Sunday, July 3, 2022

Early High-Speed Experimental Planes

Updates may be added.  They will be appended below at the end.  Watch for them.


There were lots of official government X-planes and many related craft,  that led to the first generation of supersonic-capable fighter aircraft (the “century series”,  beginning with the F-100 Super Sabre).   There were many “X-planes”,  but only some of them were aimed at this exploration of supersonic flight.  Wikipedia has a good list of all the official government X-planes.  There were some other designations besides “X-plane” that were also a part of this high-speed exploration.  Those are included here.

I have compiled a table of those experimental craft which answered questions about transonic / supersonic / low-hypersonic flight of manned aircraft.  That huge table is given in Figure 1.  To see it (or any other figure) enlarged,  click on any figure,  and you can see all of them enlarged.  There is an “X-out” feature at top right of that screen,  which takes you right back to the article.  That same enlargement procedure applies to any article on this site.  See also Figures 19 and 20 below

Figure 1 – Table of Data for Experimental Craft Intended to Explore Supersonic Flight and Related Issues

There are two very important vehicles not included in the table.  These were never actually built and flown:  the Miles M.52 and the X-20 Dyna-Soar.  Wikipedia has good articles on both of them.  See Figures 2 and 3 for what these craft might have looked like.   

The M.52 was a British effort to produce a manned aircraft that could break Mach 1.  That design had straight wings,  turbojet propulsion,  and a “flying tail”,  meaning an all-moving horizontal stabilizer instead of a fixed stabilizer with a hinged elevator attached to it.  That “flying tail” technology fed directly into the X-1 and the D-558-2 as retrofits,  and in everything since that flies supersonically.  It was a crucial enabling factor!  The M.52 did finally fly as a 30%-scale rocket-powered model,  breaking Mach 1 by a considerable 30% margin,  and only about a year after the manned US X-1 broke Mach 1. 

The X-20 Dyna-Soar was to be a suborbital-hypersonic / orbital spaceplane for the USAF.  The design dates to the late 1950’s,  and the project was cancelled in the early 1960’s,  just as the first examples were coming off the Boeing manufacturing line.  It drew upon the results of the early X-15 tests,  and also the other earlier X-planes and related craft,  plus the warhead re-entry work that went into the intercontinental ballistic missile (ICBM) efforts.  Seven astronauts were secretly chosen in 1960 to fly it,  among them Neil Armstrong and Pete Knight,  both of whom later flew the X-15.   

Figure 2 – The Miles M.52 That Was Never Built

Figure 3 – A Boeing Mockup of Its X-20 Dyna-Soar That Was Cancelled

The nosetip and leading edges of the X-20 were monolithic graphite with inserted zirconia rods to better conduct heat away from the actual stagnation zones.  The skins were to cool by re-radiation of heat,  but were also insulated from the interior to reduce inward conduction.  The pilot’s windscreen was covered by a heat shield skin until entry was over,  which was then jettisoned so the pilot could see to land.  It was to be launched by the then-planned Titan-3 ICBM.  The work on the X-20 project fed directly into the NASA Space Shuttle,  and its derivative the unmanned X-37B spaceplane now being flown by USAF.

These high-speed flight projects began toward the end of WW2 (about 1944),  based on questions,  concepts,  and hardware items that became available from Germany,  before and after that war’s end. These included not only supersonic flight of manned aircraft,  but also the use of swept wings as in the Me-262 fighter,  the merits of tail-less aircraft as in the Me-163 rocket interceptor,  the merits of the delta wing as suggested by Alexander Lippisch,  and the early concepts for spaceplanes,  such as the 1941 Silbervogel (“Silver Bird”) proposal in Nazi Germany for a trans-Atlantic spaceplane bomber. 

Initial questions to be answered:

               What would it take to break Mach 1?  

One very serious choice was “swept or straight wing?”,  with most authorities at the time (end of the war) believing that swept wings would be required to reach supersonic speeds.  The high-subsonic drag-rise regime had already been successfully treated with swept wings by the Germans during WW2,  causing the Allies to try to catch up on that technology after the war’s end.  However,  some others sidestepped the unknowns associated with swept wings,  and attempted supersonic designs with straight wings (like the Miles M.52):  notably Bell with its X-1,  and Douglas with its D-558-1 Skystreak.   

The all-moving-tail pitch control approach for breaking Mach 1 came to the US from the British Miles M.52 project as a possible solution to pitch control problems already encountered at high subsonic speeds (approaching transonic) in a variety of aircraft,  including some fatal power dives in the P-38 “Lightning” just before the US entry into WW2.  As soon as the “all-flying tail” approach was verified,  it was incorporated as retrofits to the Bell X-1 and the Douglas D-558-2 Skyrocket.  The all-moving tail has ever since been the standard for all supersonic craft from the X-2 on,  excepting those with delta wings. 

A final choice for breaking Mach 1 was “what propulsion to use?”  The choices at the time were rocket,  gas turbine,  or mixed (parallel-burn) rocket-and-gas turbine propulsion.  The very early gas turbines used in these craft were still rather weak in thrust performance,  especially if not equipped with afterburners.  Only a little while later,  turbine performance had very rapidly and markedly improved.

The Miles M.52 and the Douglas D-558-1 Skystreak were planned as gas turbine aircraft,  while the Bell X-1 was planned as a rocket-propelled aircraft.  The Skystreak (flown from 1947 to 1953) did break Mach 1,  but only in a dive,  and that was after the X-1 broke Mach 1 October 14,  1947,  in level flight.  The X-1 in its original form flew from 1946 to 1951.  The Douglas D-558-2 Skyrocket was designed as a mixed propulsion craft,  and was actually flown as gas turbine only,  and as rocket only,  as well as mixed rocket and gas turbine propulsion,  in the time frame from 1949 to 1956.   It was a swept-wing design. 

The answers to these initial multiple questions directly supported the “century-series” fighter designs,  notably the F-100 Super Sabre,  the F-101 Voodoo,  the F-104 Starfighter,  and the F-105 Thunderchief.  See Figures 4-6 for images of the X-1,  the D-558-1 Skystreak,  and the D-558-2 Skyrocket.  There are good Wikipedia articles about all of them.  There is even a separate Wikipedia article about the X-1 that lists all the flights it made. That last one covers all three examples of the aircraft,  the date of the flight,  its purpose,  and the identity of the pilot.

Figure 4 – The Bell X-1 (straight-wing,  rocket-powered)

Figure 5 – The Douglas D-558-1 Skystreak (straight-wing, turbojet-powered)

Figure 6 – The Douglas D-558-2 Skyrocket (swept-wing,  mixed propulsion)

               What are the other effects of swept wings besides transonic drag reduction? 

Adverse pitch-up and other instabilities at slow speeds were soon identified while testing experimental aircraft with swept wings.  Understanding these problems helped explain the “Sabre dance” behavior that resulted in several crashes at landing speed with the F-100 Super Sabre.  The possible solutions to low-speed pitch-up and instability with swept wings were wing fences,  leading edge slat devices,  inverse taper,  and variable sweep wings.  All of these but inverse taper eventually became “standard design practices” for swept wing designs,  whether supersonic or not,  although not all together.

The D-558-2 Skyrocket swept wing design was originally intended to reduce transonic drag while reaching Mach 2,  but it mostly got used to investigate pitch stability issues and solutions between transonic and almost Mach 2,  as well as supersonic underwing stores carriage,  during its flights between 1949 and 1956.  It did break Mach 2 in a very shallow dive,  but never in level flight. 

