Tuesday, November 2, 2021

The “Warm Brick” Ramjet Device

One of the most unusual ramjet projects I ever worked on was a non-propulsive device.   This was a very miniature ram-fed airbreathing combustor,  that was to be the hot gas generator for an infrared (IR) decoy.  This decoy was to be towed behind an aircraft in lieu of a whole series of dispensed flares.  It was intended to work by having enough IR output to cause the aircraft to drop out of the missile field-of-view first.  See Figure 1 for that concept.  I was working for my friend Byron Hinderer doing this.

Figure 1 – Towed IR Decoy Concept,  called “Warm Brick” at Tracor in 1984

I did this at what was then Tracor Aerospace,  in Austin,  Texas,  during 1984.  We called this decoy “Warm Brick”,  and my job was to determine if this concept was even feasible (it was).  Our idea was to heat a porous refractory material until it glowed brightly in the IR.  We preferred fuel-air combustion to minimize decoy mass,  and ram combustion is the simplest of the airbreathers.  Plus,  I had lots of experience with ramjet combustion at what was then Rocketdyne/Hercules in McGregor,  Texas.

To the very best of my knowledge,  no patent was ever taken out on this concept,  and Tracor never did anything at all with it.  Even if there had been a patent,  and it had been renewed,  any such patent would have run out by now.   So,  what I reveal here should offend no one,  and infringe no patents.

As implemented for the feasibility tests,  this concept took the form of a “gasoline lantern mantle” made out of commercial ceramic fire curtain cloth,  as the IR emitter.  This was to be mounted behind a wake-producing spoiler,  mounted at the aft end of the burner and inlet assembly.  The decoy might carry its own fuel tank,  or it receive fuel down its tow line,  if a heavier tow line could be tolerated. 

To test the scientific and engineering feasibility,  I designed a very generalized inlet and burner hardware set that was flexible enough to allow evaluation with a variety of gaseous and liquid fuels.  See Figures 2 and 3.  The intended flight conditions were relatively low altitude from mild subsonic to barely-supersonic speeds,  typical of an attack aircraft threatened by surface-to-air missiles.  

Figure 2 – Assembly Sketch for the Initial Version “Warm Brick” Ram Combustor Test Device

The assembly sketch clearly depicts the long fuel injection-and-mixing duct allowed between the inlet diffuser and the sudden dump into the combustor.  There was an inlet piece and a fuel injector piece,  both made of aluminum for ease of rework,  and an inlet tube and a combustor shell,  both made of steel.  The combustor shell was sized for fabrication from 2-inch schedule-40 pipe,  but ended up being made of 300-series stainless to those same dimensions.  We tried automotive-style spark ignition.

One can easily see how the molded low-density ceramic liner insert was to be trapped in place by the nozzle block.  The arrangement shown in the assembly sketch of Figure 2 (directly-pinned nozzle block) was quickly replaced by a pinned steel nozzle shell ring,  as shown in the hardware photo (Figure 3).  This revision happened about the same time that the first (unreinforced) liner was replaced with the second liner (reinforced ceramic composite).  

Figure 3 – Photo of the “Warm Brick” Ram Combustor Test Hardware as Revised

The design concept called for a small combustor fed by a simple pitot inlet,  with a convergent-only nozzle that would likely function unchoked at most conditions.  I chose a center-duct coaxial air entry with sudden-dump flame stabilization,  similar to the successfully-flown ASALM-PTV liquid-fueled ramjet test vehicle.   Geometric ratios were initially set equal to those used in ASALM.

Based on Reference 1,  I chose a minimum ¼-inch (6 mm) step height around the dump.  The combustor length was sized “empirically” (rules of thumb based on ASALM-PTV geometry) so that the annular separation bubbles would close,  and the axial core would be “burned out”,  before any of these flows entered the nozzle.  That was basically an assumed 11-degree spreading angle,  on both sides of the mixing layer between the entering mixture and the recirculated flame.   That’s too crude,  in hindsight.

