Summary of Work Completed:

1. AB scale model tests were continued.
2. Negotiations were initiated with Westinghouse Electric & Manufacturing Company for testing the full-size AB in their laboratories.
3. Preliminary design and layout of the full-size experimental AB were completed.
4. Detail drawings and shop work on full-size AB were started.

Testing Details:

A. Equipment: The setup was modified from the prior arrangement (See Figs. 2 – 9). This was done to achieve greater exit velocities. The velocity now reached was 850 ft/sec and 10 inHg total pressure at 1,500°F and was considered the best that could be obtained with the available equipment. Although lower than the actual engine, the conditions were adequate for the determination of the flame holding and combustion characteristics of the special step configuration being tested. Higher pressure in the full-scale AB would compensate for the higher velocity to give comparable performance. A water-cooled total pressure rake was constructed to determine the AB exit total pressures.
B. Fuel Injection and Distribution Tests: Initial tests used atomizing nozzles in the center cone and investigation revealed that when the AB became hot, alternate vapor and solid fuel flowed through the nozzles. The vapor lock causing this was eliminated by constructing a hollow center cone with the atomizing nozzles removed. Injection of the fuel through twelve 1/8" tubing fingers was substituted and these gave good fuel distribution and were used for the remaining scale tests.
C. Outer Shell Tests: The unknown time interval between ignition and complete combustion made scale model combustion chamber length analysis to simulate a comparative length performance in the full-size AB difficult. If it was insignificant, the scale model should give an external flame length comparative to that of the full-size AB. Three different combustion chamber lengths were tried; the shortest showed considerable external burning, the medium and longest showing very little and of nearly equivalent lengths. This gave a 36" minimum full-size AB length and a 54" length from turbine flange to exhaust outlet. All subsequent tests used the long model combustion chamber AB in order to prevent adding any secondary effects to the test data.
D. Outer Flame Holder Tests: The 0.4" diameter tube flame holder reported on previously gave good burning, but had excessive pressure loss. Some was due to the scale effect of the over-scale burner used in the model, but other flame holder configurations were tested. The surface type upstream step gave equally good combustion, but a high cold flow pressure loss similar to that of the round ring flame holder. The AB shell was heated to a red-white color from the step to the exit nozzle. The shielded upstream step flame holder gave performance similar to the surface type, except that the AB shell did not become red-white hot until approximately 12" downstream of the flame holder. The downstream step flame holder gave good burning with a similar high cold flow pressure loss, as in the other types. Starting was difficult at low fuel flows and the range of fuel flows for smooth combustion was narrow. This was the only outer flame holder model showing lean and rich mixture limits. An annulus flame holder gave similar burning and skin temperatures, but a lower cold flow pressure loss than the shielded upstream step flame holder.
E. Diffuser cones and inner flame holder tests: The short diffuser cone with a 0.45" downstream step flame holder was tested without an outer flame holder in order to obtain the cold flow pressure losses caused by the diffusion and inner cone flame holder. The pressure loss was found to be 2.44%. Burner tests in this configuration were satisfactory with good burning being obtained over a wide fuel flow. Flame flickering occurred at low fuel flows and a loud resonating noise occurred at rich mixtures. The shell was red hot only at the exit nozzle. The long diffuser cone with a 0.25" upstream step was tested in a similar manner to the short cone. The pressure loss was 2.37%. The downstream step diffuser cone was expected to have less pressure loss in the full-size AB. Burning tests gave similar results to the short cone.

Summary of Work Completed:

1. Scale model tests of basic AB configurations were completed.
2. A clamshell exit nozzle scale model was constructed.
3. Equipment for clamshell exit nozzle flow test was designed and was being constructed.
4. Negotiations for testing in the Westinghouse Electric & Manufacturing Company laboratories were continued.
5. Detail drawings of full-size AB test models were completed.
6. Construction of full-size AB test model was continued.