The Bell X-2 was a swept wing design originally intended to probe the high supersonic realm from Mach 2 to Mach 3,  between 1954 and 1956.  It and the revised X-1 designs identified inertia coupling as a serious issue,  much aggravated by the thin air at high altitudes.  The XF-91 investigated the inverse taper distribution solution to low-speed pitchup problems,  flying between 1949 and 1954.  See Figures 7 and 8 for images of the X-2 and the XF-91.  There are good Wikipedia articles about both of them.

Figure 7 – Bell X-2 Being Dropped From a B-50 (rocket-powered)

Figure 8 – Republic XF-91 with Inverse Taper Swept Wings (mixed propulsion)

               Of what utility might a delta wing be? 

The Convair XF-92 (flown from 1948 to 1954) was a gas turbine-powered delta wing design.  Being a very first flying example,  it proved to have poor flying qualities,  and was not at all popular with the test pilots.  But,  it did uncover the surprising fact that thin-section delta wings can provide extremely-high angle of attack capability at low landing speeds!  This unexpectedly very high stalling angle-of-attack (near 45 degrees) was induced by strong vortices on the suction side of the wing,  just downstream of the leading edge.  These are directly linked with the more usual wing root and tip vortices.  Non-delta wings (thin or not) do not exhibit this behavior.

This work supported the Convair F-102 and Convair F-106 delta-wing “century-series” fighters,  the Convair B-58 “Hustler” bomber,  the Rockwell XB-70,  and the Convair F2Y “Sea Dart” for the US Navy,  later re-designated as the YF-7A.  This work has also since supported many of the modern tail-less aircraft designs. 

Figure 9 depicts an XF-92.  There is a good Wikipedia article about it.  

Figure 9 – The Convair XF-92 (delta-wing,  turbojet-powered)

               Are tail-less designs useful at high speeds? 

With simple flight controls,  powered or not,  the answer turned out to be “no”,  as demonstrated by the Northrup X-4 “Bantam”,  and the related history of the US Navy with the Vought F7U “Cutlass”.  The X-4 flew from 1950 to 1953.  Figure 10 is an image of it.  There is a good Wikipedia article about it.

The related experiences of the Navy with the F7U are also contained in a rather good Wikipedia article.  It was initially underpowered,  had unreliable hydraulics powering its flight controls,  and exhibited too many difficulties (verging on instabilities) landing on an aircraft carrier.  That is why the Navy has since avoided most manned tail-less designs except for delta-wings,  such as on the A-4 “Skyhawk”.

However,  the answer could be “yes”,  but only with fully-computerized flight controls,  as found later with fundamentally-unstable things like X-29,   and related to many unmanned craft since. 

Figure 10 – The Northrup X-4 “Bantam” (swept wings,  tail-less,  turbojet-powered)

Questions that came up during testing:

               Inertia coupling in high-altitude thin air 

This was a problem with X-1A-on (including the Chuck Yeager flight to Mach 2.4 in the X-1A in 1953,  that barely avoided a crash),  the X-2 (leading to a fatal crash in 1956),  the NF-104 (leading to the famous ejection and helmet fire experienced by Chuck Yeager in 1963),  the X-15 (also leading to a fatal crash in 1967),  and it is still a problem today with most big fighters at high wing loadings.  See Figure 11 for an image of the Bell X-1A,  Figure 12 for the Lockheed NF-104,  and Figure 13 for the North American X-15.  The Bell X-2 was already depicted in Figure 7.  

Figure 11 – Bell X-1A (X-1B,  X-1C,  X-1D,  and X-1E Were Very Similar)

Figure 12 – The Lockheed NF-104,  As Remanufactured From An F-104A

Figure 13 – The North American X-15,  As Equipped With the “Big Engine”

While termed “inertia coupling”,  this is fundamentally a situation where the forces induced by the aerodynamic flight control surfaces are simply too small for the corresponding inertias of the vehicle,  resulting in too-low (or zero) achievable recovery accelerations.  This is more commonly experienced at extreme altitudes in the very thin air,  where the control forces are inherently low due to low density,  despite the high speeds.  But at very high wing loadings (weight per square foot of wing area),  it can be experienced even at low altitudes. 

With the “Bell X-1A-on” series:  the X-1A,  X-1B,  X-1D,  and X-1E actually flew.  The X-1C was never built.  All of these had a revised pilot cockpit with a more conventional canopy,  and an ejection seat,  unlike the original X-1.  These models represent changes in wing section and instrumentation.  The X-1B was the first of these craft equipped with attitude control thrusters,  that were later included as part of the X-15 baseline design.  Why?  These experimental craft were reaching altitudes above 100,000 feet (30 km),  where the air was simply too thin to permit normal control by means of aerodynamic surfaces.  The X-1E was a rebuild of the X-1 #2 aircraft (“Glamorous Glennis”).  The rest were “built from scratch”.

The NF-104 was a modified Lockheed F-104A Starfighter to add rocket propulsion and attitude control thrusters.  It was intended as a trainer for astronauts who were to fly space planes for USAF,  either sub-orbitally,  or orbitally,  but it was not specifically built as an experimental aircraft,  in and of itself. 

By the time this craft flew,  gas turbine engines had advanced to point of enabling runway takeoff to a rather high-altitude flight envelope “cap” near 50,000 feet (15 km).  In a high path angle “zoom climb”,  adding rocket thrust took the craft well beyond the top of its normal flight envelope,  to near the edge of space,  above (even well above) 100,000 feet (30 km). 

Attitude control thrusters were simply required up there,  for flight control,  so that a nose-first “re-entry” into denser air could be made,  in turn allowing a windmill re-start of the gas turbine engine,  which in turn restarted the aerodynamic flight control hydraulics.  Any failure of this sequence could well prove fatal!  So,  a de-spin chute mounted in the tail was included in the design.

There was a JF-104 prototype for this craft,  which had the attitude thrusters,  but not the added rocket engine.  It first flew in 1959,  reaching around 80,000 feet.  There were 3 mixed-propulsion NF-104 craft delivered.  2 of these flew from 1963 to 1971.  The number 3 craft was destroyed in the famous Chuck Yeager crash in 1963.  The other 2 reached altitudes of 100,000 to 130,000 feet multiple times.

               Is variable sweep really a feasible thing to do? 

The answer turned out to be “yes”,  as demonstrated by the Bell X-5,  flown experimentally from 1951 to 1955,  and as a “convenient chase plane” to 1958.  The wings did not just pivot,  the pivot point also translated longitudinally,  to keep the center of lift near the same relation to the shifting center of gravity,  as the sweep angle changed.   This work supported the swing-wing designs of the General Dynamics F-111,  the Grumman F-14,  and the Rockwell B-1A and B-1B bombers.  There is a good Wikipedia article about the X-5.  A composite image of it at various sweep angles is shown in Figure 14.  

Figure 14 – Bell X-5 Variable-Sweep Craft (turbojet-powered)

               How do we stop adverse pitch-up & instability with swept wings? 

The “band-aids” finally identified are flow fences and leading edge slat devices.  These were explored with the D-558-2 (and inverse taper with the XF-91),  which supported all the modern fighters after the later “century series” (F-105),  and most modern jet airliners.  Inverse taper was,  and still is today,  not used;  fences and slats are,  whether the craft is supersonic or not.  Leading edge offset (as on the F8U “Crusader”) was found later,  and is also often used.