We wanted sufficient porosity in the emitter so that the burner operation would be unaffected by the presence or absence of the emitter.  The fire curtain cloth gave us that,  in the sizes tested,  because the surface area of the ellipsoidal shape was so large relative to the final burner throat area.  Its effective porosity-driven “free” open area was very much larger than any of the burner throat areas we tested.   

There were two crucial unanswered questions:  (1) emitter/hot gas coupling (could we really get the emitter hot enough to radiate effectively?),  and (2) obtaining stable combustion at all in a burner that small,  with any fuel whatsoever!  There was an extensive paper trade study done,  to determine the desired fuels.  In test,  these fuels,  and some other fuels that were easier to use,  were investigated. 

This combustor was nominally 1.5 inch (38 mm) inside diameter,  as insulated,  and 3 inches (76 mm) long inside.  The smallest size ramjet combustor in my experience up to that point had been some heavyweight solid-propellant ducted rocket ramjet work (in a completely-different geometry) at 4.6 inch (117 mm) inside diameter,  and length/diameter 6-to-8.  The largest was ASALM-PTV at a 20 inch (51 cm) combustor case diameter.  “Warm Brick” was smaller than anything of which I had any knowledge

I didn’t want to periodically replace an ablative liner in the test burner,  and I didn’t want to attempt an air-cooled liner shell for full-rich combustion in something that small.  So I opted for an unknown,  inspired by the Space Shuttle’s heat shield tiles.  Could I put a low-density ceramic insulator in this combustor,  and not melt it?  The answer turned out to be “yes”,  but it took some adaptive effort.

The project operated in three logical parts:  (1) obtain stable combustion with a variety of fuels in the burner alone,  (2) add the emitter and determine how best to shape,  fabricate,  and attach it,  and (3) document infrared radiometric output.  The real prerequisite for part (1) was the combustor insulator,  since we started with gaseous fuels,  thereby avoiding the fuel vaporization issue. 

I selected free-jet test mode as the best way to accomplish all three parts of this project with the same hardware and test setup (see Figure 4).   All that I had personally done while at Rocketdyne/Hercules was direct-connect testing,  but I knew about free-jet testing,  both from my research,   and some experimental association with Marquardt,  while I was with Rocketdyne/Hercules. 

We used a commercially-rented air compressor trailer as our air source,  to be run real-time.  In 1984,  this 750 SCFM unit was the largest of its kind in Texas.  It fed a PVC pipe stilling chamber,  terminating in a simple convergent-only nozzle block made (conveniently) of wood.  

Figure 4 – Test Setup:  Stilling Chamber Exhausting To Left,  Fed From Right

The test article was bolted to a heavy pipe stand-and-sting,  with its inlet immersed in the free jet of air.  That free jet typically measured 190 F (88 C) stagnation temperature,  at full-power compression conditions. 

The first part of the investigation began with bottled hydrogen gas fuel (series 1).  This and all the other trials are summarized in Table 1 below.  Series 1 wasn’t very successful for two reasons:  (1) the nozzle was too wide open for a stable flame,  and (2) free jet air speeds higher than about 0.25 Mach blew the spark column out from the electrodes of the spark plug,  even though it was located flush within the annular recirculation zone.  

The device didn’t ignite at all until I obstructed the nozzle with a scrap of wood,  and it still went out after ignition,  if I removed the obstruction.  So,  I built a smaller-throat nozzle block.  We still had to ignite at low airspeed and gradually work up to higher speeds,  limited at that time to about half a Mach number by the stilling chamber nozzle.  I also tried liquid ethanol unsuccessfully at this time (series 2). 

Somewhere in all of this,  I first drove the combustor into what proved to be a very violent rich blowout instability,  and completely shattered my first (unreinforced) liner!  The combustor visibly shook on its sting,  and it spit the pieces of its liner out the nozzle,  igniting a local grass fire!  Later,  we estimated a pressure amplitude near 0.8 atm,  at audio frequencies (a few hundred Hertz),  for this instability. 

A photo of the liner molding tools that I used is given in Figure 5,  which includes the basic combustor shell as the outer forming tool for the combustor liner.  Both it and the nozzle block were laid up as (commercial) low-density molding compound troweled onto the wooden plug,  and inserted into the corresponding shell for cure.  I used Cotronics Corp. 360M low-density molding compound for this.  

