The classified Sabre engine was a descendent of the RB545 developed for the abandoned HOTOL space launcher. The Sabre was designed to deliver a high air-breathing thrust/weight ratio with moderate SFC while reverting to a high specific impulse rocket engine at transition. Since the air-breathing mode operated on a turbomachinery based cycle the engine was capable of generating static thrust (unlike ramjet cycles) and engine development could therefore take place on open test bed facilities. Optimum transition from air-breathing to rocket mode with this type of powerplant occurred at around Mach 5 and 26km after which the vehicle climbed steeply out of the atmosphere to minimise drag losses. The resulting ascent trajectory was relatively benign to both engine and airframe, leaving a reasonable choice of airframe materials capable of withstanding the ascent and reentry temperatures without active cooling. The Sabre engine was designed with state of the art technology for turbomachinery, pumps and combustion chambers etc. Current materials were specified for the engine machinery while the nacelle shell was manufactured in Sic reinforced glass and the bypass system in C/Sic.
By employing the rocket combustion chamber, nozzle and pumps in both modes the mass penalty of adding a separate air-breathing engine was reduced, while also eliminating the base drag penalty of the 'dead' rocket engine during air-breathing ascent. To generate reasonable thrust levels during air-breathing while preserving high nozzle area ratio the airflow had to be pumped up to typical rocket chamber pressures (i.e.: approx. 100-200bar). The aim of the thermodynamic cycle therefore, was to provide this high pressure airstream with minimum fuel flow, assisted by the remarkable properties of liquid hydrogen (temperature and specific heat). By treating the engine as a 'black box' it was possible to show that the minimum fuel/air ratio to achieve a 150bar air deliver y was 0.0433, at which the combined entropy rise of the two streams was zero.
However practical cycles could only achieve a fuel/air ratio approximately twice this value partly due to their thermodynamics and partly due to component inefficiencies and finite temperature differences. In order to reduce the cycle power requirement and to achieve reasonable air compressor outlet temperatures it was necessary to cool the incoming airflow, particularly at high Mach no's. However the airflow could also be viewed as a source of heat to drive a thermodynamic cycle operating between the high inlet air temperature and the low hydrogen stream 'sink' temperature. Then to allow the engine to operate over a range of speeds, the reduced incoming air enthalpy at low Mach no required 'topping up' by combustion heat release in a preburner to ensure constant turbine inlet temperature. The resulting engines ran at a nearly constant turbomachinery operating point and fuel flow over the whole air-breathing Mach no range.
The first attempt at designing an engine embodying the principles outlined above was the RB545 (HOTOL) engine. In this cycle the high pressure hydrogen flow was used to cool the airstream directly, following which the hydrogen stream split; approximately 1/3rd passing to the combustion chamber via the preburner while the remainder was expanded through the turbocompressor turbine prior to exhaust. This cycle variant had a higher fuel flow than the Sabre engine particularly at high Mach no's due to precooler metal temperature limitations caused by hydrogen embrittlement. In addition the precooler frost control system was relatively crude, resulting in significant payload penalty.
The Sabre engine traded simplicity for a lower fuel flow by interposing a Brayton cycle power loop in between the 'hot' airstream and the 'cold' hydrogen stream. The closed power loop required a working fluid that remained gaseous at temperatures as low as 30K, restricting the choice to hydrogen or helium. Helium was chosen since: helium's higher gamma reduced the loop pressure ratio; helium's higher molecular weight resulted in lighter turbines with fewer stages; and helium's inert nature permits a wider choice of precooler alloys. The air entered the cycle via the air intake where it was cooled to low temperature prior to entry into the air compressor. Following compression the airflow was divided, part flowing to t he main combustion chamber and the remainder passing to the fuel rich preburner. The hot gas from the preburner passed through HX3 to raise the helium outlet temperature from the precooler to a constant turbine entry temperature. The preburner gas then fl owed to the main combustion chamber to complete its combustion with the remaining air prior to expansion through the exhaust nozzle. From HX3 the helium was expanded through the main drive turbines to drive the air compressor, and then to HX4 where it was cooled by hydrogen delivered from the hydrogen pump. The cold helium then flowed to the circulator and then to the precooler to pass around the cycle once more. The warm hydrogen emerging from HX4 drove the hydrogen turbopump and then the helium circulator before passing into the preburner.