Variable wing sweep is the other basic approach,  explored with the X-5.  This work supported the F-111,  F-14,  and the B-1A/B-1B swing-wing craft.  It works well,  too,  but is heavier.

               What happens above Mach 1? 

This regime was thoroughly explored with the X-1A-on series 1953-1958, the D-558-2 in the interval 1949-1956,  and was attempted with the Douglas X-3 “Stiletto” 1952-1956. 

Of these,  the X-3 was far less than successful,  mostly because the turbine engines of the time were far too limited in performance.  It exceeded Mach 1 in a steep dive,  but never came close to Mach 2,  much less the Mach 3 for which it was originally intended. 

The X-3 was a very “pointy” shape powered by two gas turbine engines,  and fitted with stub wings.  There is a good Wikipedia article on it.  An image of this craft is given in Figure 15.

Work in this regime with the other planes supported the “century-series” fighters,  and later. 

Figure 15 – The Douglas X-3 “Stiletto” (turbojet-powered)

               What happens above Mach 2? 

This regime was explored by the X-1A-on 1953-1958,  and the X-2 1952-1956.  The X-3 was intended to explore this regime up to Mach 3,  but proved to be very unsuccessful above Mach 1.  This work supported the later “century-series” fighters,  and everything faster since then.

               What happens above Mach 3? 

This regime is quite extreme,  and was explored with the X-2 1952-1956,  and by the X-15 1959-1968 in 199 flights.  Test pilot Mel Apt was killed in the X-2 in 1956,  after reaching Mach 3.2 at 65,500 feet.  (At somewhat slower speeds,  this same craft exceeded 120,000 feet.)  The vehicle tumbled upon Apt attempting a coordinated turn at those high-speed,  thin-air conditions,  due to inertia coupling.  He attempted to escape,  but that escape system design proved to be utterly inadequate!  The cockpit section did indeed separate,  but Apt was unable to bail out of it before its impact.  I’ve seen the cockpit camera film of him being pinned and pounded,  by too much cockpit section spin,  to bail out.  His death in that crash ended the X-2 project,  which at the time was the only way to reach Mach 3.

The X-15 rolled out on 1958,  and first flew in 1959,  with the “two small engines” (XLR-11’s).  There were 3 vehicles built.  It first flew with the single “big engine” (the XLR-99) in 1960.  From there it explored the regime from Mach 3 to Mach 6.7 at around 100-120,000 feet,  and the really high altitudes of 50-60+ miles at somewhat slower speeds (nearer 2700 mph  ~ Mach 4). 

The number 3 vehicle was lost along with its pilot Michael Adams on flight 191 in 1967,  due to the effects of inertia coupling (a “hypersonic spin” as the Wikipedia article terms it).  The aircraft finally broke up during re-entry at about 60,000 feet,  scattering debris across 50 square miles.  Adams was later posthumously awarded his astronaut wings for exceeding 50 statute miles altitude on this flight. 

The number 2 vehicle was rebuilt after an incident,  with a little bit of extra length,  a ceramic heat shield coating,  and two external propellant drop tanks,  as the X-15A-2.  Pilot Pete Knight took this X-15A-2 version to a record speed of about Mach 6.7 (4520 mph) at around 19.3 miles altitude (near 100,000 feet) in 1967,  with a scramjet test article affixed to the ventral fin stub.  See Ref. 3.

This record-setting max speed flight found shock-impingement heating to be a very real danger,  with a scramjet test article as an adjacent nacelle on the ventral-fin stub.  This was because the shock wave off the test article inlet spike very nearly cut the tail section off this craft,  with high shock impingement-amplified heating rates.  As it turns out,  the localized heating at the shock wave impingement location was around factor 5+ times higher than what one would otherwise expect at that Mach 6+ flight speed. 

The 199 X-15 test flights supported the design of the Lockheed SR-71/YF-12A,  and also earlier the Boeing X-20,  and later-on the NASA Space Shuttle.  An image of the X-15A-2 is shown in Figure 16.  The envelope of speeds and altitudes achieved during the X-15 program is given in Figure 17.   

Figure 16 – The North American X-15A-2 With a Scramjet Test Article On Its Ventral Fin Stub

Figure 17 – Speeds and Altitudes Achieved During the X-15 Program

               How can we train space plane pilots? 

In the late 1950’s through the mid 1960’s this topic was a real issue for the USAF.  After all manned spaceflight was taken from USAF,  and given to NASA,  starting about 1965,  this became much less of an issue.  It remained only a bit of an issue with the lifting body research vehicles through the end of the 1960’s into the 1970’s,  for USAF.  It took some time (to about 1969-1970) before the last of the USAF manned space programs (the “Manned Orbiting Laboratory” or MOL),  was terminated.  A lot of different reasons for this shift to NASA were involved.  That is not explored here.

Training the space plane pilots took two forms:  (1) simulating the low-speed landing characteristics of low lift/drag vehicles,  and (2) flying the “typical” high-speed/high-altitude flight profiles for sub-orbital / orbital ascents.  The item (1) issues were directly relevant to the NASA Space Shuttle vehicle and the lifting bodies.  The JF-104 and “factory-stock” F-104’s were adequate to address this. 

The item (2) profiles were based on the flight profiles for the X-1,  the D-558-2,  and especially the X-15.  Until the X-15,  these involved flying to higher altitude at high subsonic speed,  diving to gain supersonic speed,  then pulling up into a supersonic “zoom climb” at a high path angle above horizontal.  From there,  rocket propulsion took one out of the sensible atmosphere.  Upon re-entry,  the idea was to restart the turbine engine,  and fly back to a conventional landing.  That is the design flight profile of the NF-104.  It was the zoom climb portion into the extremely thin air that simulated space plane ascents. 

It was actually the X-15 that ended that dive/pull-up/zoom climb approximation.  Its “big engine” XLR-99 propulsion was “good enough” to delete the dive to attain supersonic speed.  The X-15 simply accelerated the vehicle to higher supersonic speeds at lower altitudes,  or to lower supersonic speeds at higher altitudes.  X-15 or X-15A-2,  it made no real difference,  except in the exact numbers achieved.  The “big” XLR-99 rocket engine in the X-15 really was that good an engine/airframe combination!   

Spaceplane pilot training in the USAF began with “stock” F-104A aircraft,  and the NASA-modified JF-104 that had attitude thrusters.  It continued with the 3 delivered NF-104 craft that had mixed rocket / turbine propulsion and attitude thrusters.  These used the Rocketdyne AR2-3 engine that burned jet fuel with high test hydrogen peroxide.  These flights supported the X-20 Dyna-Soar that was cancelled,  and also the NASA Space Shuttle.  The USAF astronauts mostly went to NASA for the Space Shuttle program.

Rocket and Turbine Engines Used In These Planes

A variety of rocket and turbine engines were used in these early experimental planes.  Those data are summarized in Figure 18.  There are good Wikipedia articles about all of these engines.

In particular,  gas turbine technology was in its infancy when these efforts started at war’s end.  Pressure ratios and stage counts were low,  thrust-to-weight ratios were low,  and thrust-specific fuel consumptions were high.  The very earliest engines did not have afterburners. 

Performance factors improved at a very rapid pace.  By the time the J-79 was installed in the F-104 fighters from which the JF-104 and the NF-104 were derived,  performance had approached what we might consider relatively modern levels.  This showed up in vastly-improved aircraft performance.  