Figure 5 – Tools Used for Molding Ceramic Combustor Insulation Liner Inserts

These parts were cured at 215 F (102 C) in an oven to drive off the water.  The circuitous paths for exiting steam led to a low density ceramic matrix.  The resulting parts were coated with a paint-like ceramic cement (Cotronics Corp. 901),  and cured again,  in the same oven.  The unreinforced liner showed evidence of hot gas flow behind the insulation,  and into the cracks,  shown in Figures 6 and 7.  

Figure 6 – Recovered Pieces of Shattered Unreinforced Liner,  Bonded Together

Figure 7 – View of Fracture Surface,  Showing Hot Gas Flow Damage with Sooting

I built a second ceramic composite liner reinforced by layers of the fire curtain cloth (woven from 3M Nextel 312),  which survived all instabilities and any other test abuses thereafter.  It survived many hours of accumulated burn time in near-pristine condition,  as seen in Figures 8 and 9.  The shrinkage cracks did not preclude functionality.  There was some melting evident in the throat of the nozzle.  

Figure 8 – View Into Near-Pristine Reinforced Liner,  After Hours of Burn Time

Figure 9 – View Into Reinforced Nozzle Block,  After Hours of Burn Time

Once we had the burner working at all,  we tried some test sample pyrometers in its exhaust plume,  with both propane and acetone as fuel (series 3,  and acetone proved worthless as a fuel).  These pyrometers would be old nails,  or else planar samples of potential emitter materials.  We even tried gasoline as fuel (series 4),  but results were poor,  and it became very obvious that poor vaporization was the cause!  I tried propane again (series 5) as the most successful fuel,  and got enough radiometer output to be encouraging,  from a sample of the fire curtain cloth immersed in the jet exhaust. 

So,  I created a fuel-line hot-soak bucket to correct the poor fuel vaporization problem for test purposes.  This took the form of an electrically-heated bucket of old motor oil,  in which a coil of the fuel supply line was immersed.  That rig is shown in Figure 10.  It may resemble a moonshine still,  but it is not!  

Figure 10 – Fuel Vaporization Preheat Bucket Rig

At this point,  I had a crudely-successful burner,  but an unproven fuel supply method.  I checked out the combined burner and fuel vaporization bucket,  first on propane (series 6),  then on aviation gasoline (series 7),  and finally on a “home-made version of JP-4”  that was actually half Jet-A and half aviation gasoline (series 8).  Plus,  I added instrumentation to the burner unit (enough manometer pressures and thermocouples to attempt an actual “engine” cycle analysis). 

Results,  including the exhaust pyrometer samples,  were favorable enough to warrant continuing the project further.  It still required a lower-airspeed ignition.  I stood in the jet blast for all these tests,  looking directly into the flame zone,  and sniffing for unburned fuel,  to set mixture.  That “settled” the fuel injection and ignition issues well enough to test emitter coupling issues for the very first time

The first actual emitter was made of Nextel 312 fire curtain cloth,  coated with the Cotronics 901 adhesive as a “paint”.  It was sewn together,  with alumino-silicate thread,  from bias-cut gores much like a balloon,  to form an elongated semi-ellipse approximation.  The seams were left on the outside of this first emitter,  as shown in Figure 11.  It was the first of several series 9 tests with pre-heated propane,  at air speeds up to about Mach 0.47.  Those test conditions are depicted in Figure 12.

Figure 11 – Test Setup for First-Article Emitter

For all subsequent tests,  the seams in the sewn emitters were placed to the inside,  as is depicted in Figure 13.  That photo shows post-burn appearance of two series 12 emitters tested with ethanol fuel,  but all the internal-seam emitters appeared similar,  regardless of series and fuel. 

These articles were brittle and fragile post-test,  as expected for alumino-silicate materials soaked to temperatures exceeding the solid phase-change temperature of about 2350 F (1290 C).  That fragility alone confirmed a high surface temperature for radiation purposes!  This was also verified by radiometric measurement,  which also indicated very “non-gray” behavior,  in that the effective color temperature (radiation peak wavelength) was substantially cooler than the actual temperature. 