Testing Details:

A. Two scale model ABs were constructed and tested, design based on previous reported results.
B. A new water-cooled total pressure rake was used to obtain the final pressure loss data. A new installation location removed the effect of the exit cone pressure loss and the radial components of velocity from the data.
C. The one-third scale Model A-1 (Fig. 11) AB had the best overall performance to date. It had good combustion over a wide range of fuel/air ratios. The external flame length was 4 to 6 times the exit diameter and blue-orange in color. Pressure loss, corrected to full scale conditions, averaged 4 – 5%. The Model A-2 AB proved that it had good possibilities when scaled to full-size at an extra short length. The pressure loss was slightly higher than the Model A-1. Burning tests were inconclusive as the scale size flame holder steps were too small to hold the flame. It was considered likely that the full-sized AB would have the flame front starting at the step and combustion would be completed in the shorter overall AB length.
E. Analysis of Clamshell Nozzle Flow Constants: The full-size AB would incorporate a clamshell type variable nozzle. The model had different coefficients of discharge and velocity at different openings and these had to be determined by calibration. The resulting curves could be used in analysis of afterburner performance. A 1/3 size scale clamshell nozzle was selected for calibration using the method shown in Figure 10.

Summary of Work Completed:

1. The clamshell variable exit nozzle was calibrated.
2. The fabrication of (Full Size) models A1-5000 and A1-5100 ABs was completed.
3. The fabrication of an alternate forward section for the A1-5000 AB was started.
4. The fabrication of an alternate clamshell exit nozzle of Inconel “X” material was started.
5. Ignition test on the scale model was completed.
6. Test work at Westinghouse was started.

Experimental Work:

A. Calibration of the variable exit nozzle was completed in accordance with the method described in the prior report. Figures 14 and 15 depict experimental equipment set up.

Fig. 14. Calibration Blower and Mass Flowmeter Measurement Equipment	Fig. 15. Scale AB Clamshell Calibration Equipment

 

B. While all scale testing did not require direct ignition, the hot gases entering the AB causing automatic ignition, direct ignition was added to preclude problems when full‑size AB testing began. A sparkplug with a 0.125" gap was added in each of the inner and outer flameholder steps of Model A-1. The entering gas temperatures were reduced until self-ignition no longer occurred. At that point, with temperatures between 600 – 980°F, either sparkplug gave instantaneous smooth ignition at low fuel flows. Below 600°F, there was a delay in the ignition between steps if only one sparkplug was used. At a normal blower temperature of 120°F without any preheat, satisfactory electric ignition occurred. When only one sparkplug was used at the ambient temperature, the opposite step would not ignite.
C. Preliminary setup for the full-scale tests was started at Westinghouse on 9 December 1947, with instrumentation of one test cell completed. Calibration of the engine was started but was delayed when hot spots were observed on the turbine. A clogged fuel nozzle was discovered.
D. Full Sized Test Models:

a. The test models A1-5000 and A1-5100 were completed and delivered to Westinghouse for testing.
b. An alternate forward section of the A1-5000 AB was being fabricated. The construction was similar to the original section except that 0.043" thick, Type 347, stainless steel sheet was used instead of 0.064" thick, Type 309.
c. An alternate clamshell exit nozzle was being fabricated of Inconel “X” steel.

Summary of Work Completed:

1. The initial phase of static performance tests on the A1-5000 and A1-5100 AB configuration was completed.
2. The A1-5000 AB with 1" flameholder steps was redesigned for further testing.

Experimental Work:

A. Initial testing was delayed due to hot spots in the engine, rectified by 29 December 1947.
B. Comparisons of data taken before and after the AB tests indicated that the AB caused no unusual deterioration of the engine performance.
C. A cold flow (now termed “dry”) calibration test with the A1-5000 installed with 3/4" flameholder steps, 53" long showed unexpectedly low thrust losses of 3% at 12,500 rpm and an SFC increase of 3.5%.