The rocket engines were initially rather small and not throttleable in the modern sense.  The initial designs used the same propellants as were used in the German V-2 ballistic missile (ethyl alcohol and liquid oxygen).  The XLR-11 that powered the X-1 and some other craft (including the initial X-15’s) addressed throttleability by being 4-chambered,  with the ability to run any or all of the chambers at their fixed thrust levels.  That gave the XLR-11 a sort of “incremental throttleability”,  ranging from no thrust to 6000 lb of thrust,  in 1500 lb increments. 

Later,  with the XLR-25,  XLR-99,  and AR2-3 engines,  continuous throttleability from 50 to 100% became possible.  In particular,  the XLR-25 in the X-2 had two independently-operating chambers of two different max thrust sizes (10,000 and 5000 lb),  each throttleable from 50% to 100%.  Essentially,  by selecting which chamber was active,  and what its throttle setting was,  the engine could be managed anywhere from 2500 lb of thrust all the way up to the maximum sum of 15,000 lb of thrust. 

The biggest of these was the XLR-99 that powered the X-15 as its “big engine”.  This had enough max thrust to simply push the craft directly to the speed and altitude desired for the test.  It was throttleable from 50% to 100% of its 57,000 lb thrust rating.  This engine burned LOX and anhydrous ammonia.

Figure 18 – Data About the Engines Used

Some Extra Details

The data table in Figure 1 is quite large.  I have broken it into two figures here,  for the avid reader.  Figure 19 is a data table relating which vehicles answered which questions,  and what the outcomes were.  That is a sort of testing logic “diagram”,  if you will.  

Figure 19 – Which Craft Did What?

Figure 20 is a timeline diagram showing in which years each of these planes flew.  Be aware that the end of the timeline scale is nonlinear.  

Figure 20 – When Did They Do It?

On a personal note,  I was born in 1950,  right in the middle of the X-1,  D-558-1,  XF-91,  XF-92,  and D-558-2 tests.  I clearly remember Sputnik.  I started aeronautical engineering school in 1968,  just as the X-15 flights were ending.  My Dad was an aeronautical engineer.  I quite literally grew up with this stuff. 

How Might These Early Experiences Support Today’s Interest In Hypersonics For Ascent to Orbit?

The key thing for that issue is the plot in Figure 17,  showing the flight envelope conditions explored with the X-15,  relative to the feasible paths for reaching orbit with aerodynamic ascent.  That set of aerodynamic ascent paths is actually quite narrow!  100 miles is about 161 km,  for those who prefer metric.  Similarly,  1000 mph is about 1610 km/hour (divide that km/hour by 3600 to get km/sec).

Fly too low at too high a speed,  and the aeroheating is simply too excessive to endure,  in any practical way.  That is exactly what the figure says!  Fly too high,  and the air is too thin to provide the lift for an aerodynamic ascent of any kind.   That is also what the figure says!  And,  if you don’t have enough air to generate lift,  you also won’t have any thrust,  if you are trying to use airbreathing propulsion in the form of a scramjet.  See Ref. 1 for more information about that.  If rocket,  the free-expansion nozzle designs are also limited by streamline divergence effects at very high altitudes.  That is covered in Ref. 2.

One should note that vertically-launched rocket vehicles fly well above this aerodynamic ascent path:  they leave the sensible atmosphere at relatively low speeds,  and fly thrusted gravity-turn trajectories to orbit,  essentially in vacuum most of the way.  Such vehicles usually are moving only around 1500 to 2000 mph,  at altitudes around 25-30 miles or so.  That is actually well to the left of the spaceflight conditions explored by the X-15 above the aerodynamic ascent corridor. 

The X-15 achieved its maximum speed of about 4520 mph a bit over 19 miles up (about Mach 6.7 at about 100,000 feet),  which is right at the left end of the “straight” portion of the aerodynamic ascent corridor.  That is exactly where it identified shock impingement heating as a very extreme risk,  thus severely constraining the vehicle configurations that might be feasible for such an ascent (or re-entry)

The estimated effective temperature of the shocked plasma sheath surrounding the vehicle at this flight condition is near 2000 K = 3150 F.  That is the effective driving temperature for the convective heat transfer to the vehicle,  anywhere on its surfaces. Only the heat transfer coefficient varies across these locations,  by somewhere around a factor of 10+ between stagnation zones and wake zones.

Parallel-mounted nacelles affixed to pylons or wings simply cannot be made to work in this flight regime,  as the localized heating where the nacelle shock wave impinges on the wing or pylon leading edge is some factor 5+ higher than one would otherwise expect at that speed and altitude,  even for leading edge stagnation zones!  Which in turn is why I don’t think the proposed Skylon airframe,  with its wingtip-mounted engine nacelles,  would ever survive re-entry from orbit!  The inlet spike shocks impinging upon the adjacent wing leading edges,  will simply cut the wings clean off!  In a matter of seconds!  That kind of rapid damage is what the X-15 found in that 4520 mph flight!  Again,  see Ref. 3.

Some will point to the shape of the SR-71/YF-12A to deny what I say about shock impingement heating risks.  But they are wrongthat craft was NOT hypersonic!  It typically flew Mach 3.2 at about 85,000 feet altitude.  If it ever exceeded Mach 3.3,  the pilots immediately took action to slow back down,  precisely to avoid engine and airframe overheat damage!  I know,  I have talked to a couple of them. 

Despite the shock impingement effects magnifying the heating rates (not the temperatures!),  the SR-71 could withstand that heating,  because the driving temperature was only about 810 F = 700 K at that M3.3 overspeed limitation.  Even beta-phase titanium could withstand being fully soaked out to that temperature.  But no more than thatThe amplified heating rates just got it fully soaked-out faster!

The far end of the “straight” portion of the corridor for aerodynamic ascent is about 13,000 mph at about 25 miles altitude,  according to Figure 17.  Conditions there are quite a bit more extreme.  The effective temperature of the shocked plasma sheath about the vehicle is some 5810 K = 10,000 F.  Heat transfer coefficients are reduced a bit relative to the left end,  because the air pressure is lower at that higher altitude.  This condition is actually probably at least fairly close to the max heating rate point. 

On this aerodynamic ascent corridor,  the actual “entry” toward low orbit is just about 17,000 mph at just about 70 miles.  The shocked plasma sheath effective temperature there is about 7600 K = 13,200 F.  The density is so low that the heat transfer coefficients are actually quite low.  Therefore,  the peak heating rates are lower down,  earlier in the trajectory.  No surprise there!  (Re-entry is similar.) 

However,  these numbers do so very clearly indicate a very severe thermal problem to be managedif an aerodynamic ascent to orbit is the goal.  This is rather like re-entry in reverse,  except the exposures are longer,  especially if your acceleration capability is limited!  And it will be very limited,  if you use only airbreathing propulsion for this task.  Ref. 1 explains why,  in terms of the “service ceiling” effect.


I wrote all of these.  They are all here on this “exrocketman” site.  For finding references located on this site,  simply jot down the date and title.  There is a navigation tool on the left side of the page.  Click on the year,  then the month,  then the title if need be.  Ref. 4 has a much more comprehensive list of articles,  arranged by topic,  concerning many different topics.  Feel free to look at any or all of them. 