The spoiler just ahead of the emitter clamp mounting provided protection from direct wind blast forces.  Plus,  it also provided effective hot gas recirculation effects external to the emitter surface.  Both acted to raise emitter material soak temperature,  and therefore IR output,  quite successfully.    

Figure 12 – Test Conditions Explored with Series 9 Propane

Two tests were made as series 10 in this same configuration with the “home-made JP-4” fuel.  Results were similar to the series 9 propane runs,  except for a small liquid-wet “cold spot” at the very end of the emitter bulb.  This was due to still-unvaporized kerosene hitting the emitter on-axis.  

Figure 13 – Post-Test Emitter Appearance from Series 12 Ethanol Tests

Sometime during this checkout process before the series 9 propane runs,  I successfully modified the inlet to a larger lip radius, in order to decrease its “buzz” instability tendencies at higher backpressures.  That also greatly improved ignition characteristics,  and it further pushed the rich blow-out instability limits to richer mixtures!  The test set-up for cold-flow inlet calibration is shown in Figure 14. 

Both the original and modified (larger lip radius) inlets were cold-flow tested with this rig.  Data were cross-plotted in a variety of ways.  The data plot format for “typical” supersonic ramjets was rather undiscriminating at these subsonic speeds:  stream tube area ratio versus Mach and stagnation pressure recovery ratio versus Mach.  Plots in the more primitive-variable format were actually more useful for this mostly-subsonic system.  These included the diffused Mach to freestream Mach ratio,  and the static “pressure gain” ratio. 

These results guided the 1984-vintage data reductions of the series 9 propane runs with emitters.  From those,  installed hot-burn test inlet performance data matched the cold-flow tests.  The streamtube area recovery ratio shows a very strong influence of the so-called “highlight” area versus the true minimum area,  when used as the reference area for the calculation. 

After the fact,  this was entirely expected,  based on Reference 2,  which (of course) recommends the highlight definition.  At the time I did these tests,  I had used something pretty close to the minimum area for the reference.   It shows explicitly in the data,  as a recovery ratio substantially greater than unity,  which is completely out-of-line with the usual expectations for ramjet inlets. 

See Figures 15 and 16.  

Figure 14 – Cold-Flow Inlet Calibration Test Rig

After the series 9 and 10 tests,  the air nozzle in the stilling chamber was replaced with a second wooden unit of slightly smaller throat diameter,  as depicted in Figure 17.  This enabled free jets of nearly Mach 1 speed at the maximum compressor output.  Two more test series were conducted with this change,  specifically to obtain data at those higher simulated air speeds.  These were series 11,  using both propane and hydrogen fuels,  and series 12,  which used the finally-selected ethanol fuel. 

The series 12 tests employed both radiometer measurements,  and imaging with a thermal imager camera.  The fuel vaporizer rig was less successful with a high latent heat pure-substance fuel (ethanol),  than it had been with distillate fuels,  or with the easily-vaporized propane.  With ethanol,  it was essentially long-period unstable,  with an oscillating fuel flow output.  The cycling time was a few seconds. 

Nevertheless,  using ethanol fuel produced an output spectral power distribution closer to what is needed from the non-gray decoy.  The radiometer data clearly showed this.  We attributed this difference (with a high degree of confidence) to the lack of yellow carbon glare in the ethanol flame.  This yellow carbon glare was quite noticeable in the propane tests,  and even more so when using gasoline or jet fuel.  The series 12 ethanol runs looked to the eye “positively white” in comparison.

The ethanol fuel injector was stopwatch-and-bucket calibrated for those series 12 tests. Those calibration data are shown in Figure 18.

Figure 15 – Calibrated Inlet Performance Derived from Series 9 Data,  Part 1

Figure 16 -- Calibrated Inlet Performance Derived from Series 9 Data,  Part 2

Figure 17 – Air Nozzle Re-Work for Higher-Airspeed Test Capability

Figure 18 – Flow Calibration Data for the Series 12 Ethanol Fuel Runs

After these tests,  the fuel vaporization problem was conceptually addressed as a hot-gas tap from the forward end of the combustor to the lower-pressure zone at the minimum area of the inlet.  Fuel would be injected into this very hot recirculated gas stream to effect rapid vaporization.  While the design analysis looked good,  that concept never received any testing due to budgetary constraints that essentially stopped all experimental work on the project after late 1984.  Some prototype flyable hardware was designed,  and a few of those parts manufactured,  before all work on the project was completely stopped.  It never resumed.  So NOTHING is confirmed about any of this!