At transition the air inlet closed and the turbocompressor was rundown while simultaneously the liquid oxygen turbopump was runup. The preburner temperature was reduced in rocket mode reflecting the reduced power demand of the lox turbopump. The high pressure liquid oxygen was evaporated in the main chamber cooling jacket in order to allow the same oxidiser injectors to be used in both modes.
The majority of the Sabre engine components (e.g.: combustion chamber, nozzle, pumps, turbocompressor) were relatively conventional. Lightweight high power heat exchangers however were a new feature peculiar to this type of engine and posed a challenging manufacturing problem.
The precooler was an efficient counterflow design, consisting of many thousand small bore thin wall tubes. The cold helium flowed inside the tubes while the hot air was arranged in external crossflow. Due to the high pressure difference between the two fluid streams plain tubes gave the lightest matrix design. External fins and turbulence stimulators did not prove mass effective despite the low airside heat transfer coefficient.
The matrix was divided into two parts, the first high temperature matrix was manufactured in nickel alloy while the cooler second matrix was aluminium. The matrix dimensions resulted from an optimisation exercise whereby effectiveness and airside pressure drop were traded off against matrix mass. The resulting matrix had a large airside frontal area but was relatively shallow. Consequently the matrix was wrapped into a tapered cylinder in order to minimise the nacelle cross section. Proprietary frost control technology was used.
To maximise the core engine thrust in air-breathing mode the engine was operated at constant chamber pressure, which resulted in nearly constant core airflow. The intake capture area was determined by the requirement to supply this airflow at the maximum air-breathing Mach no, notionally M5 @26km. The compressor flow capacity was determined by specifying that the engine should achieve full chamber pressure at Mach 0.6 sea level. This specification yielded the required intake recovered pressure to achieve full chamber pressure along the trajectory (nearly constant at 1.3bar). The intake pressure recovery schedule (PRF) had to be carefully chosen in order to match the airframe drag characteristic to the engine demand. A relatively low PRF intake (1 oblique and 1 normal shock) was capable of supplying air at the correct pressure while following a trajectory that was close to the airframe minimum drag schedule.
The simple shock structure enabled a lightweight axisymmetric intake to be designed which resulted in efficient intake/engine integration. The intake face was positioned ahead of the wing in close to freestream conditions thereby preventing wing or fuselage boundary layer ingestion and removing the possibility of wing shock interaction. The intake axis was drooped 7deg nosedown relative to the fuselage centreline to minimise the effects of the vehicle incidence. At transition the intake centrebody was translated forward while picking up 3 conical frustrums thereby sealing the nacelle for the rocket ascent and reentry.
At low Mach nos a fixed capture intake ingested more air than was required to feed the core engine. Spilling the excess air around the outside of the nacelle resulted in excessive drag therefore the Sabre nacelle contained an internal bypass system. Above Mac h 2 the intake centrebody position was adjusted to focus the oblique shock on the intake lip thereby ensuring max capture with zero forespill.
The captured airflow divided inside the nacelle part flowing to the core engine while the remainder was diverted down an annular bypass duct. The bypass flow (and intake recovered pressure) were controlled by the variable throat area bypass nozzle. In order to minimise the momentum drag penalty for ingesting the bypass flow, it was necessary to maximise the exhaust velocity of the bypass system. Expanding the bypass flow at the intake recovered temperature would result in net bypass drag due to the intake shock and internal flow losses. However since the core engine was overfuelled it was possible to warm the bypass stream by diverting hydrogen from the core engine to a bypass burner system without increasing the overall fuel flow.
The bypass combustion temperature (and hence thrust) were maximised along the trajectory subject to nozzle choking and structural limits (18 00K). The performance of the bypass system could be further enhanced by operating the intake supercritically (swallowing the normal shock into the expanding subsonic diffuser) giving freedom to select a trajectory that maximised the overall nacelle net thrust while leaving the core engine unaffected. This new trajectory dipped below the original between Mach 2 and 5 resulting from a tradeoff between bypass thrust and intake cowl drag.
Status: Developed -1995.
Specific impulse: 700 s.
Specific impulse sea level: 2,000 s.
Burn time: 730 s.