#1. 1 June 2022,  “About Hypersonic Vehicles” (covers both rocket and airbreathing ascent)

#2. 12 November 2018,  “How Propulsion Nozzles Work” (covers conventional and free-expansion)

#3. 12 June 2017, “Shock Impingement Heating Is Very Dangerous” (Mach 6.7 in X-15A-2)

#4.  21 October 2021,  “Lists of Some Articles By Topic Area” (much more comprehensive lists)

ADDENDUM (update 7-4-2022)

Here are some images of some of the aircraft these early experiments supported or confirmed.  With one exception (the XB-70),  these were not “experimental planes”,  although perhaps a few of them should have been experimentals.  There were serious difficulties with the F-102 Delta Dagger and the F7U Cutlass.  The Dagger had to have its fuselage area-ruled and its intakes revised.  The F7U was quickly replaced by the F8U Crusader.

First generation swept-wing supersonic fighters

North American F-86 “Sabre” transonic fighter (Mach 1.02-capable),  a mainstay in Korea

Initial versions of the F-86 had leading edge slats.  Those did not suffer from low-speed pitch-up and instability problems.  Later versions deleted the leading-edge slats in favor of fixed leading edges for more fuel volume and range.  Not surprisingly,  those versions suffered from the infamous “Sabre dance” that afflicted the F-100 Super Sabre.  Your only option is never to fly that slow at landing (at that high an angle of attack) with plain swept wings.


North American F-100A Super Sabre (served through Vietnam)

McDonnell F-101 Voodoo  (Mach 1.72-capable fighter bomber and recon)

Both of these aircraft (F-100 and F-101) lacked leading edge slats and suffered severely from low-speed pitch-up and instability problems.  Yet both served successfully for long times.  You just don’t fly slow at landing in plain swept-wing craft like these.


Lockheed F-104G Starfighter (this one with Netherlands Air Force)

F-104’s were notoriously difficult and dangerous to fly,  not because of swept wing problems,  but because of inertia-coupling problems induced at high wing loadings by the stub wings,  plus inherently high landing and takeoff speeds.  The X-3 suffered the same troubles.  

Republic F-105D Thunderchief with Full Load of 750 LB Bombs in Vietnam

These still had low-speed pitch-up and instability problems,  because they had plain swept wings.  You just do not fly so slow at landing,  and thus avoid the troubles.


Vought F8U “Crusader” – served as top fighter 1957-1987

The “Crusader” introduced the offset wing leading edge as an alternative to wing fences.  It worked.  The other innovation was the variable-incidence wing,  which reduced nose-high visibility problems during landing,  extremely critical for landing on a carrier deck.  This aircraft entered service armed only with four 20 mm cannons.  Missiles were added later,  after they became available.  It served some 30 years as the Navy’s top combat air patrol fighter.


Grumman F11 Tiger (Mach 1.1-capable,  replaced subsonic F9 Cougar)

These F-11’s had leading edge devices,  as can be seen in the picture.  These were good-handling aircraft,  and were flown by the “Blue Angels”.  These were capable of low supersonic speeds,  replacing the swept-wing F9F-6 Cougar,  which was high subsonic.  The earlier F9 series were the subsonic straight-wing Panthers.  All were carrier fighters and attack planes.

Early delta-wing aircraft

Convair F-102 Delta Dagger with Area-Ruled Fuselage and Revised Intakes

This one entered service with severe drag and thrust problems.  It took applying the “area rule” to solve the drag problem.  Revising the air intakes fixed the thrust problem.  Weapons were carried internally in a weapons bay.  The usual weapon was the conventional-tipped air-to-air “Falcon”,  but the nuclear-tipped “Genie” could also be carried.  This aircraft was intended as an interceptor against incoming enemy bombers.

Convair F-106 Delta Dart

A bigger engine and better design from the outset made this one far more successful than its predecessor,  the F-102.  Both are the same basic idea,  however:  a delta-wing supersonic interceptor with internal weapons storage.  Again,  the weapons were the “Falcon” and the “Genie”.  The advantage of this design approach is at landing,  where very high angle of attack capability reduces touchdown speed.  There is a nose-high visibility problem,  aggravated by the high angle of attack at touchdown. 


Convair B-58 “Hustler” Supersonic Bomber (served,  but soon replaced by older Boeing B-52 as strategic bomber)

This one entered service as a supersonic-capable bomber (for penetrating Soviet air defenses) that otherwise cruised at very high subsonic speeds.  It could not maintain supersonic flight on 3 engines,  though.  Loss of an engine would cause a fatal problem with center of gravity vs center of lift,  as the aircraft suddenly decelerated subsonic.  It would turn broadside and break up.  

This was due to an extreme shift in the center of pressure between what obtained subsonically,  and what obtained supersonically.  Once supersonic,  fuel was pumped aft for a more rearward center of gravity.  If it then suddenly went subsonic,  the aircraft was unstable,  pitched up broadside,  and broke up faster than its crew could react. 

Despite initial pilot resistance,  the solution was an automatic fuel pumping system to adjust center of gravity forward again,  immediately upon detecting an engine problem.  The real issue to resolve was that this had to be done much faster than humans could react. It had to be automatic.

The B-58 entered service after the B-52,  but eventually was replaced by the older B-52 as the preferred strategic bomber.  This is because the higher speeds and lower internal volume of the B-58 incur more drag and more fuel consumption with less fuel capacity,  rendering the B-58 capable of only a 1-way flight from the US to Russia.  Orbiting at the fail-safe point,  awaiting the “go code”,  was simply not an option with the B-58.  But it was,  with the B-52.  Which loiter capability is really what made the B-52 the preferred strategic bomber.

Rockwell XB-70 Valkyrie (Mach 3 bomber experimental prototype – 2 built,  1 crashed)

This design (XB-70) was intended for an even more-invulnerable high-supersonic penetration of Russian air defenses,  similar to the B-58,  but much faster.  It suffered the same basic mission problems as the B-58:  (1) it was a one-way trip,  and (2) there was no orbiting at a fail-safe point awaiting a “go code”.  Development was terminated in favor of Mach 3 experimental work,  accordingly.  The Russian Mig-25 was intended to be the interceptor to counter it.  That interceptor was the fastest gas turbine-powered aircraft ever flown,  at Mach 3.5 max,  but with an inherently short engine life.

Convair F2Y Sea Dart (dual-ski version),  more-or-less an experiment

This one probably should have been an “experimental”.  The need for waterborne ski takeoffs “evaporated” in favor of catapult launch from aircraft carriers with angled flight decks.  

Douglas A-4 “Skyhawk” (delta wing but also with horizontal tail)

This one combines the high angle of attack capability of the delta wing with the positive pitch control of actually having a horizontal tail.  It was a very,  very successful design.  One nickname was “Heineman’s Hot Rod”.  My Dad working at Vought knew Ed Heineman at Douglas. 

I have seen one of these A-4’s making multiple low high-speed passes at the Mojave airport.  It was doing just about 0.9 Mach,  at just about 3 or 4 feet above the runway.  The aircraft would have had to climb,  in order to lower its landing gear.  Very impressive!

Tailless designs


Vought F7U “Cutlass” (carrier fighter,  not so successful,  replaced by Vought F8U “Crusader”)

Landing on aircraft carrier flight decks in all conditions proved to be a problem with tail-less aircraft designs,  if fitted with simple flight controls,  powered or not.  The deserved nickname for this one (the F7U) was “widowmaker”.  The Navy has since for several decades avoided tail-less designs for its manned carrier aircraft.  Tail-less was a sort of “design fad” in the mid to late 1950’s,  based on the notion of parasite drag reduction.  It proved to be less than desirable,  ultimately.   Much more effective pitch control with a horizontal tail proved to be far more important.  