The ceramic liner material was never characterized,  it “just worked”.  Density,  strength,  and thermal conductivity were never measured in any way!  However,  it handled as if it were about as dense as commercial Styrofoam products.  The strength was considerable,  considering the rich blow-out instability abuse it endured.  Immersed in a 190 F (88 C) air stream,  the combustor shell would “barely boil spit” after an hour-long burn test at full rich mixture (theoretically around 3800 F or 2100 C),  with but 0.2 inch (5 mm) thickness of the insulation!  That indicated very low thermal conductivity indeed! 

Table 1 – Summary of “Warm Brick” Burner Tests

In recent years,  I developed further those basic cycle analysis techniques applicable to a low-speed ramjet system,  or a subsonic nonpropulsive item like “Warm Brick”.  In particular,  I programmed them into an “Excel” spreadsheet,  and reanalyzed the “typical” series 9 propane run at 0.47 Mach air speed and full-rich mixture.  The spreadsheet setup is shown in Figure 19,  and the spreadsheet results in Figure 20.  Since then,  I have created a real low-speed ramjet cycle analysis code.  It works just fine. 

These recent compressible-flow cycle analysis results defined the bulk flow conditions inside the combustor well enough to attempt a heat transfer model with a reasonable expectation of success.  That model was cylindrical convective-conductive,  and based on standard compressible flow models inside and outside the combustor shell.  Radiative loss was near zero,  as there was no effective path by which thermal radiation could leave the interior.  The shell radiation cooling potential was very low. 

While the steel shell has a well-known thermal conductivity,  the ceramic composite liner did not,  so I ran this model parametrically versus conductivity values from “very low” to “very high”.  The “best” value of thermal conductivity was that which matched both my recollections of perceived shell temperature,  and my observation that the liner surface was often close to melting (3250 F, 1790 C). 

Those thermal conductivity results are given in Figures 21 and 22.  The highlighted value of 0.02 BTU/hr-ft-F equates to 0.035 W/m-C.  Density and strength still lack actual characterization!  I have often wondered whether this material might serve as a re-entry heat shield material,  the way that the somewhat-similar low-density ceramic Shuttle tile did.  But that is another topic for another venue. 


#1. Curran,  Edward T.,  “An Investigation Of Flame Stability In A Coaxial Dump Combustor” (dissertation,  AFAPL/RJ WPAFB,  Dayton,  OH),  AFIT/AE/DS 79-1,  Jan. 1979. 

#2. Seddon, J.,  and Goldsmith,  E. L.,  “Intake Aerodynamics”,  AIAA Education Series,  1985,  ISBN 0-930403-03-7.

Figure 19 – Spreadsheet Setup for “Warm Brick” Cycle Analysis at Series 9 Propane Conditions

Figure 20 – Spreadsheet Cycle Analysis Results for “Warm Brick” at Series 9 Propane Conditions

Figure 21 – Heat Transfer Model Results for “Warm Brick” Liner Thermal Conductivity

Figure 22 – Heat Transfer Model Results Plotted vs Radius

Epilogue:  Some Practical Combustion Device Lessons Learned

Cycle analysis with one-dimensional flow models turned out to be less important than the actual scale-dependent physical chemistry of flame stability,  for this “Warm Brick”  device.  Residence time is proportional to dimension,  all else equal,  while chemical reaction rates are scale-independent.  This alone suggests there is a minimum size below which a thing “just won’t work” with a particular fuel. 

Mixing is another very strong determinant of flame stability.  Mixing is not proportional to scale,  nor is it scale-independent,  but it is something in-between.  Again,  this also suggests that there is a size below which a thing “just won’t work” with any particular fuel.  That is precisely one issue (of many) in flameholding.