Fig. 23. Engine WE002062 Calibration Test Results	Fig. 24. Engine WE002062 with A1-5000 AB Attached, Dry Calibration Results

D. Burning tests began on 5 January 1948 and continued until 4 February. The excessive noise produced delayed testing and required muffling. Only 3 hours of AB burning were recorded during the period.

Ignition: The worst internal flow conditions were used for ignition tests. The fixed exit was 18 ¾" and produced low outlet temperatures and high airflow. No AB starts could be obtained above 4,000 rpm. With 1" flameholder steps, spark ignition range was extended to 12,000 rpm at the outer step and 11,500 at the inner step. It was concluded that a sparkplug located at the downstream corner of the step was a satisfactory ignition method. Addition of an independent fuel supply to the sparkplug was expected to raise the smooth ignition range to the required 12,500 rpm.
Total Pressure Recovery: The water-cooled total pressure rake operated satisfactorily for several tests, but finally failed at the tube welds. Vapor pockets had likely formed and permitted overheating.
Combustion: The tests indicated a need for a larger step than the ¾" step being tested. Ignition was difficult and burning was unstable at high rpm. When the 1" inner flameholder step was substituted, the flame held at the step and displayed good stability over a wide range of fuel flows. Several successful tests were made with the ¾" outer step and 1" inner step combination. The exit flames were red-blue in color at the AB exit and extended out in a cone shape for approximately 5 feet. Most of the external flame was transparent blue in color. Tests with 1" inner and outer steps indicated that configuration was the most stable burning of those tests.
Performance: No maximum performance data was obtained during the initial tests. The AB flow rate was limited both by inadequate fuel pressures and the small exit diameters used on the AB.

Fig. 25. Initial Test Data for McDonnell A1-5000 AB

Construction: The initial construction of the first A1-9000 AB caused some delays. The flange gaskets on the combustion shell burned out after short running times. Flanges buckled and caused warping of the shell because of the differential thermal expansion. Simple design changes would eliminate the failures. Structural failure of the inner step flange was caused by fatigue failure of three fuel injection tubes. These likely failed as a result of vibration of the inner fuel manifold which was not anchored firmly. It was observed that the nose contour of the streamlined island supports caused a sharply diverging airflow which deflected the fuel from the nearby nozzles and caused a cool streak in the section of the combustion chamber behind the islands. The initial tests on the A1-5100 AB were unsatisfactory. No stable burning above 5,000 rpm could be obtained. Increasing the flame holder step size might correct the problem, but because the A1-5000 was performing so well, it was decided to postpone the A1-5100 tests until a later date.

Afterburner Redesign for Further Testing: In early February, the A1-5000 was redesigned. The second prototype was a single, all welded unit incorporating the best features of the first model. The 1" inner and outer flameholder steps were used with a constant area combustion chamber of 2.7" diameter and a 26° exit cone angle. The overall length without shroud was 46" from turbine flange to exit. To avoid weakness, gusset reinforcements were added to both the inner and outer steps. The sparkplugs were moved closer to the downstream of the steps and an independent fuel supply was added for each sparkplug. An alternate inner cone fuel injector system of the surface type as was used in the A1-5100 AB was designed. This fuel system had good fuel distribution characteristics and appeared to be worth further testing. An ejector type cooling shroud and a radiation shield were added. It was hoped the assembly would reduce the AB skin temperatures and also protect the surrounding airplane structure from excessive temperatures during flight.
Work Planned for the Next Period:

1. Fabricate redesigned AB.
2. Start second phase of testing at Westinghouse on 1 March 1948.
3. Continue design of AB for flying tests:

a. Modify cooling shroud design as indicated by test results.
b. Design two-position exit nozzle.
c. Design bellows type joint to permit +/-2° AB misalignment.
d. Design control system.
e. Design a quick disconnect joint.
f. Design AB mounting lugs.