Vought Regulus 1 Missile (successful but superseded by SLBM) 

Vought Regulus 2 Missile (from USS Grayback,  superseded by SLBM)

Not having to land on an aircraft carrier flight deck,  the two tail-less Regulus missile designs were actually rather successful despite the pitch control deficits incurred by the tail-less designs.  They were simply superseded by the submarine-launched ballistic missile (SLBM),  initially Polaris.  Regulus 1 made operational missile patrols on US Navy submarines.  The faster Regulus 2 was to replace it,  but the submarine-launched cruise missile concept was superseded by SLBM’s before Regulus 2 could be fielded.


Northrup SM-62 Snark Missile (superseded by SAC ICBM)

The tail-less Snark strategic cruise missile had several real control problems,  including pitch control,  leading to the expression “Snark-infested waters” at the site where it was tested.  Fortunately,  it was also superseded by the intercontinental ballistic missiles (ICBM’s) of the Strategic Air Command (SAC).

Swing-wing designs

Unlike the other aircraft pictured here,  these were not first generation anything.  The swing-wing concept was not used in first generation designs.  Its use came later,  starting with the B-1 bombers.   

North American Rockwell B-1A Prototype (4 Built)

The four B-1A prototypes proved-out the basic concept.  Some changes were made to enhance low observability for the B-1B production model.  Avionics systems integration in the B-1B was done by USAF instead of Rockwell.  That did not turn out so well:  there are 3 systems critical to successful penetration of very dense air defenses.  Only 2 of these systems can be “on” simultaneously,  because of severe interference problems that “crash” all of them.  Fixing this required all-new avionics,  which were about 2/3 of the aircraft price,  so that was never done.  Accordingly,  the B-1B cannot be,  and is not,  used for penetrating very dense air defenses. 

North American Rockwell B-1B “Lancer” (100 Built and Most Still Serving)

General Dynamics F-111 “Aardvark” Fighter-Bomber,  mostly used as a bomber,  now retired

The F-111 design came from the TFX (“tactical fighter experimental”) effort aimed at a fighter-bomber that both USAF and USN were to use.  The Navy never accepted the F-111B version intended for carrier usage,  because it was simply too heavy for the ship’s flight decks.  The F-111A version served many years for the USAF.  It was the first to feature a terrain-following radar (TFR) to control the aircraft at very low altitudes,  for better penetration success into dense air defenses.


Grumman F-14D “Tomcat”,  long-serving but now retired

The F-14 is famous enough to need no comment here.  Its replacement is the fixed-wing FA-18,  which fills not only the fighter role,  but also the attack bomber role.

Swing-wing designs work,  they are just heavier than fixed-wing designs.  If you have enough thrust,  you can fly pretty-much fly anything that is heavy,  albeit with due regard to inertia coupling.  Swing-wings do allow much slower landing speeds for a craft that can fly high supersonic with highly-swept wings.  You just won’t like the range and payload penalties induced by the weight and volume of the swing-wing mechanism.   And you might not like the enhanced risk of low-altitude inertia coupling,  if you let the wing loading get too high. 

Wednesday, June 1, 2022

About Hypersonic Vehicles

For hypersonic flight vehicles of any kind there are two fundamental problems that require solutions:  aeroheating and propulsion.  Of these,  the aeroheating is the more important!  If the design concept does not have a thermal management solution for the extreme aeroheating,  then regardless of any propulsion,  the design concept has no credibility at all!  See References 1 and 2 for more information.

The simplest propulsion solution is rocket.  This can take two distinctly-different forms:  (1) a hypersonic glider vehicle propelled by a large ballistic missile that is staged-off after doing its job,  and (2) a tactical-size hypersonic missile with its rocket propulsion on-board (which does not preclude adding a small staged-off booster rocket).  The range and top speed of the big ballistic missile-boosted concept is not inherently limited,  while the tactical-size vehicle with on-board rocket propulsion is very limited by the weight and volume constraints pertaining to whatever launches it.

However,  if you have a rocket on board the hypersonic vehicle,  you will have to protect it from the heat that conducts inward from the hot lateral skins (and the nose tip and any leading edges).  Same is true of any payload-related items.  These skin surfaces will have to operate in the 1000-1500 F (540-820 C) range in order to radiate enough heat away,  to balance the aeroheating input at stratospheric altitudes,  since the inward conduction simply must be interrupted to protect the rocket (and any other) internal components. 

The maximum recommended service temperature for titanium is 750-800 F (400-430 C).  It is very definitely NOT a high-temperature material,  despite what so many seem to think based on its use in the SR-71.  That vehicle was limited to speeds under Mach 3.3,  it was definitely NOT hypersonic!

Failing that re-radiation balance,  you simply will have to actively-cool those skins!  Why?  It is probably very infeasible to dump such large amounts of inward-conducted heat into the rocket propellant,  particularly if it is solid propellant,  and for several very compelling reasons.  See Figure 1 below. 

Hypersonic Airbreathing Propulsion

If one uses airbreathing propulsion to extend the range of a tactical-size hypersonic vehicle,  that inherently opens multiple further aeroheating problem issues,  that are simply not faced by a rocket vehicle or a glider.  You will (at the very least) have air inlet features and a combustor and nozzle to consider. 

The air capture cowl is aeroheated both inside and outside,  but can effectively radiate only from the outside,  and opportunities for conduction are geometrically absent,  so a higher equilibrium material temperature than a lateral skin is simply inevitable!  Internal inlet ducts obviously cannot radiate to the environment,  and must not conduct inward,  so they will require active cooling!  The combustor and nozzle will also require active cooling,  being either within the airframe unable to radiate,  or else actively aeroheated on the external surface if exposed.  Figure 1 below illustrates these items,  too.

Your choices for airbreathing propulsion are really quite limited:  ramjet,  scramjet,  and some sort of combined cycle propulsion (rocket/turbine,  ramjet/turbine,  or scramjet/turbine).  Turbine alone will simply not work at hypersonic speeds:  the fastest operational gas turbine engine was the short-life design in the Mig-25,  at Mach 3.5 maximum in the stratosphere!  At Mach 5 in the stratosphere,  the captured air temperature exceeds most turbine inlet temperature limits without burning any fuel at all!


Ramjet works quite well at Mach 3 to 4 in the stratosphere,  and can be readily designed to survive Mach 5 speeds with the modern technologies.  If the vehicle is low drag and the ramjet engine is ~100% of the vehicle frontal blockage area,  Mach 6 is demonstrably attainable,  maybe Mach 7.  However,  this is possible only with great difficulty solving the Mach 6-7 aeroheating problems,  especially those associated with the inlet capture,  internal inlet duct,  and inlet duct-mounted fuel injection hardware!  The only currently-viable technology solutions for heat protection at conditions like these are one-shot ablatives.  See References 3 and 4 for lots more information about subsonic-combustion ramjet.


Scramjet may well now be almost ready-to-apply technologically.  It has flown experimentally,  but not yet in vehicles with full aeroheat protections in place.  These tests were conducted at altitudes so high that the heat transfer coefficients were reduced by the low air density,  reducing the severity of the thermal management problem for the experimental designs (X-43A and X-51A,  plus an earlier Australian test).  At such altitudes,  airbreather frontal thrust densities are too low to provide any climb rate,  or any acceleration capability in level flight.  These are 100,000 to 130,000 foot (30-40 km).  

That high altitude effect has very serious implications for using airbreathing propulsion for flight-to-orbit,  since the airbreather (any airbreather!) will always have insufficient thrust to fly,  as the air thins further,  just because the ambient pressure is so low!  This is really why the X-30 project failed!  (Rocket actually has slightly higher thrust at altitude than at sea level,  but only if a conventional nozzle is used;  see Reference 5 for why that last statement is true.) 