Those considerations explain why the required nozzle contraction ratio (and internal flow velocities) were so low in the “Warm Brick” device for stable ignition and burning,  relative to everything I knew about,  before I attempted this project.  However,  these experiences with the Warm Brick subminiature combustor predate the in-depth understanding of flameholding and flame stability that I was later able to achieve,  after returning to Rocketdyne/Hercules.  That knowledge is summarized in the “exrocketman” article titled “Ramjet Flameholding” (on this site) and dated 3 March 2020.

The vaporization of fuels of different latent heats and boiling behavior revealed a surge instability in the hot-bucket fuel rig (referring again to the crude hardware in Fig. 10 above).  The basic layout was a source of fuel at pressure,  led through a copper line coiled in the hot bucket,  and from there to the metering orifices inside the test article.  See the cartoon in Fig. 23. 

The source of fuel-at-pressure was a standard 5-gallon propane bottle (usually around 200-250 psig),  or a welding gas bottle (initially 2200 psig),  or a pressure tank of liquid fuel pressurized with compressed dry nitrogen (usually pressurized in the 100-300 psig range).  All of these pressurization schemes are regulator-controlled.  That regulator was physically located about 5-to-10 feet downstream of the test article,  and within arm’s reach of the exhaust plume.  This allowed me to manually adjust the fuel flow during the test by varying the regulated pressures,  while standing immersed in the exhaust where I could smell for unburned fuel.  For the open-nozzle tests,  I could literally see the flame up the tailpipe.    

Fig. 23 – Conceptual Layout and Operation of Fuel Supply

When using hydrogen directly from the welding gas bottle,  there was no vaporization problem,  as this was simply compressed hydrogen gas.  We did not use a pre-heater bucket with this fuel,  but the rest of the component layout in Fig. 22 is correct. 

With propane in the 0.47 Mach air tests,  we found the line just downstream of the regulator,  and the sides of the propane bottle,  to be cold.  This is because the vaporizing pool of liquid propane in the bottle must draw about 150 BTU/lbm of latent heat from itself and from its surroundings,  mostly from itself (gets cold).  If it cannot draw sufficient heat to vaporize,  then it won’t vaporize,  pressure drop notwithstanding!  The energy to change phase (latent heat) simply must come from somewhere!

There was a cold-line risk of re-condensation on the way to the test article,  which we “cured” with the hot oil bucket preheater.  We kept the line length from bottle to preheater as short as practical.  We also found bottle “freeze-up” occurred at the higher flow rates with the Mach 0.9 airstream tests.  We “cured” that by the camper’s expedient of putting the propane bottle in a tub of hot water. 

With gasoline and jet fuel,  the driving pressures helped us pre-heat the liquid fuel without getting any boiling in the fuel line.  Without preheat,  there was insufficient air stream heat in the test article to get the fuel to vaporize and burn.  With about 300 F preheat,  we got all but the “tag-end” of the distillation curve to vaporize upon being injected,  due to combined atomization and pressure-drop boiling. 

With the gasoline and 300 F preheat,  our nominal 100-300 psig driving pressure was apparently barely enough to prevent any significant boiling in the line,  so we did not encounter any noticeable problems with vapor lock-induced fuel flow rate surges.  With the jet fuel and its lower volatility,  we had no real risk of vapor lock surging,  but we did see a little more “tag end” unvaporized fuel,  indicating a higher preheat temperature was really needed.  Both of these are about 150 BTU/lbm latent heat materials.

We did have a real fuel surge problem running neat ethanol as fuel.  This material has a far higher latent heat at about 378 BTU/lbm,  and it has a single normal boiling point,  instead of distillation behavior.  At our delivery conditions,  the pressure was insufficient to prevent boiling in the line,  leading directly to vapor lock-induced flow rate surging!  Fuel delivery rates oscillated through about a factor of two,  on a long period of several seconds.  It would vapor lock,  unlock,  and relock to cause this surging.

We could not reduce preheat temperatures and still expect to get any flash vaporization upon injection,  in hindsight due to that higher latent heat.   We could not increase the feed pressures to preclude the boiling without re-working the test article for much smaller injection orifices.  That latter is the real design solution to this problem,  but we did not use it for these tests!  We were able to get our infrared radiometer data from the high points of the oscillating-intensity burn. 

