage payload included in that accounting.  I assumed stage 1 burnout outside the “sensible atmosphere”,  at 3.05 km/s achieved velocity.  LEO was assumed to be 7.79 km/s achieved velocity.  I completely ignored the “boost” effect of the Earth’s eastward rotation.

I did some crude engine ballistics based on characteristic velocity c* at 1000 psia chamber pressures (c* = 5900 ft/s for LOX-RP1,  from the vintage 1970 version of the old Pratt & Whitney “Vest-Pocket Handbook”),  and a bell divergence-corrected thrust coefficient CF chart computed for specific heat ratio 1.20.  The divergence thrust correction factor is 0.983,  pretty much an average 15-degree half angle,  and equivalent to most modern curved expansion bells. 

The first stage is expanded “perfectly” to 14.7 psia (101.3 Kpa,  1013 mbar,  760 mm Hg) backpressure.  The chart says CF = 1.57 at expansion ratio 9.00 when read at pressure ratio 68.  There is a simple relationship among CF,  c*,  and Isp,  leading to Isp = 287.9 s for the first stage.  A very slightly-oversimplified model then estimates exhaust velocity as 2.835 km/s. 

For the second stage,  the ambient backpressure is zero (out in the vacuum).  I looked at the chart for pressure ratio set to an arbitrary 1000:1,  and got CF = 1.826 at expansion ratio 65:1.  For that same 1000 psia LOX-RP1,  c* = 5900 ft/s,  I got Isp = 334.8 s,  and exhaust velocity 3.284 km/s. 

5% gravity loss plus 5% drag loss is a total 10% loss for the first stage,  making the effective required velocity not 3.05 but 3.355 km/s.  That corresponds to a mass ratio of 3.2820,  and a propellant fraction of 0.6953.  The corresponding stage payload fraction is 0.2597.  For a nominal liftoff weight of 500 metric tons,  the stage 1 payload (stage 2 plus the “real payload”) is 129.85 tons.  That ignores any interstage weights,  which might be around a ton,  for a second stage ignition weight of 128.85 tons. 

5% gravity loss on the second stage velocity increment of 4.74 km/s results in an effective velocity requirement on the second stage of about 4.977 km/s.  At the higher second stage Isp,  the mass ratio is 4.55183,  and the propellant fraction for that stage is 0.7803.  For the same stage 4.5% inert fraction,  that leaves 0.1747 for the stage 2 “real payload” fraction.   That’s near 22.51 metric tons for a stage 2 ignition weight of 128.85 tons. 

That payload has to ride inside some sort of protective,  aerodynamic shroud.  For the sake of argument,  assume that shroud also weighs about 1 ton.  Therefore,  the real delivered payload is nearer 21 metric tons for a vehicle massing 500 tons at launch. 

The real Falcon-9 is a little over 500 tons at ignition,  and is rated to deliver 13 metric tons to LEO.  My 21 tons is in the ballpark,  clearly,  but still quite a ways “off” for purposes of “exact” estimates.  All in all,  I’d say these ballpark estimating techniques are actually quite good,  especially for relative-comparison calculations.  But it takes a better model than this,  to really “pin things down”. 

For example,  use the 5%-5% loss factors on the entire trajectory to LEO,  with 4.5% inerts,  and the lower-performing first stage engine performance,  but as a single-stage to orbit vehicle.  It really does calculate as technically feasible,  just at only 1.48 ton payload,  and with propellant tanks “stretched” by 37% (volume) to maintain a 500 ton ignition weight. 

That 1.48 tons gets compared to the 21 tons for the 500 ton two-stage bird.  For one stage versus two,  the launch cost could only be factor 2 lower for the one-stage bird,  at the very most.  On a per-unit-delivered payload basis,  the one stage version will always deliver less mass for a higher cost.  That basic effect is why staging was invented,  over 6 decades ago.