Combined-Cycle Propulsion Issues

There are two fundamental problems with any (and all) combined-cycle engine designs that use gas turbine as one component.  Problem (1):  the inlet diffuser and nozzle geometries required by the gas turbine are fundamentally incompatible with scramjet (but not necessarily ramjet),  as illustrated in Figure 2 below.  Problem (2):  there must be zero airflow through the turbine component,  once the max safe speed for it (only about Mach 3 to 3.5) is surpassed.  The risk is overheating the turbomachinery.

Regarding problem (1),  gas turbines require low subsonic delivery speed at the compressor face,  which means the post-capture inlet is a divergent diffuser duct that is nearly all-subsonic.  Ramjet demands something similar,  although not geometrically identical.  (See Reference 6 for an explanation of those differences.)  Scramjet demands a nearly constant-area “isolator duct”,  that is all-supersonic to its outlet!  The very small divergence in that “isolator duct” merely offsets boundary layer thickening.  Variable-inlet-geometry hardware is well known to be both voluminous and heavy.  See again Figure 2. 

The turbine outlet speed from a gas turbine is also generally subsonic.  The nozzle must neck-down to a minimum throat area to reach sonic speed,  and may have a very modest supersonic expansion ratio.  Ramjet demands something very similar,  but usually somewhat larger.  Jet fighters use moving “turkey feathers” that are air-cooled to accomplish the throat and exit area variations needed.  Scramjet cannot have a neck-down to a min-area throat,  only a supersonic expansion!   And,  at high supersonic and hypersonic speeds,  there is simply no such thing as “cooling air”,  which means no variable geometry nozzle technological solutions exist in anything resembling a ready-to-apply form!  Again,  see Figure 2.

Regarding problem (2),  there must be designed-in some way to bypass all (ALL!) the inlet air around the gas turbine component,  directly to the ramjet or the scramjet component.  Otherwise,  the hot high-supersonic and hypersonic air will simply destroy the turbomachinery,  even if it is not turning!  This diversion geometry is hard enough to do subsonically for a ramjet component,  and pretty-much technologically impossible to do supersonically for a scramjet component,  due to the shock-down risk.   And,  any such variable geometry inlet hardware is going to be voluminous and heavy,  as already stated.

Rocket-based combined cycles avoid the turbomachinery gas temperature problems.  These are essentially variations on the old ejector ramjet,  and conceptually could transition to scramjet,  if the ramjet uses a thermal choke instead of a physical convergence to a minimum throat area.  However,  no ramjet vehicle ever flew with a thermal choke,  instead always with a physical nozzle!  More than half a century ago,  tests clearly showed that thermal chokes resulted in too low a combustor pressure to ever get any effective performance out of the ramjet.

The true state-of-the-art for these combined cycle approaches is only concept design with finite-element computer analyses.  Not much real testing has been done,  and those results were always less than expected.  The design analyses were (and are) usually made with computational fluid dynamics (CFD),  which is still notoriously subject to both the garbage-in/garbage-out (GIGO) law,  and serious problems recognizing fully-converged numerical solutions. 

Here’s the real problem with CFD models:  there’s a lot more going on inside any engine than just compressible fluid flow with this-or-that turbulence model.  The physics of combustion are usually inadequately modeled,  and the physics of flameholding are usually NOT modeled at all,  in most CFD codes.  Yet these effects really dominate the physics in the engine!  The “gold standard” is thus still real test data with real hardware,  and there is actually precious little of that with most of these concepts. 

That test data objection applies to both the turbine-based and the rocket-based combined-cycle concepts.  These are thus nowhere-near ready-to-fly,  generally speaking.   Which in turn is why you cannot go to a propulsion company,  and just buy one off-the-shelf!  These notions get proposed a lot for government R&D funded efforts,  but none have ever completed any actual development programs.

Effective Propulsion Solutions For Hypersonic Flight

The real solution to these propulsive geometry dilemmas probably has more to do with “parallel burn” of separate propulsion devices,  than with any sort of combined-cycle engine approach.  Another name for “parallel burn” is “mixed propulsion”,  which craft such as the Douglas “Skyrocket”,  the NF-104,  and the XF-91 had.  This took the form of a rocket engine and a gas turbine engine,  on-board separately.

One possible example could be a ramjet vehicle with a built-in rocket booster,  but one able to burn both engines simultaneously after launch and at very high altitudes,  and at landing.  As separate propulsion devices,  the geometry and performance of each component can be optimized.  Forced to share otherwise-incompatible geometries,  both components will inherently end up far from optimal. 

Hypersonic Airbreathing Propulsion for Orbital Ascent?

The real problem is that thrust of an airbreather (any airbreather!),  and the vehicle lift and drag,  are roughly proportional to ambient atmospheric pressure,  while vehicle weight is not!  At very high altitudes in the thin air,  there is not enough lift to oppose the normal weight component,  and not enough thrust-minus-drag net force available to overcome the axial weight component,  and so thus the vehicle to fails to fly steady-state,  much less climb and/or accelerate.  This is shown in Figure 3. 

This effect is the source of the “service ceiling” for an aircraft powered by an airbreathing engine (any type of airbreathing engine!).  This is usually specified to be the altitude at which the rate of climb (R/C) falls under about 200 feet per minute,  which is 3.33 ft/sec vertical velocity.  At Mach 5 (about 5000 ft/sec velocity),  that would be a climbing path angle near 0.03 degrees.  At only 500 ft/sec flight velocity (V),  that would be a climb path angle nearer 0.30 degrees.  Neither is very discernible above horizontal. And THAT is the point here:  effectively that is no rate of climb capability,  no matter how long you try.

Lower down,  where the air is thicker,  these forces become far more favorable,  but the aeroheating and drag problems to overcome are far worse.  This is because the heat transfer coefficients are roughly proportional to atmospheric density raised to a fractional power near 0.8,  not 1.  That assessment comes from the usual formulation of the Nusselt number correlations from Reynolds number:  constant x Reynolds number-to-the-0.8 x Prandtl number-to-the-1/3,  for turbulent flow.  See again Figure 3.   

That aeroheating effect and the drag are what drive the need to be at really high altitudes as the vehicle speed approaches orbital values.  Yet for thrust-relative-to-weight purposes,  the airbreather needs to be at very much lower altitudes!  That fundamental design requirements incompatibility is quite stark! 

You overcome it by using rocket propulsion,  not airbreathing propulsion,  at those very high altitudes in the really thin air.  Or at least use rocket propulsion simultaneously with your airbreather.  Yes,  the airbreather makes its thrust at high specific impulse,  but it just does not make very much thrust in that thin air!  The rocket does.  Which is why one should prefer the rocket,  when leaving the atmosphere!

For a two-stage launch system where the first stage is an airplane of some kind,  there are three important variables to consider at your intended staging condition.  In order of importance,  they are (1) highest possible speed,  (2) path angle at about 45 (or more) degrees above horizontal,  and (3) highest possible altitude.  Speed has the greatest beneficial effect,  altitude the least. 

Path angle is important so that the second stage may fly a non-lifting gravity-turn trajectory at minimum drag loss,  to orbit.  Pulling up at high speed is a large-radius turn requiring a lot of gees and incurring a lot of drag loss due to a very high lift requirement.   If your first stage can do its trajectory to arrive at both high speed and path angle,  that minimizes the second stage drag loss and impulse requirement! 