While high pressure preheat to get flash vaporization from an atomizing injector is an approach that really works,  the equipment to do it is usually large and heavy,  too much so for a miniature decoy.  The alternative would be to mix the fuel with hot combustion gas to get vaporization,  downstream of the metering point.  The design difficulty is then to get good mixing of the fuel-rich gas stream with the inlet airstream,  without suffering large pressure losses.  That seemed the better approach for the flight decoy design.  We were never able to test this,  though!  It is still just a concept!

For an aero-engine application,  high-pressure fuel pre-heat with atomizing flash vaporization is likely the better design approach.  The sizing of required preheat depends upon raising the liquid to a temperature such that the enthalpy drop across the injection orifice exceeds the latent heat of vaporization.  The size of the orifice and the feed line pressure determine flow rate.  But,  the feed line pressures must always exceed fuel vapor pressure at that high preheat temperature!  If this is not done,  then vapor lock-induced surging will occur,  and at very significant magnitudes.  Fuel control then becomes impossible. 

As indicated,  we never got to test the concept of vaporization by injection into a hot combustion gas stream,  followed by injecting that hot mixed stream into inlet air.  There is a lot of promise in that notion,  but it is fraught with practical difficulties,  as well.

Final Comments:  IR Emission Characteristics and Towed Decoy Physics

The IR emission characteristics topic has been mostly ignored here,  except to say these ceramics were decidedly “non-gray” in their spectrally-dependent emissivity properties.  They were non-gray enough to reduce expected radiation in the 1-2 micron band very markedly,  to near what they emitted in the 3-5 micron band,  despite operating at a temperature somewhere near 3000 F (1650 C).  The effective “color temperature” (really the wavelength at peak spectral distribution of radiation) was much closer to typical tailpipe temperatures at full power (but less than those with full afterburning).

Suffice it to say that a great deal of infrared power was radiated by a very small object,  whose color temperature and radiated-power in-band looked like a very large jet engine tailpipe at full power.  This little emitter would blister my face with radiated heat from some 20 feet away.  The large radiated power would be the temperature-to-fourth-power effect,  while the color temperature would be the non-gray emissivity effect.  Both are critical effects.

Exploring this IR emission topic in more detail would be the subject of some future article,  or perhaps even a book relating these experiences.  This is an application of otherwise well-established physics.

Another unaddressed topic is aeromechanical in nature:   how to tow hard-body decoys stably on towlines,  at speeds from very subsonic to low supersonic.  The answers are not what one would expect,  based on the towed gunnery targets that have been flown for some decades now.  Straight tow is the easiest to achieve at all speeds,  meaning the tow line extends mostly straight back from the aircraft,  although you DO NOT tow the body by its nose! Low or high tow are far,  far more difficult to achieve,  especially as speeds become high subsonic and the aero forces exceed the weight force.  Stable side tow is nearly impossible,  even at low subsonic.  This applies to radar decoys as well as IR decoys.

Exploring how to tow hard bodies behind aircraft might be the topic of a future article or articles,  or even part of a book.  The basic rules were invented by my friend Byron Hinderer.  I researched the details,  and documented what did not work,  as well as what did,  in the wind tunnel while at Tracor.

The final unaddressed topic deals with what is called “engagement analysis”,  where the geometry of the aircraft,  the tow,  the approach geometry of the attacking aircraft or missile,  and the characteristics of the decoy (IR or radar) and the seeker,  all interplay.  The desired result is an estimate of the kill probability for the attack.  The decoy designer wishes to reduce that kill probability to near zero.

Exploring engagement analysis with IR decoys and IR threats might be some future article.  Or it might also be part of a book on these experiences.   This topic I learned and practiced while at Tracor.


  1. I am very surprised that no-one else has ever come up with this!

    Thanks for sharing.

    1. This is stuff from some decades ago, but it is a topic area not often presented to the public. Towed radar decoys were under development at the same time, and I worked on those, too. Those are now "standard equipment" in recent years, but the towed IR decoy never achieved that status. -- GW