But,  at all but the very lowest altitudes,  this will require rocket propulsion as the entire propulsion system,  or at least in combination with the airbreather in parallel burn.   No airbreather of any kind used alone will be able to do this kind of beneficial first stage trajectory,  precisely because of the thin-air “service ceiling” effect at higher altitudes.

How Big A Threat Is This New Hypersonic Weapon Stuff,  Really?

So,  there are very good reasons why the new “hypersonic weapons” currently being ballyhooed in the press are just rocket-powered tactical missiles with a peak speed above Mach 5.  Otherwise,  they are just hypersonic gliders dispensed from a large ballistic missile flown on a low,  rather flat,  trajectory.  The “scramjet missiles” are still experimental items,  not really fieldable weapons,  for a while yet. 

The old,  retired AIM-54 “Phoenix” rocket-powered missile had a peak speed of just about Mach 5,  way back in the 1970’s.  So,  what is so “new” or threatening about reprising that?  Nothing!

We have had maneuverable re-entry vehicles as space capsules since the 1960’s,  and as ballistic missile warheads since the 1980’s.  The space shuttle was another.  There is nothing “new” there! 

For the “hypersonic gliders” to be much of a military threat,  the launcher has to fly a much lower trajectory,  at a very shallow angle below horizontal,  very unlike the usual strategic ballistic missile.  Otherwise, there is no time to maneuver the glider before it impacts.  That’s just high school physics!

This “new hypersonic glider threat” is really no big deal,  if you know to watch for those depressed ballistic missile trajectories.  And we do.  I just told you,  if no one else did.


In addition to the six references cited above,  there is a seventh very useful item,  for those who wish to research this topic further and deeper.  It is included in the list here as Reference 7.  That one contains lists of articles sorted by the topic area. 

One of those topic areas is “aerodynamics and heat transfer articles”,  where I put the high-speed aerothermodynamics stuff,  among some other things.  The hypersonics-related stuff is there,  right up to entry heating models.  There’s also topic areas for “ramjet” and “rocket” stuff,  and much more.  All of these are articles that I wrote and published on this “exrocketman” site over the last several years. 

To find any such article quickly,  use the navigation tool on the left side of the page.  You will need the posting date and the title (jot them down).  Click on the year,  then month,  then the title (if need be). 

You can click on any figure in an article to see enlargements of all of the figures in the article.  There is an X-out option at top right of that page,  which takes you right back to the article itself. 

#1. 2 January 2020,  “High Speed Aerodynamics and Heat Transfer”  (physics and calculation models)

#2. 12 June 2017,  “Shock Impingement Heating Is Very Dangerous” (physics with X-15 as an example)

#3. 10 December 2016,  “Primer on Ramjets” (basic concepts and fundamentals)

#4. 21 December 2012,  “Ramjet Cycle Analyses”  (how these things are best calculated)

#5. 12 November 2018,  “How Propulsion Nozzles Work” (covers conventional and free-expansion)

#6. 9 November 2020,  “Fundamentals of Inlets” (same components used quite differently for ramjets and gas turbines)

#7. 21 October 2021,  “Lists of Some Articles By Topic Area” (dates and titles arranged by topic)

Figure 1 – Heat Transfer Issues With Hypersonic Flight

Figure 2 – Geometric Incompatibilities Among Airbreathing Concepts For Hypersonic Flight


Figure 3 – Thin-Air Effects On Thrust,  Lift,  Drag,  and Aeroheating at High Altitudes

Addendum 6-11-22:

Here is a plot about the flight test space covered by the X-15 program,  relative to the aerodynamic ascent path to low Earth orbit.  The lower altitudes correspond to higher speeds reached by the X-15.  The higher altitudes correspond to lower speeds achieved.  

The X-15A-2 variant was able to reach 4520 mph (Mach 6.7) at 19.3 miles altitude,  still near the left end of that ascent corridor.  It was carrying a scramjet test article on its ventral fin stub during this flight.  The shock wave off the scramjet inlet compression spike nearly cut the tail off the bird from shock impingement heating effects. 

Now bear in mind that shock impingement effects multiply (considerably) the heating rates,  but not the plasma temperatures themselves.  All that means is that the structures affected are going to soak out very quickly,  to very near the plasma driving temperature,  pretty much regardless of what the designer might have done in the way of cooling provisions.  

At the X-15A-2 peak speed conditions,  that plasma sheath temperature was some 3070-3080 F,  far beyond the capability of even the Inconel-X skin material.  The white ceramic coating applied to this particular variant would have been able to do very little about this fast soak-out effect at the shockwave impingement locations.  So,  it is not surprising at all (in retrospect) that the craft suffered so much damage.  

What this figure says so very clearly is that if you fly too high,  you will not get enough lift to fly aerodynamically in such thin air.  And if you fly too low,  you will suffer unendurable aeroheating,  no matter how you attempt to construct your craft.  

And the X-15A-2 experience cited here says that you must avoid any shapes that might cause a shockwave from one structure to impinge upon an adjacent structure.  That absolutely rules out parallel-mounted nacelles.  

It is even worse near the orbital entry point end of the aerodynamic ascent corridor.  That would be about 17,000 mph at about 70 miles altitude.  The estimated plasma sheath driving temperature there is just about 13,200 F.  The peak heating point lies in between these two point (the X-15A-2 test and the orbital entry point).  

Your nosetip and leading edges will have to be ablatives.  And there is no allowing any significant soak-out for lateral skins.  Not at plasma temperatures in the 3000-13,000 F range!

By way of comparison,  the vertically-launched rocket vehicles we currently use to reach orbit (and that includes the Space Shuttle),  leave the sensible atmosphere at about 30 miles,  but only about 1500-2000 mph at that point.  They stay well above the aerodynamic ascent corridor,  essentially flying most of the way in vacuum,  without severe heating.

It is re-entry from orbit that essentially flies this aerodynamic ascent corridor in reverse.  It's just that the exposure times are a lot shorter for entry than they are for ascent.  So the overall heating exposures for aerodynamic ascent are much worse than for entry.

Those are the fundamental things you have to consider,  if you wish to fly an aerodynamic ascent to orbit,  regardless of the propulsion you may be considering. 

Update 6-12-22:

A lot of folks are using the word “hypersonic” very loosely.  Too loosely.  I have even seen the SR-71 termed “hypersonic”,  when it most definitely was not,  at a max allowable cruise speed of Mach 3.2 at some 85,000 feet.  Pilots would pull up and reduce throttle to slow down quickly,  if it ever reached Mach 3.3,  to avoid engine and airframe damage.  I have talked to a couple of them,  and that is what they told me.

There is a formal definition:  “hypersonic” is that speed at which the shape of the shock wave system about the vehicle no longer changes its shape appreciably as flight speed increases further.  The speed at which that happens depends upon the shape of the vehicle. 

For “pointy” shapes likes missiles and aircraft,  it is just about Mach 5,  or a velocity that is 5 times the speed of sound.  For really blunt shapes like space capsules traveling heat shield forward,  it is just about Mach 3.  Those two values are the rules of thumb for “hypersonic”,  for those two classes of shape.

Here is a further update to the figure showing the ascent corridor and the X-15 data.  I have added a typical vertical-launch trajectory,  and some effective temperatures in the plasma sheath about any vehicle.  I also show how precipitously the heat transfer coefficient drops as altitude extremizes.  That is why the peak heating rate point is somewhere in the middle of the ascent corridor,  not its high-speed end.